| 1 |
Guile NEWS --- history of user-visible changes. |
| 2 |
Copyright (C) 1996-2020 Free Software Foundation, Inc. |
| 3 |
See the end for copying conditions. |
| 4 |
|
| 5 |
Please send Guile bug reports to [email protected]. |
| 6 |
|
| 7 |
|
| 8 |
|
| 9 |
Changes in 2.2.7 (since 2.2.6) |
| 10 |
|
| 11 |
* New interfaces and functionality |
| 12 |
|
| 13 |
** (texinfo plain-text) now exports '*line-width*' fluid |
| 14 |
|
| 15 |
The new '*line-width*' fluid allows users to specify the width of a line |
| 16 |
for the purposes of line wrapping. See "texinfo plain-text" in the |
| 17 |
manual. |
| 18 |
|
| 19 |
* Bug fixes |
| 20 |
|
| 21 |
** Reduce GC pressure when using bignums |
| 22 |
|
| 23 |
Guile no longer installs a finalizer on each bignum (large integer) it |
| 24 |
creates. This significantly improves speed and memory usage on |
| 25 |
applications that make heavy use of bignums, such as the compiler |
| 26 |
itself. |
| 27 |
|
| 28 |
** Fix peval bug that ignored excess arguments |
| 29 |
|
| 30 |
In an expression like: |
| 31 |
|
| 32 |
((lambda () |
| 33 |
(define (add1 n)(+ 1 n)) |
| 34 |
(add1 1 2))) |
| 35 |
|
| 36 |
the compiler (specifically 'peval') would silently ignore the excess |
| 37 |
argument to 'add1'. |
| 38 |
|
| 39 |
** Respect thread local fluid defaults |
| 40 |
|
| 41 |
Previously (fluid-ref (make-thread-local-fluid #t)) would return #f. |
| 42 |
This is now fixed. |
| 43 |
|
| 44 |
** Fix non-deterministic crash in 'finalization_thread_proc' |
| 45 |
(<https://bugs.gnu.org/37757>) |
| 46 |
|
| 47 |
** texinfo properly renders @acronym in plain text |
| 48 |
(<https://bugs.gnu.org/37846>) |
| 49 |
|
| 50 |
** 'scm_port_poll' honors "w" flags |
| 51 |
(<https://bugs.gnu.org/36709>) |
| 52 |
|
| 53 |
** Do not record LDFLAGS in .pc file |
| 54 |
(<https://bugs.gnu.org/36339>) |
| 55 |
|
| 56 |
** Fix Readline configure check for the sake of libedit |
| 57 |
|
| 58 |
This fixes builds on macOS against the system-provided libedit. |
| 59 |
|
| 60 |
** Fix build on platforms where the stack grows upwards |
| 61 |
|
| 62 |
|
| 63 |
|
| 64 |
Changes in 2.2.6 (since 2.2.5) |
| 65 |
|
| 66 |
* Bug fixes |
| 67 |
|
| 68 |
** Fix regression introduced in 2.2.5 that would break HTTP servers |
| 69 |
|
| 70 |
Guile 2.2.5 introduced a bug that would break the built-in HTTP server |
| 71 |
provided by the (web server) module. Specifically, HTTP servers would |
| 72 |
hang while reading requests. See <https://bugs.gnu.org/36350>. |
| 73 |
|
| 74 |
** 'strftime' and 'strptime' honor the current locale encoding |
| 75 |
|
| 76 |
Until now these procedures would wrongfully assume that the locale |
| 77 |
encoding is always UTF-8. See <https://bugs.gnu.org/35920>. |
| 78 |
|
| 79 |
** Re-export 'current-load-port' |
| 80 |
|
| 81 |
This procedure was erroneously removed in the 2.2 series but was still |
| 82 |
documented. |
| 83 |
|
| 84 |
** Minor documentation mistakes were fixed |
| 85 |
|
| 86 |
|
| 87 |
|
| 88 |
Changes in 2.2.5 (since 2.2.4): |
| 89 |
|
| 90 |
* Notable improvements |
| 91 |
|
| 92 |
** Greatly improved performance of bidirectional pipes. |
| 93 |
|
| 94 |
The performance of bidirectional pipes, as created using 'open-pipe' or |
| 95 |
'open-pipe*' in OPEN_BOTH mode, has been greatly improved. When reading |
| 96 |
large blocks of binary data from a bidirectional pipe, the maximum |
| 97 |
bandwidth has been increased by a factor of ~10^3 in some cases. |
| 98 |
|
| 99 |
** New 'get-bytevector-some!' I/O primitive. |
| 100 |
|
| 101 |
This new I/O primitive is similar to 'get-bytevector-some' from the |
| 102 |
R6RS, except that it writes its data to a user-specified range of |
| 103 |
indices in an existing bytevector. As a corollary, it is also now |
| 104 |
possible to specify a maximum number of bytes to read. Note that |
| 105 |
'get-bytevector-some', and now 'get-bytevector-some!', are unique among |
| 106 |
Guile's I/O primitives in their support of efficient binary reads of |
| 107 |
potentially large blocks while also allowing for short reads, to avoid |
| 108 |
undesired blocking. Now these operations can be performed while also |
| 109 |
avoiding heap-allocation. |
| 110 |
|
| 111 |
'get-bytevector-some!' is needed to efficiently implement the new |
| 112 |
bidirectional pipes, which are built upon R6RS custom binary |
| 113 |
input/output ports. |
| 114 |
|
| 115 |
** get-bytevector-{n!,some,some!} now support suspendable I/O. |
| 116 |
|
| 117 |
Scheme implementations of 'get-bytevector-n!', 'get-bytevector-some', |
| 118 |
and 'get-bytevector-some!' have been added to (ice-9 suspendable-ports). |
| 119 |
As a result, these I/O operations now support suspendable I/O. |
| 120 |
|
| 121 |
* Compiler improvements |
| 122 |
|
| 123 |
** guild compile: Add -Wshadowed-toplevel. |
| 124 |
|
| 125 |
Top-level definitions that shadow previous top-level definitions from |
| 126 |
the same compilation unit will now trigger a compile-time warning, if |
| 127 |
-Wshadowed-toplevel is enabled. It is enabled by default. |
| 128 |
|
| 129 |
** guild compile: Add '-x' flag. |
| 130 |
|
| 131 |
Passing "-x EXT" to 'guild compile' will now cause EXT to be recognized |
| 132 |
as a valid source file name extension. For example, to compile R6RS |
| 133 |
code, you might want to pass "-x .sls" so that files ending in ".sls" |
| 134 |
can be found. |
| 135 |
|
| 136 |
* Miscellaneous improvements |
| 137 |
|
| 138 |
** Bootstrap optimization |
| 139 |
|
| 140 |
eval.go and psyntax-pp.go are now built before the rest of the .go files |
| 141 |
so that they are processed by a fast macro expander. This saves time |
| 142 |
when using parallel builds. |
| 143 |
|
| 144 |
** put-u8 now always writes a single byte, regardless of the port encoding. |
| 145 |
|
| 146 |
Previously, (put-u8 PORT OCTET) worked as expected only when writing to |
| 147 |
binary ports, i.e. those with port encoding "ISO-8859-1" a.k.a. Latin-1. |
| 148 |
Strictly speaking, this meets the requirements of the R6RS 'put-u8', |
| 149 |
which need only support binary ports. However, Guile in fact allows |
| 150 |
binary I/O to be performed on any port, and yet 'put-u8' behaved in a |
| 151 |
surprising way with other port encodings: it would perform a _textual_ |
| 152 |
I/O operation, writing the character with Unicode scalar value OCTET. |
| 153 |
Now, 'put-u8' always writes a single byte with value OCTET, regardless |
| 154 |
of the port encoding. |
| 155 |
|
| 156 |
** Optimize fixnum exact integer square roots. |
| 157 |
|
| 158 |
'exact-integer-sqrt' now avoids heap allocation when applied to a |
| 159 |
fixnum. 'sqrt' now avoids heap allocation when applied to a fixnum |
| 160 |
that's a perfect square. Fewer heap allocations are now required when |
| 161 |
applying 'sqrt' to a square of an exact rational whose numerator or |
| 162 |
denominator are fixnums. |
| 163 |
|
| 164 |
** scm_mkstrport: Optimize the POS -> BYTE_POS conversion. |
| 165 |
|
| 166 |
scm_mkstrport now avoids an unnecessary heap allocation and conversion |
| 167 |
to UTF-8, when STR is provided and POS is non-zero. |
| 168 |
|
| 169 |
** SRFI-19: Support ~N in string->date. |
| 170 |
|
| 171 |
Support for the ~N escape, which allows fractions of a second to be |
| 172 |
parsed, is now supported in SRFI-19 'string->date'. |
| 173 |
|
| 174 |
** SRFI-19: Update the leap second table. |
| 175 |
|
| 176 |
The leap on 1 January 2017 was added to SRFI-19's leap second table. |
| 177 |
|
| 178 |
** stexi->shtml: Add support for @i, @math, @tie and @dots. |
| 179 |
|
| 180 |
stexi->shtml is now able to convert @i, @math, @tie and @dots to HTML. |
| 181 |
|
| 182 |
** Define AT_SYMLINK_NOFOLLOW, AT_NO_AUTOMOUNT, and AT_EMPTY_PATH. |
| 183 |
|
| 184 |
AT_SYMLINK_FOLLOW, AT_NO_AUTOMOUNT, and AT_EMPTY_PATH are now available |
| 185 |
from Scheme, if supported on the platform. |
| 186 |
|
| 187 |
** Improvements to the 'time' macro from (ice-9 time). |
| 188 |
|
| 189 |
The 'time' macro now supports expressions that return multiple values. |
| 190 |
It has also been rewritten as a hygienic 'syntax-rules' macro. |
| 191 |
Previously, it was built using 'define-macro', and was therefore |
| 192 |
unhygienic. This is not merely an internal implementation detail, but |
| 193 |
is potentially relevant to any user of the 'time' macro, since it could |
| 194 |
lead to unintended variable capture and other problems. |
| 195 |
|
| 196 |
** Clarify the documentation for 'nil?'. |
| 197 |
See commit b44f505f1571fc9c42e58982f161a9cfc81fb7f4. |
| 198 |
** Clarify the manual's "Processes" section. |
| 199 |
See commit 8cdd3a0773930ca872a13aada7a1344f03bb382b. |
| 200 |
** Avoid 'with-latin1-locale' in binary I/O tests. |
| 201 |
See commit 162a031e5f2c64cd23fcf069fb7b5071196f9527. |
| 202 |
** Update user-visible copyright years. |
| 203 |
|
| 204 |
* Bug fixes |
| 205 |
|
| 206 |
** Avoid regexp ranges in HTTP inter-protocol exploitation check. |
| 207 |
|
| 208 |
The regular expression used to check for HTTP inter-protocol |
| 209 |
exploitation attacks previously used a character range '0-9', whose |
| 210 |
meaning depends on the current locale. This has now been fixed. |
| 211 |
|
| 212 |
** Fixes to the SRFI-19 time/date library. |
| 213 |
|
| 214 |
*** TAI-to-UTC conversion leaps at the wrong time. |
| 215 |
<https://bugs.gnu.org/21911> |
| 216 |
*** time-utc->date shows bogus zone-dependent leap second. |
| 217 |
<https://bugs.gnu.org/22034> |
| 218 |
*** Manual incorrectly describes Julian Date. |
| 219 |
<https://bugs.gnu.org/21902> |
| 220 |
*** date->string duff ISO 8601 negative years. |
| 221 |
<https://bugs.gnu.org/21903> |
| 222 |
*** date->string duff ISO 8601 format for non-4-digit years. |
| 223 |
<https://bugs.gnu.org/21904> |
| 224 |
*** julian-day->date negative input breakage. |
| 225 |
<https://bugs.gnu.org/21906> |
| 226 |
*** time-duration screws up negative durations. |
| 227 |
<https://bugs.gnu.org/26162> |
| 228 |
*** time-difference doesn't detect error of differing time types. |
| 229 |
<https://bugs.gnu.org/26163> |
| 230 |
|
| 231 |
** Improve overflow checks in bytevector, string, and I/O operations. |
| 232 |
|
| 233 |
Several numerical computations, performed using primitive C arithmetic |
| 234 |
in Guile's core bytevector, string, and I/O operations, have been |
| 235 |
rewritten to avoid overflows. |
| 236 |
|
| 237 |
** Fix type inferencing for 'nil?' and 'null?' predicates. |
| 238 |
|
| 239 |
Previously, the compiler would sometimes miscompile certain combinations |
| 240 |
of 'nil?' and 'null?' predicates present within the same top-level form. |
| 241 |
See <https://bugs.gnu.org/33036>. |
| 242 |
|
| 243 |
** Fix 'atomic-box-compare-and-swap!'. |
| 244 |
|
| 245 |
Previously, 'atomic-box-compare-and-swap!' would sometimes spuriously |
| 246 |
fail on architectures based on Load-Linked/Store-Conditional (LL/SC) |
| 247 |
synchronication primitives (e.g. ARM, PowerPC, and MIPS) in a way that |
| 248 |
was undetectable by the caller. See <https://bugs.gnu.org/32786>. |
| 249 |
|
| 250 |
** Make URI handling locale independent. |
| 251 |
|
| 252 |
Previously, procedures in (web uri) would misbehave in some locales |
| 253 |
including sv_SE. See <https://bugs.gnu.org/35785>. |
| 254 |
|
| 255 |
** Strings, i18n: Limit the use of alloca to approximately 8 kilobytes. |
| 256 |
|
| 257 |
Previously, 'string-locale-ci=?', 'string-locale-ci<?', |
| 258 |
'string-locale<?', 'string-locale-{downcase,upcase,titlecase}' and |
| 259 |
'string-normalize-{nfd,nfc,nfkd,nfkc}' would overflow the C stack when |
| 260 |
applied to very large strings. |
| 261 |
|
| 262 |
** Fix documentation of R6RS 'binary-port?' to reflect reality. |
| 263 |
|
| 264 |
Previously, the documentation incorrectly stated that 'binary-port?' |
| 265 |
always returns #t. In fact, it returns #t if and only if the port |
| 266 |
encoding is "ISO-8859-1". The documentation for 'binary-port?' and |
| 267 |
'textual-port?' has been rewritten to reflect the current reality, and |
| 268 |
also to leave open the possibility of changing the behavior of these |
| 269 |
predicates in a future version of Guile. |
| 270 |
|
| 271 |
** Avoid passing NULL to 'memcpy' and 'memcmp'. |
| 272 |
|
| 273 |
Previously, NULL was sometimes passed to 'memcpy' or 'memcmp' when the |
| 274 |
size argument was 0. |
| 275 |
|
| 276 |
** Save and restore errno in the signal handler. |
| 277 |
|
| 278 |
Previously, Guile's synchronous C signal handler failed to save and |
| 279 |
restore 'errno', although it might change its value. This could |
| 280 |
potentially lead to spurious corruptions of 'errno' within threads |
| 281 |
interrupted by signals. |
| 282 |
|
| 283 |
** scm_to_stringn: Avoid passing NULL to c_strcasecmp. |
| 284 |
<https://lists.gnu.org/archive/html/guile-user/2019-05/msg00070.html> |
| 285 |
** r6rs-ports: Accept 'port-position' values greater than 2^32. |
| 286 |
<https://bugs.gnu.org/32161> |
| 287 |
** r6rs-ports: 'put-bytevector' accepts 64-bit integers. |
| 288 |
Fixed in commit 741c45458da0831a12a4f8d729814bf9f2cb6571. |
| 289 |
** Fix R6RS call-with-{input,output}-file to open textual ports. |
| 290 |
<https://bugs.gnu.org/32329> |
| 291 |
** Update (ice-9 match) to include selected bug fixes from upstream. |
| 292 |
*** ice-9/match named match-let is not working |
| 293 |
<https://bugs.gnu.org/22925> |
| 294 |
** open-process: Fix dup(2) and execvp(2) error handling. |
| 295 |
Fixed in commit 521f1ab4709217407496004019c00005d2a82f78. |
| 296 |
** bytevectors: Support large indices in integer accessors. |
| 297 |
Fixed in commit b9cf3517efd4643670d970d2692bc7bede9a85e8. |
| 298 |
** bytevectors: Fix list validation of *list->bytevector procedures. |
| 299 |
<https://bugs.gnu.org/32938> |
| 300 |
** Gracefully handle huge shift counts in 'ash' and 'round-ash'. |
| 301 |
<https://bugs.gnu.org/32644> |
| 302 |
** In 'ash' and 'round-ash', handle right shift count of LONG_MIN. |
| 303 |
<https://bugs.gnu.org/21901> |
| 304 |
** Use 'scm_from_utf8_{string,symbol,keyword}' for C string literals. |
| 305 |
<https://bugs.gnu.org/33044> |
| 306 |
** web: Add support for HTTP header continuation lines. |
| 307 |
Fixed in commit 73cde5ed7218a090ecee888870908af5445796f0. |
| 308 |
** scm_seed_to_random_state: Support wide string arguments. |
| 309 |
<https://bugs.gnu.org/33044> |
| 310 |
** Do not warn the user when 'madvise' returns ENOSYS. |
| 311 |
Fixed in commit 45e4ace6603e00b297e6542362273041aebe7305. |
| 312 |
** Add 'texinfo' as a dependency in the README. |
| 313 |
Fixed in commit 1bbce71501198c3c7abdf07941f5cdc1434858c2. |
| 314 |
** Don't mutate read-only string in ports test. |
| 315 |
Fixed in commit 552f007e91a97f136aad1b22918688b61d03a4a3. |
| 316 |
** Remove redefinition of when & unless in snarf-check-and-output-texi. |
| 317 |
Fixed in commit 1ba5d6f47a54dceee4452a1e7726d2635e5b3449. |
| 318 |
** Fix strftime when Guile is built without threading support. |
| 319 |
Fixed in commit 139c702fc8b61fdeb813c3428fef3701ea8677f9. |
| 320 |
** Avoid leaking a file descriptor in test-unwind. |
| 321 |
Fixed in commit 1437b76777e576b3d000e2f80c5ecdb33a74ac33. |
| 322 |
** Fix binary output on files created by mkstemp!. |
| 323 |
Fixed in commit 78468baa118d316050a27e43250966e52ffd3d54. |
| 324 |
** Fix crypt-on-glibc test error. |
| 325 |
Fixed in commit 27ffbfb0235de466016ea5a6421508f6548971b6. |
| 326 |
** Fix race when expanding syntax-parameterize and define-syntax-parameter. |
| 327 |
<https://bugs.gnu.org/27476#102> |
| 328 |
** Add a fallback value for the locale-monetary-decimal-point. |
| 329 |
Fixed in commit 9ba449643d4c2ac1d2174befca7d765af222bcc0. |
| 330 |
** Handle newlib C library's langinfo constant names. |
| 331 |
Fixed in commit 92105d13ad1363b511214589b7d62d95304beb17. |
| 332 |
** Make locale monetary conversion tests be less strict on terminal whitespace. |
| 333 |
Fixed in commit 2a3ccfb66714efc1c081ea6e921336f80b756d3c. |
| 334 |
** Disable test for current value of setitimer on Cygwin. |
| 335 |
Fixed in commit 3a64c504caaf83e9faf2ec9b7d0e031e1a6a09b9. |
| 336 |
** Fix gc.test "after-gc-hook gets called" failures. |
| 337 |
<https://bugs.gnu.org/31776#17> |
| 338 |
** Update iconv.m4 from gnulib, to fix an iconv leak during configure. |
| 339 |
<https://lists.gnu.org/archive/html/guile-devel/2019-05/msg00011.html> |
| 340 |
** guild compile: Add missing newline in "unrecognized option" error message. |
| 341 |
Fixed in commit 85c5bae4fd94f8686d26fd792b7c0f588c23bd94. |
| 342 |
** 'basename' now correctly handles "/" and "//". |
| 343 |
Fixed in commit 36ad1d24b3d2c174a64c445502a36f19605dbd65. |
| 344 |
** Make srfi-71 visible through 'cond-expand'. |
| 345 |
Fixed in commit 59a06d8392234fbec8b3605cec266a7a0a7b7a56. |
| 346 |
|
| 347 |
|
| 348 |
|
| 349 |
Changes in 2.2.4 (since 2.2.3): |
| 350 |
|
| 351 |
* New interfaces and functionality |
| 352 |
|
| 353 |
** SRFI-71 (Extended LET-syntax for multiple values) |
| 354 |
|
| 355 |
Guile now includes SRFI-71, which extends let, let*, and letrec to |
| 356 |
support assigning multiple values. See "SRFI-71" in the manual for |
| 357 |
details. |
| 358 |
|
| 359 |
** (web client) export 'http-request' procedure |
| 360 |
|
| 361 |
The 'http-request' procedure is the generalized procedure underneath |
| 362 |
'http-get', 'http-post', etc. |
| 363 |
|
| 364 |
** GDB support now registers the 'guile-backtrace' GDB command |
| 365 |
|
| 366 |
The 'guile-backtrace' GDB command displays a backtrace of the VM stack |
| 367 |
for the current thread. |
| 368 |
|
| 369 |
** Recognize RISC-V compilation targets in (system base target) |
| 370 |
|
| 371 |
* Bug fixes |
| 372 |
|
| 373 |
** Fix stack-marking bug affecting multi-threaded programs |
| 374 |
(<https://bugs.gnu.org/28211>) |
| 375 |
|
| 376 |
** Add missing SYNC_IP calls in the VM |
| 377 |
|
| 378 |
These could cause multi-threaded code to crash. |
| 379 |
|
| 380 |
** Fix multi-threaded access to modules |
| 381 |
(<https://bugs.gnu.org/30602>, <https://bugs.gnu.org/31879>, |
| 382 |
and <https://bugs.gnu.org/31878>) |
| 383 |
|
| 384 |
** (ice-9 match) now has better documentation |
| 385 |
|
| 386 |
** 'get-bytevector-n' and 'get-bytevector-n!' can now read more than 4 GB |
| 387 |
|
| 388 |
** Fix cross-compilation support for elisp |
| 389 |
|
| 390 |
** Fix error reporting in 'load-thunk-from-memory' |
| 391 |
|
| 392 |
** Fix GOOPS 'instance?' to work on objects that aren't structs |
| 393 |
(<https://bugs.gnu.org/31606>) |
| 394 |
|
| 395 |
** Fix type inference for bitwise logical operators |
| 396 |
(<https://bugs.gnu.org/31474>) |
| 397 |
|
| 398 |
** Avoid inexact arithmetic in the type inferrer for 'sqrt' |
| 399 |
|
| 400 |
** Fix floating point unboxing regression in 2.2.3 |
| 401 |
(<https://bugs.gnu.org/30020>) |
| 402 |
|
| 403 |
** Fix eta-conversion edge cases in peval (<https://bugs.gnu.org/29520>) |
| 404 |
|
| 405 |
** Correctly interpret SRFI-18 timeout parameters |
| 406 |
(<https://bugs.gnu.org/29704>) |
| 407 |
|
| 408 |
** 'select' returns empty sets upon EINTR and EAGAIN |
| 409 |
(<https://bugs.gnu.org/30368>) |
| 410 |
|
| 411 |
** Restore pre-2.2.3 '%fresh-auto-compile' behavior |
| 412 |
|
| 413 |
This reverts an incorrect fix for <https://bugs.gnu.org/29226>. |
| 414 |
|
| 415 |
|
| 416 |
|
| 417 |
Changes in 2.2.3 (since 2.2.2): |
| 418 |
|
| 419 |
* New interfaces and functionality |
| 420 |
|
| 421 |
** (web uri) module has better support for RFC 3986 |
| 422 |
|
| 423 |
The URI standard, RFC 3986, defines additional "relative-ref" and |
| 424 |
"URI-reference" data types. Thanks to Daniel Hartwig, Guile's support |
| 425 |
for these URI subtypes has been improved. See "Universal Resource |
| 426 |
Identifiers" in the manual, for more. |
| 427 |
|
| 428 |
** `struct-ref/unboxed' and `struct-set!/unboxed' |
| 429 |
|
| 430 |
These procedures should be used when accessing struct fields with type |
| 431 |
`u' (unboxed). See "Structure Basics" in the manual, for full details. |
| 432 |
|
| 433 |
** Improved support for arrays with non-zero lower bounds |
| 434 |
|
| 435 |
Thanks to work by Daniel Llorens, Guile no longer exhibits buggy |
| 436 |
behavior in "sort" or "sort!" on arrays with non-zero lower dimension |
| 437 |
bounds. Arrays with non-zero lower dimension bounds are now allowed for |
| 438 |
array-slice-for-each, and truncated-print now supports bitvectors and |
| 439 |
arrays with non-zero lower bounds. General arrays are now supported as |
| 440 |
well for random:hollow-sphere!. |
| 441 |
|
| 442 |
** Add `uintptr_t' and `intptr_t' to FFI types. |
| 443 |
|
| 444 |
See "Foreign Types" in the manual for full details. |
| 445 |
|
| 446 |
* Compiler improvements |
| 447 |
|
| 448 |
** Improve speed of compiler backend for functions without loops |
| 449 |
|
| 450 |
This is a marginal speed improvement, especially for code compiled with |
| 451 |
optimization level "-O1" or below. |
| 452 |
|
| 453 |
** Disable slot pre-coloring for optimization level "-O1" or below |
| 454 |
|
| 455 |
This improves the speed of the compiler backend. |
| 456 |
|
| 457 |
** Improve complexity of constant subexpression elimination pass |
| 458 |
|
| 459 |
This is a large speed improvement when compiling large files with the |
| 460 |
default "-O2" pass. |
| 461 |
|
| 462 |
** CPS conversion avoids generating return arity adapters if possible |
| 463 |
|
| 464 |
In Guile, the expression in (define a EXP) may return 1 or more values. |
| 465 |
This value elision in "value" context is implicit earlier in the Guile |
| 466 |
compiler, in Tree-IL, but is made explicit in the CPS middle-end |
| 467 |
language by the addition of the equivalent of explicit call-with-values |
| 468 |
continuations that ignore additional values. However in many cases we |
| 469 |
can avoid generating these extra continuations if we know that EXP is |
| 470 |
single-valued, as is the case for example for constants or variable |
| 471 |
references or the like. |
| 472 |
|
| 473 |
Although these "arity-adapting continuations" would be removed by dead |
| 474 |
code elimination at optimization level "-O2" or above, they were still |
| 475 |
being needlessly generated in the first place. Guile now avoids |
| 476 |
generating them, speeding up not only the optimizer at -O2 but also the |
| 477 |
entire compiler pipeline at -O1 or below, as well as improving the |
| 478 |
residual code at -O1 or below. |
| 479 |
|
| 480 |
* New deprecations |
| 481 |
|
| 482 |
** Using `uri?' as a predicate on relative-refs deprecated |
| 483 |
|
| 484 |
If you don't care whether the URI is a relative-ref or not, use |
| 485 |
`uri-reference?'. If you do, use `uri-reference?' and `relative-ref?'. |
| 486 |
In the future `uri?' will return a true value only for URIs that specify |
| 487 |
a scheme. |
| 488 |
|
| 489 |
** Struct tail arrays deprecated |
| 490 |
|
| 491 |
Guile's structures used to have a facility whereby each instance of a |
| 492 |
vtable can contain a variable-length tail array of values. The length |
| 493 |
of the tail array was stored in the structure. This facility was |
| 494 |
originally intended to allow C code to expose raw C structures with |
| 495 |
word-sized tail arrays to Scheme. |
| 496 |
|
| 497 |
However, the tail array facility was confusing and doesn't work very |
| 498 |
well. It was very rarely used, but it insinuates itself into all |
| 499 |
invocations of `make-struct'. For this reason the clumsily-named |
| 500 |
`make-struct/no-tail' procedure can actually be more elegant in actual |
| 501 |
use, because it doesn't have a random `0' argument stuck in the middle. |
| 502 |
|
| 503 |
Tail arrays also inhibit optimization by allowing instances to affect |
| 504 |
their shapes. In the absence of tail arrays, all instances of a given |
| 505 |
vtable have the same number and kinds of fields. This uniformity can be |
| 506 |
exploited by the runtime and the optimizer. The presence of tail arrays |
| 507 |
make some of these optimizations more difficult. |
| 508 |
|
| 509 |
Finally, the tail array facility is ad-hoc and does not compose with the |
| 510 |
rest of Guile. If a Guile user wants an array with user-specified |
| 511 |
length, it's best to use a vector. It is more clear in the code, and |
| 512 |
the standard optimization techniques will do a good job with it. |
| 513 |
|
| 514 |
For all of these reasons, tail arrays are deprecated in Guile 2.2 and |
| 515 |
will be removed from Guile 3.0. Likewise, `make-struct' / |
| 516 |
`scm_make_struct' is deprecated in favor of `make-struct/no-tail' / |
| 517 |
`scm_make_struct_no_tail'. Perhaps one day we will be able to reclaim |
| 518 |
the `make-struct' name! |
| 519 |
|
| 520 |
** Struct "self" slots deprecated |
| 521 |
|
| 522 |
It used to be that you could make a structure vtable that had "self" |
| 523 |
slots. Instances of that vtable would have those slots initialized to |
| 524 |
the instance itself. This can be useful in C code where you might have |
| 525 |
a pointer to the data array, and want to get the `SCM' handle for the |
| 526 |
structure. However this was a little used complication without any use |
| 527 |
in Scheme code. To replace it, just use "p" slots and initialize the |
| 528 |
slot values manually on initialization. |
| 529 |
|
| 530 |
** Struct fields with opaque ("o") protection deprecated |
| 531 |
|
| 532 |
Struct fields are declared with a "protection", meaning read-only ('r'), |
| 533 |
read-write ('w'), or opaque ('o'). There is also "hidden" ('h') which |
| 534 |
is read-write but which isn't initialized by arguments passed to |
| 535 |
`make-struct/no-tail', but that's a detail. Opaque struct fields were |
| 536 |
used to allocate storage in a struct that could only be accessed by C. |
| 537 |
This facility was very rarely used (unused in Guile itself) but now that |
| 538 |
we are implementing more and more in Scheme, it is completely useless. |
| 539 |
|
| 540 |
To enforce permissions on struct fields, instead layer on an abstraction |
| 541 |
at a higher level, in the same way that immutable record fields are |
| 542 |
simply those which don't have an accessor. |
| 543 |
|
| 544 |
** Using `struct-ref' and `struct-set!' on unboxed fields is deprecated |
| 545 |
|
| 546 |
Use the new `struct-ref/unboxed' and `struct-set!/unboxed' instead. |
| 547 |
|
| 548 |
* Bug fixes |
| 549 |
|
| 550 |
** guile.m4 now checks for Guile 2.2 by default |
| 551 |
|
| 552 |
Before, it was still preferring Guile 2.0. It now also supports the |
| 553 |
Guile 3.0 prereleases. |
| 554 |
|
| 555 |
** Fix setting breakpoints from the REPL |
| 556 |
|
| 557 |
** Allow GDB support to be used with GDB linked against Guile 2.0. |
| 558 |
|
| 559 |
** Fix deadlock in `readdir' on error. |
| 560 |
|
| 561 |
** Fix crash on ia64 during thread switches. |
| 562 |
|
| 563 |
** Fix bug inferring range of `logand' computations with negative numbers |
| 564 |
|
| 565 |
** Fix bug when issuing HTTP requests through proxies. |
| 566 |
|
| 567 |
** Refactor weak hash table implementation to be more robust |
| 568 |
|
| 569 |
Guile 2.2's weak hash table implementation had three big problems. The |
| 570 |
first was a bug causing these tables to leak memory when they would be |
| 571 |
resized. The second was that the implementation was designed so that |
| 572 |
tables should be visited by the mark phase of the garbage collector in |
| 573 |
one big piece. This could cause the garbage collector to see too many |
| 574 |
newly marked objects at once, causing inefficies in garbage collection. |
| 575 |
Finally, the way in which lost weak references were ultimately removed |
| 576 |
from weak tables caused a race between the finalizer threads and the |
| 577 |
mutator threads, leading to unbounded excess space retention in |
| 578 |
pathological cases. All of this problems have been fixed. |
| 579 |
|
| 580 |
** Allow garbage collection of revealed file ports |
| 581 |
|
| 582 |
Guile can mark a file port as "revealed" if Scheme has been given access |
| 583 |
to the file descriptor. In that case, the file descriptor will not be |
| 584 |
closed when the port is garbage-collected. However we had a bug that |
| 585 |
for revealed ports prevented the port from ever being garbage-collected, |
| 586 |
leading to memory leaks of Guile's internal port buffers. This is now |
| 587 |
fixed. |
| 588 |
|
| 589 |
** Fix put-bytevector, unget-bytevector with start == bytevector length |
| 590 |
|
| 591 |
** Enable GNU Readline 7.0's support for "bracketed paste". |
| 592 |
|
| 593 |
Before, when pasting an expression that contained TAB characters into |
| 594 |
Guile's REPL with GNU Readline support enabled, the pasted TAB |
| 595 |
characters would trigger autocompletion in Readline. This was never |
| 596 |
what you wanted. Guile now sets the new "bracketed-paste" option in GNU |
| 597 |
Readline 7.0 to on by default, making readline treat pastes into the |
| 598 |
terminal as atomic units without control characters. See "Readline |
| 599 |
Options" in the manual for full details. |
| 600 |
|
| 601 |
** Fix time-monotonic from SRFI-19; broken in 2.2.1. |
| 602 |
|
| 603 |
|
| 604 |
Changes in 2.2.2 (since 2.2.1): |
| 605 |
|
| 606 |
* Bug fixes |
| 607 |
|
| 608 |
** Syntax objects are once more comparable with 'equal?' |
| 609 |
|
| 610 |
The syntax object change in 2.2.1 had the unintended effect of making |
| 611 |
syntax objects no longer comparable with equal?. This release restores |
| 612 |
the previous behavior. |
| 613 |
|
| 614 |
** Restore libgc dependency |
| 615 |
|
| 616 |
The change to throw exceptions when mutating literal constants partly |
| 617 |
relied on an interface that was added to our garbage collector (BDW-GC) |
| 618 |
after its 7.2 release. Guile 2.2.2 adds a workaround to allow Guile to |
| 619 |
continue be used with libgc as old as 7.2. |
| 620 |
|
| 621 |
** SRFI-37 bug fix to not error on empty-string arguments. |
| 622 |
|
| 623 |
Thanks to Thomas Danckaert for fixing this long-standing bug. |
| 624 |
|
| 625 |
|
| 626 |
|
| 627 |
Changes in 2.2.1 (since 2.2.0): |
| 628 |
|
| 629 |
* Notable changes |
| 630 |
|
| 631 |
** New sandboxed evaluation facility |
| 632 |
|
| 633 |
Guile now has a way to execute untrusted code in a safe way. See |
| 634 |
"Sandboxed Evaluation" in the manual for full details, including some |
| 635 |
important notes on limitations on the sandbox's ability to prevent |
| 636 |
resource exhaustion. |
| 637 |
|
| 638 |
** All literal constants are read-only |
| 639 |
|
| 640 |
According to the Scheme language definition, it is an error to attempt |
| 641 |
to mutate a "constant literal". A constant literal is data that is a |
| 642 |
literal quoted part of a program. For example, all of these are errors: |
| 643 |
|
| 644 |
(set-car! '(1 . 2) 42) |
| 645 |
(append! '(1 2 3) '(4 5 6)) |
| 646 |
(vector-set! '#(a b c) 1 'B) |
| 647 |
|
| 648 |
Guile takes advantage of this provision of Scheme to deduplicate shared |
| 649 |
structure in constant literals within a compilation unit, and to |
| 650 |
allocate constant data directly in the compiled object file. If the |
| 651 |
data needs no relocation at run-time, as is the case for pairs or |
| 652 |
vectors that only contain immediate values, then the data can actually |
| 653 |
be shared between different Guile processes, using the operating |
| 654 |
system's virtual memory facilities. |
| 655 |
|
| 656 |
However, in Guile 2.2.0, constants that needed relocation were actually |
| 657 |
mutable -- though (vector-set! '#(a b c) 1 'B) was an error, Guile |
| 658 |
wouldn't actually cause an exception to be raised, silently allowing the |
| 659 |
mutation. This could affect future users of this constant, or indeed of |
| 660 |
any constant in the compilation unit that shared structure with the |
| 661 |
original vector. |
| 662 |
|
| 663 |
Additionally, attempting to mutate constant literals mapped in the |
| 664 |
read-only section of files would actually cause a segmentation fault, as |
| 665 |
the operating system prohibits writes to read-only memory. "Don't do |
| 666 |
that" isn't a very nice solution :) |
| 667 |
|
| 668 |
Both of these problems have been fixed. Any attempt to mutate a |
| 669 |
constant literal will now raise an exception, whether the constant needs |
| 670 |
relocation or not. |
| 671 |
|
| 672 |
** Syntax objects are now a distinct type |
| 673 |
|
| 674 |
It used to be that syntax objects were represented as a tagged vector. |
| 675 |
These values could be forged by users to break scoping abstractions, |
| 676 |
preventing the implementation of sandboxing facilities in Guile. We are |
| 677 |
as embarrassed about the previous situation as we pleased are about the |
| 678 |
fact that we've fixed it. |
| 679 |
|
| 680 |
Unfortunately, during the 2.2 stable series (or at least during part of |
| 681 |
it), we need to support files compiled with Guile 2.2.0. These files |
| 682 |
may contain macros that contain legacy syntax object constants. See the |
| 683 |
discussion of "allow-legacy-syntax-objects?" in "Syntax Transformer |
| 684 |
Helpers" in the manual for full details. |
| 685 |
|
| 686 |
* Bug fixes |
| 687 |
|
| 688 |
*** Fix snarfing with -ggdb3 (#25803) |
| 689 |
*** Fix spurious snarf warnings for net_db.c |
| 690 |
*** Output statprof flat display to correct port |
| 691 |
*** Document guile-2.2 cond-expand feature |
| 692 |
*** Add --with-bdw-gc for BSDs that use bdw-gc-threaded (see README) |
| 693 |
*** Documentation typo fixes (#26188) |
| 694 |
*** Fix SRFI-9 date->string bugs with ~N and ~F (#26261, #26260, #26259) |
| 695 |
*** SRFI-19 current-time-monotonic returns time of right type (#26329) |
| 696 |
*** Avoid causing GC when looking up exception handler |
| 697 |
*** Increment objcode version, in a compatible way |
| 698 |
*** Fix compile warning in (system base types) |
| 699 |
*** Only run tests that require fork if it is provided |
| 700 |
*** Speed up procedure-minimum-arity for fixed arity |
| 701 |
*** REPL server tests catch ECONNABORTED |
| 702 |
*** Avoid deprecated argument to setvbuf in (web client) |
| 703 |
*** Remove non-existent 'open-connection-for-uri' export from (web client) |
| 704 |
|
| 705 |
|
| 706 |
Changes in 2.2.0 (changes since the 2.0.x stable release series): |
| 707 |
|
| 708 |
* Notable changes |
| 709 |
|
| 710 |
** Speed |
| 711 |
|
| 712 |
The biggest change in Guile 2.2 is a complete rewrite of its virtual |
| 713 |
machine and compiler internals. The result is faster startup time, |
| 714 |
better memory usage, and faster execution of user code. See the |
| 715 |
"Performance improvements" section below for more details. |
| 716 |
|
| 717 |
** Better thread-safety |
| 718 |
|
| 719 |
This new release series takes the ABI-break opportunity to fix some |
| 720 |
interfaces that were difficult to use correctly from multiple threads. |
| 721 |
Notably, weak hash tables and ports are now transparently thread-safe. |
| 722 |
See "Scheduling" in the manual, for updated documentation on threads and |
| 723 |
communications primitives. |
| 724 |
|
| 725 |
** Better space-safety |
| 726 |
|
| 727 |
It used to be the case that, when calling a Scheme procedure, the |
| 728 |
procedure and arguments were always preserved against garbage |
| 729 |
collection. This is no longer the case; Guile is free to collect the |
| 730 |
procedure and arguments if they become unreachable, or to re-use their |
| 731 |
slots for other local variables. Guile still offers good-quality |
| 732 |
backtraces by determining the procedure being called from the |
| 733 |
instruction pointer instead of from the value in slot 0 of an |
| 734 |
application frame, and by using a live variable map that allows the |
| 735 |
debugger to know which locals are live at all points in a frame. |
| 736 |
|
| 737 |
** Off-main-thread finalization |
| 738 |
|
| 739 |
Following Guile 2.0.6's change to invoke finalizers via asyncs, Guile |
| 740 |
2.2 takes the additional step of invoking finalizers from a dedicated |
| 741 |
finalizer thread, if threads are enabled. This avoids concurrency |
| 742 |
issues between finalizers and application code, and also speeds up |
| 743 |
finalization. If your application's finalizers are not robust to the |
| 744 |
presence of threads, see "Foreign Objects" in the manual for information |
| 745 |
on how to disable automatic finalization and instead run finalizers |
| 746 |
manually. |
| 747 |
|
| 748 |
** Better locale support in Guile scripts |
| 749 |
|
| 750 |
When Guile is invoked directly, either from the command line or via a |
| 751 |
hash-bang line (e.g. "#!/usr/bin/guile"), it now installs the current |
| 752 |
locale via a call to `(setlocale LC_ALL "")'. For users with a unicode |
| 753 |
locale, this makes all ports unicode-capable by default, without the |
| 754 |
need to call `setlocale' in your program. This behavior may be |
| 755 |
controlled via the GUILE_INSTALL_LOCALE environment variable; see |
| 756 |
"Environment Variables" in the manual, for more. |
| 757 |
|
| 758 |
** Complete Emacs-compatible Elisp implementation |
| 759 |
|
| 760 |
Thanks to the work of Robin Templeton, Guile's Elisp implementation is |
| 761 |
now fully Emacs-compatible, implementing all of Elisp's features and |
| 762 |
quirks in the same way as the editor we know and love. |
| 763 |
|
| 764 |
** Dynamically expandable stacks |
| 765 |
|
| 766 |
Instead of allocating fixed stack sizes for running Scheme code, Guile |
| 767 |
now starts off each thread with only one page of stack, and expands and |
| 768 |
shrinks it dynamically as needed. Guile will throw an exception for |
| 769 |
stack overflows if growing the stack fails. It is also possible to |
| 770 |
impose a stack limit during the extent of a function call. See "Stack |
| 771 |
Overflow" in the manual, for more. |
| 772 |
|
| 773 |
This change allows users to write programs that use the stack as a data |
| 774 |
structure for pending computations, as it was meant to be, without |
| 775 |
reifying that data out to the heap. Where you would previously make a |
| 776 |
loop that collect its results in reverse order only to re-reverse them |
| 777 |
at the end, now you can just recurse without worrying about stack |
| 778 |
overflows. |
| 779 |
|
| 780 |
Using the stack also allows more code to be continuation-safe. For |
| 781 |
example, returning multiple times from a `map' procedure in Guile 2.0 |
| 782 |
would change the value of previously returned result lists, because |
| 783 |
`map' built its result list in reverse order then used `reverse!' to |
| 784 |
return the proper result. Now in Guile 2.2, `map' is implemented using |
| 785 |
straightforward recursion, which eliminates this bug while maintaining |
| 786 |
good performance as well as good space complexity. |
| 787 |
|
| 788 |
** Out-of-memory improvements |
| 789 |
|
| 790 |
Instead of aborting, failures to allocate memory will now raise an |
| 791 |
unwind-only `out-of-memory' exception, and cause the corresponding |
| 792 |
`catch' expression to run garbage collection in order to free up memory. |
| 793 |
|
| 794 |
** GOOPS core reimplemented in Scheme |
| 795 |
|
| 796 |
Guile's object orientation system, GOOPS, has been mostly reimplemented |
| 797 |
in Scheme. This decreases its maintenance burden on the rest of Guile, |
| 798 |
while also makes it possible to implement new features in the future, |
| 799 |
such as method combinations or `eqv?' specializers. |
| 800 |
|
| 801 |
** Better handling of GUILE_LOAD_COMPILED_PATH |
| 802 |
|
| 803 |
It used to be that Guile would stop at the first .go file it found in |
| 804 |
the GUILE_LOAD_COMPILED_PATH. If that file turned out to be out of |
| 805 |
date, then no .go file would be loaded. Now Guile will continue to |
| 806 |
search the path for a file which is both present and up-to-date, with |
| 807 |
respect to the .scm file. |
| 808 |
|
| 809 |
** C99 required |
| 810 |
|
| 811 |
Following Emacs, you must use a C99-capable compiler when building |
| 812 |
Guile. In the future we also expect require C99 to use Guile's C |
| 813 |
interface, at least for `stdint' support. |
| 814 |
|
| 815 |
** Lightweight pre-emptive threading primitives |
| 816 |
|
| 817 |
The compiler now inserts special "handle-interrupts" opcodes before each |
| 818 |
call, return, and backwards jump target. This allows the user to |
| 819 |
interrupt any computation and to accurately profile code using |
| 820 |
interrupts. It used to be that interrupts were run by calling a C |
| 821 |
function from the VM; now interrupt thunks are run directly from the VM. |
| 822 |
This allows interrupts to save a delimited continuation and, if the |
| 823 |
continuation was established from the same VM invocation (the usual |
| 824 |
restriction), that continuation can then be resumed. In this way users |
| 825 |
can implement lightweight pre-emptive threading facilities. |
| 826 |
|
| 827 |
** with-dynamic-state in VM |
| 828 |
|
| 829 |
Similarly, `with-dynamic-state' no longer recurses out of the VM, |
| 830 |
allowing captured delimited continuations that include a |
| 831 |
`with-dynamic-state' invocation to be resumed. This is a precondition |
| 832 |
to allow lightweight threading libraries to establish a dynamic state |
| 833 |
per lightweight fiber. |
| 834 |
|
| 835 |
* Performance improvements |
| 836 |
|
| 837 |
** Faster programs via new virtual machine |
| 838 |
|
| 839 |
Guile now compiles programs to instructions for a new virtual machine. |
| 840 |
The new virtual machine's instructions can address their source and |
| 841 |
destination operands by "name" (slot). This makes access to named |
| 842 |
temporary values much faster than in Guile 2.0, and removes a lot of |
| 843 |
value-shuffling that the old virtual machine had to do. The end result |
| 844 |
is that loop-heavy code can be two or three times as fast with Guile 2.2 |
| 845 |
as in 2.0. Your mileage may vary, of course; see "A Virtual Machine for |
| 846 |
Guile" in the manual for the nitties and the gritties. |
| 847 |
|
| 848 |
** Better startup time, memory usage with ELF object file format |
| 849 |
|
| 850 |
Guile now uses the standard ELF format for its compiled code. (Guile |
| 851 |
has its own loader and linker, so this does not imply a dependency on |
| 852 |
any particular platform's ELF toolchain.) The benefit is that Guile is |
| 853 |
now able to statically allocate more data in the object files. ELF also |
| 854 |
enables more sharing of data between processes, and decreases startup |
| 855 |
time (about 40% faster than the already fast startup of the Guile 2.0 |
| 856 |
series). Guile also uses DWARF for some of its debugging information. |
| 857 |
Much of the debugging information can be stripped from the object files |
| 858 |
as well. See "Object File Format" in the manual, for full details. |
| 859 |
|
| 860 |
** Better optimizations via compiler rewrite |
| 861 |
|
| 862 |
Guile's compiler now uses a Continuation-Passing Style (CPS) |
| 863 |
intermediate language, allowing it to reason easily about temporary |
| 864 |
values and control flow. Examples of optimizations that this permits |
| 865 |
are optimal contification, optimal common subexpression elimination, |
| 866 |
dead code elimination, loop-invariant code motion, loop peeling, loop |
| 867 |
inversion, parallel moves with at most one temporary, allocation of |
| 868 |
stack slots using precise liveness information, unboxing of 64-bit |
| 869 |
integers and floating point values, and closure optimization. For more, |
| 870 |
see "Continuation-Passing Style" in the manual. |
| 871 |
|
| 872 |
** Faster interpreter |
| 873 |
|
| 874 |
Combined with a number of optimizations to the interpreter itself, |
| 875 |
simply compiling `eval.scm' with the new compiler yields an interpreter |
| 876 |
that is consistently two or three times faster than the one in Guile |
| 877 |
2.0. |
| 878 |
|
| 879 |
** Allocation-free dynamic stack |
| 880 |
|
| 881 |
Guile now implements the dynamic stack with an actual stack instead of a |
| 882 |
list of heap objects, avoiding most allocation. This speeds up prompts, |
| 883 |
the `scm_dynwind_*' family of functions, fluids, and `dynamic-wind'. |
| 884 |
|
| 885 |
** Optimized UTF-8 and Latin-1 ports, symbols, and strings |
| 886 |
|
| 887 |
Guile 2.2 is faster at reading and writing UTF-8 and Latin-1 strings |
| 888 |
from ports, and at converting symbols and strings to and from these |
| 889 |
encodings. |
| 890 |
|
| 891 |
** Optimized hash functions |
| 892 |
|
| 893 |
Guile 2.2 now uses Bob Jenkins' `hashword2' (from his `lookup3.c') for |
| 894 |
its string hash, and Thomas Wang's integer hash function for `hashq' and |
| 895 |
`hashv'. These functions produce much better hash values across all |
| 896 |
available fixnum bits. |
| 897 |
|
| 898 |
** Optimized generic array facility |
| 899 |
|
| 900 |
Thanks to work by Daniel Llorens, the generic array facility is much |
| 901 |
faster now, as it is internally better able to dispatch on the type of |
| 902 |
the underlying backing store. |
| 903 |
|
| 904 |
** All ports are now buffered, can be targets of `setvbuf' |
| 905 |
|
| 906 |
See "Buffering" in the manual, for more. A port with a buffer size of 1 |
| 907 |
is equivalent to an unbuffered port. Ports may set their default buffer |
| 908 |
sizes, and some ports (for example soft ports) are unbuffered by default |
| 909 |
for historical reasons. |
| 910 |
|
| 911 |
** Mutexes are now faster under contention |
| 912 |
|
| 913 |
Guile implements its own mutexes, so that threads that are trying to |
| 914 |
acquire a mutex can be interrupted. These mutexes used to be quite |
| 915 |
inefficient when many threads were trying to acquire them, causing many |
| 916 |
spurious wakeups and contention. This has been fixed. |
| 917 |
|
| 918 |
* New interfaces |
| 919 |
|
| 920 |
** New `cond-expand' feature: `guile-2.2' |
| 921 |
|
| 922 |
Use this feature if you need to check for Guile 2.2 from Scheme code. |
| 923 |
|
| 924 |
** New predicate: `nil?' |
| 925 |
|
| 926 |
See "Nil" in the manual. |
| 927 |
|
| 928 |
** New compiler modules |
| 929 |
|
| 930 |
Since the compiler was rewritten, there are new modules for the back-end |
| 931 |
of the compiler and the low-level loader and introspection interfaces. |
| 932 |
See the "Guile Implementation" chapter in the manual for all details. |
| 933 |
|
| 934 |
** Add "tree" display mode for statprof. |
| 935 |
|
| 936 |
See the newly updated "Statprof" section of the manual, for more. |
| 937 |
|
| 938 |
** Support for non-blocking I/O |
| 939 |
|
| 940 |
See "Non-Blocking I/O" in the manual, for more. |
| 941 |
|
| 942 |
** Implement R6RS custom binary input/output ports |
| 943 |
|
| 944 |
See "Custom Ports" in the manual. |
| 945 |
|
| 946 |
** Implement R6RS output-buffer-mode |
| 947 |
** Implement R6RS bytevector->string, string->bytevector |
| 948 |
|
| 949 |
See "R6RS Transcoders" in the manual. |
| 950 |
|
| 951 |
** `accept' now takes optional flags argument |
| 952 |
|
| 953 |
These flags can include `SOCK_NONBLOCK' and `SOCK_CLOEXEC', indicating |
| 954 |
options to apply to the returned socket, potentially removing the need |
| 955 |
for additional system calls to set these options. See "Network Sockets |
| 956 |
and Communication" in the manual, for more. |
| 957 |
|
| 958 |
** Thread-safe atomic boxes (references) |
| 959 |
|
| 960 |
See "Atomics" in the manual. |
| 961 |
|
| 962 |
** Thread-local fluids |
| 963 |
|
| 964 |
Guile now has support for fluids whose values are not captured by |
| 965 |
`current-dynamic-state' and not inheritied by child threads, and thus |
| 966 |
are local to the kernel thread they run on. See "Thread-Local |
| 967 |
Variables" in the manual, for more. |
| 968 |
|
| 969 |
** suspendable-continuation? |
| 970 |
|
| 971 |
This predicate returns true if the delimited continuation captured by |
| 972 |
aborting to a prompt would be able to be resumed. See "Prompt |
| 973 |
Primitives" in the manual for more. |
| 974 |
|
| 975 |
** scm_c_prepare_to_wait_on_fd, scm_c_prepare_to_wait_on_cond, |
| 976 |
** scm_c_wait_finished |
| 977 |
|
| 978 |
See "Asyncs" in the manual for more. |
| 979 |
|
| 980 |
** File descriptor finalizers |
| 981 |
|
| 982 |
See "Ports and File Descriptors" in the manual. |
| 983 |
|
| 984 |
** New inline functions: `scm_new_smob', `scm_new_double_smob' |
| 985 |
|
| 986 |
These can replace many uses of SCM_NEWSMOB, SCM_RETURN_NEWSMOB2, and the |
| 987 |
like. See XXX in the manual, for more. |
| 988 |
|
| 989 |
** New low-level type accessors |
| 990 |
|
| 991 |
For more on `SCM_HAS_TYP7', `SCM_HAS_TYP7S', `SCM_HAS_TYP16', see XXX. |
| 992 |
|
| 993 |
`SCM_HEAP_OBJECT_P' is now an alias for the inscrutable `SCM_NIMP'. |
| 994 |
|
| 995 |
`SCM_UNPACK_POINTER' and `SCM_PACK_POINTER' are better-named versions of |
| 996 |
the old `SCM2PTR' and `PTR2SCM'. Also, `SCM_UNPACK_POINTER' yields a |
| 997 |
void*. |
| 998 |
|
| 999 |
** `TCP_NODELAY' and `TCP_CORK' socket options, if provided by the system |
| 1000 |
|
| 1001 |
** `scm_c_put_latin1_chars', `scm_c_put_utf32_chars' |
| 1002 |
|
| 1003 |
Use these instead of `scm_lfwrite'. See the new "Using Ports from C" |
| 1004 |
section of the manual, for more. |
| 1005 |
|
| 1006 |
** <standard-vtable>, standard-vtable-fields |
| 1007 |
|
| 1008 |
See "Structures" in the manual for more on these. |
| 1009 |
|
| 1010 |
** Convenience utilities for ports and strings. |
| 1011 |
|
| 1012 |
See "Conversion to/from C" for more on `scm_from_port_string', |
| 1013 |
`scm_from_port_stringn', `scm_to_port_string', and |
| 1014 |
`scm_to_port_stringn'. |
| 1015 |
|
| 1016 |
** New expressive PEG parser |
| 1017 |
|
| 1018 |
See "PEG Parsing" in the manual for more. Thanks to Michael Lucy for |
| 1019 |
originally writing these, and to Noah Lavine for integration work. |
| 1020 |
|
| 1021 |
** `make-stack' now also works on delimited continuations |
| 1022 |
|
| 1023 |
** Better URI-reference support |
| 1024 |
|
| 1025 |
The `(web uri)' module now has interfaces for handling URI references, |
| 1026 |
which might not have a scheme. The Location header of a web request or |
| 1027 |
response is now a URI reference instead of a URI. Also, |
| 1028 |
`request-absolute-uri' now has an optional default scheme argument. See |
| 1029 |
"Web" in the manual for full details. |
| 1030 |
|
| 1031 |
** formal-name->char, char->formal-name |
| 1032 |
|
| 1033 |
See "Characters", in the manual. |
| 1034 |
|
| 1035 |
* Incompatible changes |
| 1036 |
|
| 1037 |
** ASCII is not ISO-8859-1 |
| 1038 |
|
| 1039 |
In Guile 2.0, if a user set "ASCII" or "ANSI_X3.4-1968" as the encoding |
| 1040 |
of a port, Guile would treat it as ISO-8859-1. While these encodings |
| 1041 |
are the same for codepoints 0 to 127, ASCII does not extend past that |
| 1042 |
range, whereas ISO-8859-1 goes up to 255. Guile 2.2 no longer treats |
| 1043 |
ASCII as ISO-8859-1. This is likely to be a problem only if the user's |
| 1044 |
locale is set to ASCII, and the user or a program writes non-ASCII |
| 1045 |
codepoints to a port. |
| 1046 |
|
| 1047 |
** Decoding errors do not advance the read pointer before erroring |
| 1048 |
|
| 1049 |
When the user sets a port's conversion strategy to "error", indicating |
| 1050 |
that Guile should throw an error if it tries to read from a port whose |
| 1051 |
incoming bytes are not valid for the port's encoding, it used to be that |
| 1052 |
Guile would advance the read pointer past the bad bytes, and then throw |
| 1053 |
an error. This would allow the following `read-char' invocation to |
| 1054 |
proceed after the bad bytes. This behavior is incompatible with the |
| 1055 |
final R6RS standard, and besides contravenes the user's intention to |
| 1056 |
raise an error on bad input. Guile now raises an error without |
| 1057 |
advancing the read pointer. To skip over a bad encoding, set the port |
| 1058 |
conversion strategy to "substitute" and read a substitute character. |
| 1059 |
|
| 1060 |
** Decoding errors with `substitute' strategy return U+FFFD |
| 1061 |
|
| 1062 |
It used to be that decoding errors with the `substitute' conversion |
| 1063 |
strategy would replace the bad bytes with a `?' character. This has |
| 1064 |
been changed to use the standard U+FFFD REPLACEMENT CHARACTER, in |
| 1065 |
accordance with the Unicode recommendations. |
| 1066 |
|
| 1067 |
** API to define new port types from C has changed |
| 1068 |
|
| 1069 |
Guile's ports have been completely overhauled to allow Guile developers |
| 1070 |
and eventually Guile users to write low-level input and output routines |
| 1071 |
in Scheme. The new internals will eventually allow for user-space |
| 1072 |
tasklets or green threads that suspend to a scheduler when they would |
| 1073 |
cause blocking I/O, allowing users to write straightforward network |
| 1074 |
services that parse their input and send their output as if it were |
| 1075 |
blocking, while under the hood Guile can multiplex many active |
| 1076 |
connections at once. |
| 1077 |
|
| 1078 |
At the same time, this change makes Guile's ports implementation much |
| 1079 |
more maintainable, rationalizing the many legacy port internals and |
| 1080 |
making sure that the abstractions between the user, Guile's core ports |
| 1081 |
facility, and the port implementations result in a system that is as |
| 1082 |
performant and expressive as possible. |
| 1083 |
|
| 1084 |
The interface to the user has no significant change, neither on the C |
| 1085 |
side nor on the Scheme side. However this refactoring has changed the |
| 1086 |
interface to the port implementor in an incompatible way. See the newly |
| 1087 |
expanded "I/O Extensions" in the manual, for full details. |
| 1088 |
|
| 1089 |
*** Remove `scm_set_port_mark' |
| 1090 |
|
| 1091 |
Port mark functions have not been called since the switch to the BDW |
| 1092 |
garbage collector. |
| 1093 |
|
| 1094 |
*** Remove `scm_set_port_equalp' |
| 1095 |
|
| 1096 |
Likewise port equal functions weren't being called. Given that ports |
| 1097 |
have their own internal buffers, it doesn't make sense to hook them into |
| 1098 |
equal? anyway. |
| 1099 |
|
| 1100 |
*** Remove `scm_set_port_free' |
| 1101 |
|
| 1102 |
It used to be that if an open port became unreachable, a special "free" |
| 1103 |
function would be called instead of the "close" function. Now that the |
| 1104 |
BDW-GC collector allows us to run arbitrary code in finalizers, we can |
| 1105 |
simplify to just call "close" on the port and remove the separate free |
| 1106 |
functions. Note that hooking into the garbage collector has some |
| 1107 |
overhead. For that reason Guile exposes a new interface, |
| 1108 |
`scm_set_port_needs_close_on_gc', allowing port implementations to |
| 1109 |
indicate to Guile whether they need closing on GC or not. |
| 1110 |
|
| 1111 |
*** Remove `scm_set_port_end_input', `scm_set_port_flush' |
| 1112 |
|
| 1113 |
As buffering is handled by Guile itself, these functions which were to |
| 1114 |
manage an implementation-side buffer are no longer needed. |
| 1115 |
|
| 1116 |
*** Change prototype of `scm_make_port_type' |
| 1117 |
|
| 1118 |
The `read' (renamed from `fill_input') and `write' functions now operate |
| 1119 |
on bytevectors. Also the `mode_bits' argument now inplicitly includes |
| 1120 |
SCM_OPN, so you don't need to include these. |
| 1121 |
|
| 1122 |
*** Change prototype of port `close' function |
| 1123 |
|
| 1124 |
The port close function now returns void. |
| 1125 |
|
| 1126 |
*** Port and port type data structures are now opaque |
| 1127 |
|
| 1128 |
Port type implementations should now use API to access port state. |
| 1129 |
However, since the change to handle port buffering centrally, port type |
| 1130 |
implementations rarely need to access unrelated port state. |
| 1131 |
|
| 1132 |
*** Port types are now `scm_t_port_type*', not a tc16 value |
| 1133 |
|
| 1134 |
`scm_make_port_type' now returns an opaque pointer, not a tc16. |
| 1135 |
Relatedly, the limitation that there only be 256 port types has been |
| 1136 |
lifted. |
| 1137 |
|
| 1138 |
** String ports default to UTF-8 |
| 1139 |
|
| 1140 |
Guile 2.0 would use the `%default-port-encoding' when creating string |
| 1141 |
ports. This resulted in ports that could only accept a subset of valid |
| 1142 |
characters, which was surprising to users. Now string ports default to |
| 1143 |
the UTF-8 encoding. Sneaky users can still play encoding conversion |
| 1144 |
games with string ports by explicitly setting the encoding of a port |
| 1145 |
after it is open. See "Ports" in the manual for more. |
| 1146 |
|
| 1147 |
** `scm_from_stringn' and `scm_to_stringn' encoding arguments are never NULL |
| 1148 |
|
| 1149 |
These functions now require a valid `encoding' argument, and will abort |
| 1150 |
if given `NULL'. |
| 1151 |
|
| 1152 |
** All r6rs ports are both textual and binary |
| 1153 |
|
| 1154 |
Because R6RS ports are a thin layer on top of Guile's ports, and Guile's |
| 1155 |
ports are both textual and binary, Guile's R6RS ports are also both |
| 1156 |
textual and binary, and thus both kinds have port transcoders. This is |
| 1157 |
an incompatibility with respect to R6RS. |
| 1158 |
|
| 1159 |
** Threading facilities moved to (ice-9 threads) |
| 1160 |
|
| 1161 |
It used to be that call-with-new-thread and other threading primitives |
| 1162 |
were available in the default environment. This is no longer the case; |
| 1163 |
they have been moved to (ice-9 threads) instead. Existing code will not |
| 1164 |
break, however; we used the deprecation facility to signal a warning |
| 1165 |
message while also providing these bindings in the root environment for |
| 1166 |
the duration of the 2.2 series. |
| 1167 |
|
| 1168 |
** cancel-thread uses asynchronous interrupts, not pthread_cancel |
| 1169 |
|
| 1170 |
See "Asyncs" in the manual, for more on asynchronous interrupts. |
| 1171 |
|
| 1172 |
** SRFI-18 threads, mutexes, cond vars disjoint from Guile |
| 1173 |
|
| 1174 |
When we added support for the SRFI-18 threading library in Guile 2.0, we |
| 1175 |
did so in a way that made SRFI-18 mutexes the same as Guile mutexes. |
| 1176 |
This was a mistake. In Guile our goal is to provide basic, |
| 1177 |
well-thought-out, well-implemented, minimal primitives, on top of which |
| 1178 |
we can build a variety of opinionated frameworks. Incorporating SRFI-18 |
| 1179 |
functionality into core Guile caused us to bloat and slow down our core |
| 1180 |
threading primitives. Worse, they became very hard to describe; they |
| 1181 |
did many things, did them poorly, and all that they did was never |
| 1182 |
adequately specified. |
| 1183 |
|
| 1184 |
For all of these reasons we have returned to a situation where SRFI-18 |
| 1185 |
concepts are implemented only in the `(srfi srfi-18)' module. This |
| 1186 |
means that SRFI-18 threads are built on Guile threads, but aren't the |
| 1187 |
same as Guile threads; calling Guile `thread?' on a thread no longer |
| 1188 |
returns true. |
| 1189 |
|
| 1190 |
We realize this causes inconvenience to users who use both Guile |
| 1191 |
threading interfaces and SRFI-18 interfaces, and we lament the change -- |
| 1192 |
but we are better off now. We hope the newly revised "Scheduling" |
| 1193 |
section in the manual compensates for the headache. |
| 1194 |
|
| 1195 |
** Remove `lock-mutex' "owner" argument |
| 1196 |
|
| 1197 |
Mutex owners are a SRFI-18 concept; use SRFI-18 mutexes instead. |
| 1198 |
Relatedly, `scm_lock_mutex_timed' taking the owner argument is now |
| 1199 |
deprecated; use `scm_timed_lock_mutex' instead. |
| 1200 |
|
| 1201 |
** Remove `unlock-mutex' cond var and timeout arguments |
| 1202 |
|
| 1203 |
It used to be that `unlock-mutex' included `wait-condition-variable' |
| 1204 |
functionality. This has been deprecated; use SRFI-18 if you want this |
| 1205 |
behavior from `mutex-unlock!'. Relatedly, `scm_unlock_mutex_timed' is |
| 1206 |
deprecated; use `scm_unlock_mutex' instead. |
| 1207 |
|
| 1208 |
** Removed `unchecked-unlock' mutex flag |
| 1209 |
|
| 1210 |
This flag was introduced for internal use by SRFI-18; use SRFI-18 |
| 1211 |
mutexes if you need this behaviour. |
| 1212 |
|
| 1213 |
** SRFI-18 mutexes no longer recursive |
| 1214 |
|
| 1215 |
Contrary to specification, SRFI-18 mutexes in Guile were recursive. |
| 1216 |
This is no longer the case. |
| 1217 |
|
| 1218 |
** Thread cleanup handlers removed |
| 1219 |
|
| 1220 |
The `set-thread-cleanup!' and `thread-cleanup' functions that were added |
| 1221 |
in Guile 2.0 to support cleanup after thread cancellation are no longer |
| 1222 |
needed, since threads can declare cleanup handlers via `dynamic-wind'. |
| 1223 |
|
| 1224 |
** Only threads created by Guile are joinable |
| 1225 |
|
| 1226 |
`join-thread' used to work on "foreign" threads that were not created by |
| 1227 |
Guile itself, though their join value was always `#f'. This is no |
| 1228 |
longer the case; attempting to join a foreign thread will throw an |
| 1229 |
error. |
| 1230 |
|
| 1231 |
** Dynamic states capture values, not locations |
| 1232 |
|
| 1233 |
Dynamic states used to capture the locations of fluid-value |
| 1234 |
associations. Capturing the current dynamic state then setting a fluid |
| 1235 |
would result in a mutation of that captured state. Now capturing a |
| 1236 |
dynamic state simply captures the current values, and calling |
| 1237 |
`with-dynamic-state' copies those values into the Guile virtual machine |
| 1238 |
instead of aliasing them in a way that could allow them to be mutated in |
| 1239 |
place. This change allows Guile's fluid variables to be thread-safe. |
| 1240 |
To capture the locations of a dynamic state, capture a |
| 1241 |
`with-dynamic-state' invocation using partial continuations instead. |
| 1242 |
|
| 1243 |
** Remove `frame-procedure' |
| 1244 |
|
| 1245 |
Several optimizations in Guile make `frame-procedure' an interface that |
| 1246 |
we can no longer support. For background, `frame-procedure' used to |
| 1247 |
return the value at slot 0 in a frame, which usually corresponds to the |
| 1248 |
SCM value of the procedure being applied. However it could be that this |
| 1249 |
slot is re-used for some other value, because the closure was not needed |
| 1250 |
in the function. Such a re-use might even be for an untagged value, in |
| 1251 |
which case treating slot 0 as a SCM value is quite dangerous. It's also |
| 1252 |
possible that so-called "well-known" closures (closures whose callers |
| 1253 |
are all known) are optimized in such a way that slot 0 is not a |
| 1254 |
procedure but some optimized representation of the procedure's free |
| 1255 |
variables. Instead, developers building debugging tools that would like |
| 1256 |
access to `frame-procedure' are invited to look at the source for the |
| 1257 |
`(system vm frame)' module for alternate interfaces, including the new |
| 1258 |
`frame-procedure-name'. |
| 1259 |
|
| 1260 |
** Remove `,procedure' REPL command |
| 1261 |
|
| 1262 |
Not all procedures have values, so it doesn't make sense to expose this |
| 1263 |
interface to the user. Instead, the `,locals' REPL command will include |
| 1264 |
the callee, if it is live. |
| 1265 |
|
| 1266 |
** Remove `frame-local-ref', `frame-local-set!', `frame-num-locals' |
| 1267 |
|
| 1268 |
These procedures reference values in a frame on the stack. Since we now |
| 1269 |
have unboxed values of different kinds, it is now necessary to specify |
| 1270 |
the type when reference locals, and once this incompatible change needs |
| 1271 |
to be made, we might as well make these interfaces private. See |
| 1272 |
"Frames' in the manual, for more information on the replacements for |
| 1273 |
these low-level interfaces. |
| 1274 |
|
| 1275 |
** Vtable hierarchy changes |
| 1276 |
|
| 1277 |
In an attempt to make Guile's structure and record types integrate |
| 1278 |
better with GOOPS by unifying the vtable hierarchy, `make-vtable-vtable' |
| 1279 |
is now deprecated. Instead, users should just use `make-vtable' with |
| 1280 |
appropriate arguments. See "Structures" in the manual for all of the |
| 1281 |
details. As such, `record-type-vtable' and `%condition-type-vtable' now |
| 1282 |
have a parent vtable and are no longer roots of the vtable hierarchy. |
| 1283 |
|
| 1284 |
** Syntax parameters are a distinct type |
| 1285 |
|
| 1286 |
Guile 2.0's transitional implementation of `syntax-parameterize' was |
| 1287 |
based on the `fluid-let-syntax' interface inherited from the psyntax |
| 1288 |
expander. This interface allowed any binding to be dynamically rebound |
| 1289 |
-- even bindings like `lambda'. This is no longer the case in Guile |
| 1290 |
2.2. Syntax parameters must be defined via `define-syntax-parameter', |
| 1291 |
and only such bindings may be parameterized. See "Syntax Parameters" in |
| 1292 |
the manual for more. |
| 1293 |
|
| 1294 |
** Defined identifiers scoped in the current module |
| 1295 |
|
| 1296 |
Sometimes Guile's expander would attach incorrect module scoping |
| 1297 |
information for top-level bindings made by an expansion. For example, |
| 1298 |
given the following R6RS library: |
| 1299 |
|
| 1300 |
(library (defconst) |
| 1301 |
(export defconst) |
| 1302 |
(import (guile)) |
| 1303 |
(define-syntax-rule (defconst name val) |
| 1304 |
(begin |
| 1305 |
(define t val) |
| 1306 |
(define-syntax-rule (name) t)))) |
| 1307 |
|
| 1308 |
Attempting to use it would produce an error: |
| 1309 |
|
| 1310 |
(import (defconst)) |
| 1311 |
(defconst foo 42) |
| 1312 |
(foo) |
| 1313 |
=| Unbound variable: t |
| 1314 |
|
| 1315 |
It wasn't clear that we could fix this in Guile 2.0 without breaking |
| 1316 |
someone's delicate macros, so the fix is only coming out now. |
| 1317 |
|
| 1318 |
** Pseudo-hygienically rename macro-introduced bindings |
| 1319 |
|
| 1320 |
Bindings introduced by macros, like `t' in the `defconst' example above, |
| 1321 |
are now given pseudo-fresh names. This allows |
| 1322 |
|
| 1323 |
(defconst foo 42) |
| 1324 |
(defconst bar 37) |
| 1325 |
|
| 1326 |
to introduce different bindings for `t'. These pseudo-fresh names are |
| 1327 |
made in such a way that if the macro is expanded again, for example as |
| 1328 |
part of a simple recompilation, the introduced identifiers get the same |
| 1329 |
pseudo-fresh names. See "Hygiene and the Top-Level" in the manual, for |
| 1330 |
details. |
| 1331 |
|
| 1332 |
** Fix literal matching for module-bound literals |
| 1333 |
|
| 1334 |
`syntax-rules' and `syntax-case' macros can take a set of "literals": |
| 1335 |
bound or unbound keywords that the syntax matcher treats specially. |
| 1336 |
Before, literals were always matched symbolically (by name). Now they |
| 1337 |
are matched by binding. This allows literals to be reliably bound to |
| 1338 |
values, renamed by imports or exports, et cetera. See "Syntax-rules |
| 1339 |
Macros" in the manual for more on literals. |
| 1340 |
|
| 1341 |
** Fix bug importing specific bindings with #:select |
| 1342 |
|
| 1343 |
It used to be that if #:select didn't find a binding in the public |
| 1344 |
interface of a module, it would actually grovel in the module's |
| 1345 |
unexported private bindings. This was not intended and is now fixed. |
| 1346 |
|
| 1347 |
** Statically scoped module duplicate handlers |
| 1348 |
|
| 1349 |
It used to be that if a module did not specify a #:duplicates handler, |
| 1350 |
when a name was first referenced in that module and multiple imported |
| 1351 |
modules provide that name, the value of the |
| 1352 |
`default-duplicate-binding-handlers' parameter would be used to resolve |
| 1353 |
the duplicate bindings. We have changed so that instead a module |
| 1354 |
defaults to the set of handlers described in the manual. If the module |
| 1355 |
specifies #:duplicates, of course we use that. The |
| 1356 |
`default-duplicate-binding-handlers' parameter now simply accesses the |
| 1357 |
handlers of the current module, instead of some global value. |
| 1358 |
|
| 1359 |
** Fix too-broad capture of dynamic stack by delimited continuations |
| 1360 |
|
| 1361 |
Guile was using explicit stacks to represent, for example, the chain of |
| 1362 |
current exception handlers. This means that a delimited continuation |
| 1363 |
that captured a "catch" expression would capture the whole stack of |
| 1364 |
exception handlers, not just the exception handler added by the "catch". |
| 1365 |
This led to strangeness when resuming the continuation in some other |
| 1366 |
context like other threads; "throw" could see an invalid stack of |
| 1367 |
exception handlers. This has been fixed by the addition of the new |
| 1368 |
"fluid-ref*" procedure that can access older values of fluids; in this |
| 1369 |
way the exception handler stack is now implicit. See "Fluids and |
| 1370 |
Dynamic States" in the manual, for more on fluid-ref*. |
| 1371 |
|
| 1372 |
** `dynamic-wind' doesn't check that guards are thunks |
| 1373 |
|
| 1374 |
Checking that the dynamic-wind out-guard procedure was actually a thunk |
| 1375 |
before doing the wind was slow, unreliable, and not strictly needed. |
| 1376 |
|
| 1377 |
** All deprecated code removed |
| 1378 |
|
| 1379 |
All code deprecated in Guile 2.0 has been removed. See older NEWS, and |
| 1380 |
check that your programs can compile without linker warnings and run |
| 1381 |
without runtime warnings. See "Deprecation" in the manual. |
| 1382 |
|
| 1383 |
In particular, the following functions, which were deprecated in 2.0.10 |
| 1384 |
but not specifically mentioned earlier in this file, have been removed: |
| 1385 |
|
| 1386 |
*** `uniform-vector-read!' and `uniform-vector-write' have been |
| 1387 |
removed. Use `get-bytevector-n!' and `put-bytevector' from (rnrs io |
| 1388 |
ports) instead. |
| 1389 |
|
| 1390 |
** Remove miscellaneous unused interfaces |
| 1391 |
|
| 1392 |
We have removed accidentally public, undocumented interfaces that we |
| 1393 |
think are not used, and not useful. This includes `scm_markstream', |
| 1394 |
`SCM_FLUSH_REGISTER_WINDOWS', `SCM_THREAD_SWITCHING_CODE', `SCM_FENCE', |
| 1395 |
`scm_call_generic_0', `scm_call_generic_1', `scm_call_generic_2' |
| 1396 |
`scm_call_generic_3', `scm_apply_generic', and `scm_program_source'. |
| 1397 |
`scm_async_click' was renamed to `scm_async_tick', and `SCM_ASYNC_TICK' |
| 1398 |
was made private (use `SCM_TICK' instead). |
| 1399 |
|
| 1400 |
** Many internal compiler / VM changes |
| 1401 |
|
| 1402 |
As the compiler and virtual machine were re-written, there are many |
| 1403 |
changes in the back-end of Guile to interfaces that were introduced in |
| 1404 |
Guile 2.0. These changes are only only of interest if you wrote a |
| 1405 |
language on Guile 2.0 or a tool using Guile 2.0 internals. If this is |
| 1406 |
the case, drop by the IRC channel to discuss the changes. |
| 1407 |
|
| 1408 |
** Defining a SMOB or port type no longer mucks exports of `(oop goops)' |
| 1409 |
|
| 1410 |
It used to be that defining a SMOB or port type added an export to |
| 1411 |
GOOPS, for the wrapper class of the smob type. This violated |
| 1412 |
modularity, though, so we have removed this behavior. |
| 1413 |
|
| 1414 |
** Bytecode replaces objcode as a target language |
| 1415 |
|
| 1416 |
One way in which people may have used details of Guile's runtime in |
| 1417 |
Guile 2.0 is in compiling code to thunks for later invocation. Instead |
| 1418 |
of compiling to objcode and then calling `make-program', now the way to |
| 1419 |
do it is to compile to `bytecode' and then call `load-thunk-from-memory' |
| 1420 |
from `(system vm loader)'. |
| 1421 |
|
| 1422 |
** Weak pairs removed |
| 1423 |
|
| 1424 |
Weak pairs were not safe to access with `car' and `cdr', and so were |
| 1425 |
removed. |
| 1426 |
|
| 1427 |
** Weak alist vectors removed |
| 1428 |
|
| 1429 |
Use weak hash tables instead. |
| 1430 |
|
| 1431 |
** Weak vectors may no longer be accessed via `vector-ref' et al |
| 1432 |
|
| 1433 |
Weak vectors may no longer be accessed with the vector interface. This |
| 1434 |
was a source of bugs in the 2.0 Guile implementation, and a limitation |
| 1435 |
on using vectors as building blocks for other abstractions. Vectors in |
| 1436 |
Guile are now a concrete type; for an abstract interface, use the |
| 1437 |
generic array facility (`array-ref' et al). |
| 1438 |
|
| 1439 |
** scm_t_array_implementation removed |
| 1440 |
|
| 1441 |
This interface was introduced in 2.0 but never documented. It was a |
| 1442 |
failed attempt to layer the array implementation that actually |
| 1443 |
introduced too many layers, as it prevented the "vref" and "vset" |
| 1444 |
members of scm_t_array_handle (called "ref" and "set" in 1.8, not |
| 1445 |
present in 2.0) from specializing on array backing stores. |
| 1446 |
|
| 1447 |
Notably, the definition of scm_t_array_handle has now changed, to not |
| 1448 |
include the (undocumented) "impl" member. We are sorry for any |
| 1449 |
inconvenience this may cause. |
| 1450 |
|
| 1451 |
** `scm_make' is now equivalent to Scheme `make' |
| 1452 |
|
| 1453 |
It used to be that `scm_make' only implemented a hard-wired object |
| 1454 |
allocation and initialization protocol. This was because `scm_make' was |
| 1455 |
used while GOOPS booted its own, more complete `make' implementation in |
| 1456 |
Scheme. Now that we've re-implemented everything in Scheme, the C |
| 1457 |
`scm_make' now dispatches directly to Scheme `make', which implements |
| 1458 |
the full protocol. This change is incompatible in some ways, but on the |
| 1459 |
whole is good news for GOOPS users. |
| 1460 |
|
| 1461 |
** GOOPS slot definitions are now objects |
| 1462 |
|
| 1463 |
Slot definitions are now instances of a <slot> class, instead of being |
| 1464 |
specially formatted lists. To most user code, this is transparent, as |
| 1465 |
the slot definition accessors like `slot-definition-name' continue to |
| 1466 |
work. However, code that for example uses `car' to get the name of a |
| 1467 |
slot definition will need to be updated to use the accessors. |
| 1468 |
|
| 1469 |
** Class slot changes |
| 1470 |
|
| 1471 |
Class objects no longer have a `default-slot-definition-class' slot, |
| 1472 |
which was never used. They also no longer have slots for hashsets |
| 1473 |
(`h0', `h1', and so on up to `h7'), which have been unused since Guile |
| 1474 |
2.0 and were not a great idea. |
| 1475 |
|
| 1476 |
There is a new class option, `#:static-slot-allocation?'. See the |
| 1477 |
manual for details. |
| 1478 |
|
| 1479 |
** Removal of internal, unintentionally exposed GOOPS C interfaces |
| 1480 |
|
| 1481 |
These include: `scm_sys_fast_slot_ref', `scm_sys_fast_slot_set_x' |
| 1482 |
`scm_basic_basic_make_class', `scm_sys_compute_slots', |
| 1483 |
`scm_sys_prep_layout_x' `scm_t_method', `SCM_METHOD', |
| 1484 |
`scm_s_slot_set_x', `SCM_CLASS_CLASS_LAYOUT', `scm_si_slotdef_class', |
| 1485 |
`scm_si_generic_function', `scm_si_specializers', `scm_si_procedure', |
| 1486 |
`scm_si_formals', `scm_si_body', `scm_si_make_procedure', |
| 1487 |
`SCM_CLASS_CLASS_LAYOUT', `SCM_INSTANCE_HASH', `SCM_SET_HASHSET', `union |
| 1488 |
scm_t_debug_info', `scm_pure_generic_p', `SCM_PUREGENERICP', |
| 1489 |
`SCM_VALIDATE_PUREGENERIC', `SCM_VTABLE_FLAG_GOOPS_PURE_GENERIC', |
| 1490 |
`SCM_CLASSF_PURE_GENERIC', `scm_c_extend_primitive_generic', |
| 1491 |
`scm_sys_initialize_object', `SCM_CLASS_CLASS_LAYOUT', |
| 1492 |
`scm_si_redefined', `scm_si_direct_supers', `scm_si_direct_slots', |
| 1493 |
`scm_si_direct_subclasses', `scm_si_direct_methods', `scm_si_cpl' |
| 1494 |
`scm_si_slots', `scm_si_getters_n_setters', `SCM_N_CLASS_SLOTS', |
| 1495 |
`SCM_OBJ_CLASS_REDEF', `SCM_INST', `SCM_ACCESSORS_OF', |
| 1496 |
`scm_sys_allocate_instance', and `scm_sys_invalidate_class_x'. |
| 1497 |
|
| 1498 |
* New deprecations |
| 1499 |
|
| 1500 |
** `SCM_FDES_RANDOM_P' |
| 1501 |
|
| 1502 |
Instead, use `lseek (fd, 0, SEEK_CUR)' directly. |
| 1503 |
|
| 1504 |
** `_IONBF', `_IOLBF', and `_IOFBF' |
| 1505 |
|
| 1506 |
Instead, use the symbol values `none', `line', or `block', respectively, |
| 1507 |
as arguments to the `setvbuf' function. |
| 1508 |
|
| 1509 |
** `SCM_FDES_RANDOM_P' |
| 1510 |
|
| 1511 |
Instead, use `lseek (fd, 0, SEEK_CUR)' directly. |
| 1512 |
|
| 1513 |
** Arbiters |
| 1514 |
|
| 1515 |
Arbiters were an experimental mutual exclusion facility from 20 years |
| 1516 |
ago that didn't survive the test of time. Use mutexes or atomic boxes |
| 1517 |
instead. |
| 1518 |
|
| 1519 |
** User asyncs |
| 1520 |
|
| 1521 |
Guile had (and still has) "system asyncs", which are asynchronous |
| 1522 |
interrupts, and also had this thing called "user asyncs", which was a |
| 1523 |
trivial unused data structure. Now that we have deprecated the old |
| 1524 |
`async', `async-mark', and `run-asyncs' procedures that comprised the |
| 1525 |
"user async" facility, we have been able to clarify our documentation to |
| 1526 |
only refer to "asyncs". |
| 1527 |
|
| 1528 |
** Critical sections |
| 1529 |
|
| 1530 |
Critical sections have long been just a fancy way to lock a mutex and |
| 1531 |
defer asynchronous interrupts. Instead of SCM_CRITICAL_SECTION_START, |
| 1532 |
make sure you're in a "scm_dynwind_begin (0)" and use |
| 1533 |
scm_dynwind_pthread_mutex_lock instead, possibly also with |
| 1534 |
scm_dynwind_block_asyncs. |
| 1535 |
|
| 1536 |
** `scm_make_mutex_with_flags' |
| 1537 |
|
| 1538 |
Use `scm_make_mutex_with_kind' instead. See "Mutexes and Condition |
| 1539 |
Variables" in the manual, for more. |
| 1540 |
|
| 1541 |
** Dynamic roots |
| 1542 |
|
| 1543 |
This was a facility that predated threads, was unused as far as we can |
| 1544 |
tell, and was never documented. Still, a grep of your code for |
| 1545 |
dynamic-root or dynamic_root would not be amiss. |
| 1546 |
|
| 1547 |
** `make-dynamic-state' |
| 1548 |
|
| 1549 |
Use `current-dynamic-state' to get an immutable copy of the current |
| 1550 |
fluid-value associations. |
| 1551 |
|
| 1552 |
** `with-statprof' macro |
| 1553 |
|
| 1554 |
Use the `statprof' procedure instead. |
| 1555 |
|
| 1556 |
** SCM_WTA_DISPATCH_0, SCM_WTA_DISPATCH_1, SCM_WTA_DISPATCH_2, SCM_WTA_DISPATCH_N |
| 1557 |
** SCM_GASSERT0, SCM_GASSERT1, SCM_GASSERT2, SCM_GASSERTn |
| 1558 |
** SCM_WTA_DISPATCH_1_SUBR |
| 1559 |
|
| 1560 |
These macros were used in dispatching primitive generics. They can be |
| 1561 |
replaced by using C functions (the same name but in lower case), if |
| 1562 |
needed, but this is a hairy part of Guile that perhaps you shouldn't be |
| 1563 |
using. |
| 1564 |
|
| 1565 |
** scm_compute_applicable_methods and scm_find_method |
| 1566 |
|
| 1567 |
Use `compute-applicable-methods' from Scheme instead. |
| 1568 |
|
| 1569 |
** scm_no_applicable_method |
| 1570 |
|
| 1571 |
Fetch no-applicable-method from the GOOPS exports if you need it. |
| 1572 |
|
| 1573 |
** scm_class_boolean, scm_class_char, scm_class_pair |
| 1574 |
** scm_class_procedure, scm_class_string, scm_class_symbol |
| 1575 |
** scm_class_primitive_generic, scm_class_vector, scm_class_null |
| 1576 |
** scm_class_real, scm_class_complex, scm_class_integer |
| 1577 |
** scm_class_fraction, scm_class_unknown, scm_class_top |
| 1578 |
** scm_class_object, scm_class_class, scm_class_applicable |
| 1579 |
** scm_class_applicable_struct, scm_class_applicable_struct_with_setter |
| 1580 |
** scm_class_generic, scm_class_generic_with_setter, scm_class_accessor |
| 1581 |
** scm_class_extended_generic, scm_class_extended_generic_with_setter |
| 1582 |
** scm_class_extended_accessor, scm_class_method |
| 1583 |
** scm_class_accessor_method, scm_class_procedure_class |
| 1584 |
** scm_class_applicable_struct_class, scm_class_number, scm_class_list |
| 1585 |
** scm_class_keyword, scm_class_port, scm_class_input_output_port |
| 1586 |
** scm_class_input_port, scm_class_output_port, scm_class_foreign_slot |
| 1587 |
** scm_class_self, scm_class_protected, scm_class_hidden |
| 1588 |
** scm_class_opaque, scm_class_read_only, scm_class_protected_hidden |
| 1589 |
** scm_class_protected_opaque, scm_class_protected_read_only |
| 1590 |
** scm_class_scm, scm_class_int, scm_class_float, scm_class_double |
| 1591 |
** scm_port_class, scm_smob_class |
| 1592 |
|
| 1593 |
These class exports are now deprecated. Instead, look up the ones you |
| 1594 |
need from the GOOPS module, or use `scm_class_of' on particular values. |
| 1595 |
|
| 1596 |
** scm_get_keyword |
| 1597 |
|
| 1598 |
Instead from Scheme use kw-arg-ref or real keyword arguments, and from C |
| 1599 |
use `scm_c_bind_keyword_arguments'. |
| 1600 |
|
| 1601 |
** scm_slot_ref_using_class, scm_slot_set_using_class_x |
| 1602 |
** scm_slot_bound_using_class_p, scm_slot_exists_using_class_p |
| 1603 |
|
| 1604 |
Instead use the normal `scm_slot_ref' and similar procedures. |
| 1605 |
|
| 1606 |
* Changes to the distribution |
| 1607 |
|
| 1608 |
** Pre-built binary files in the tarball |
| 1609 |
|
| 1610 |
Building Guile from a tarball can now take advantage of a "prebuilt/" |
| 1611 |
tree of prebuilt .go files. These compiled files are created when a |
| 1612 |
tarball is made, and are used to speed up the build for users of |
| 1613 |
official releases. |
| 1614 |
|
| 1615 |
These pre-built binaries are not necessary, however: they are not stored |
| 1616 |
in revision control and can always be re-created from the source, given |
| 1617 |
that Guile can bootstrap itself from its minimal bootstrap C |
| 1618 |
interpreter. If you do not want to depend on these pre-built binaries, |
| 1619 |
you can "make -C prebuilt clean" before building. |
| 1620 |
|
| 1621 |
** New minor version |
| 1622 |
|
| 1623 |
The "effective version" of Guile is now 2.2, which allows parallel |
| 1624 |
installation with other effective versions (for example, the older Guile |
| 1625 |
2.0). See "Parallel Installations" in the manual for full details. |
| 1626 |
Notably, the `pkg-config' file is now `guile-2.2'. |
| 1627 |
|
| 1628 |
** Bump required libgc version to 7.2, released March 2012. |
| 1629 |
|
| 1630 |
** GUILE_PROGS searches for versioned Guile |
| 1631 |
|
| 1632 |
The GUILE_PROGS autoconf macro can take a required version argument. As |
| 1633 |
a new change, that version argument is additionally searched for as a |
| 1634 |
suffix. For example, GUILE_PROGS(2.2) would look for guile-2.2, |
| 1635 |
guile2.2, guile-2, guile2, and then guile. The found prefix is also |
| 1636 |
applied to guild, guile-config, and the like. Thanks to Freja Nordsiek |
| 1637 |
for this work. |
| 1638 |
|
| 1639 |
** The readline extension is now installed in the extensionsdir |
| 1640 |
|
| 1641 |
The shared library that implements Guile's readline extension is no |
| 1642 |
longer installed to the libdir. This change should be transparent to |
| 1643 |
users, but packagers may be interested. |
| 1644 |
|
| 1645 |
|
| 1646 |
|
| 1647 |
Changes in 2.0.14 (since 2.0.13): |
| 1648 |
|
| 1649 |
* Bug fixes |
| 1650 |
|
| 1651 |
** Builds of .go files and of Guile itself are now bit-reproducible |
| 1652 |
(<http://bugs.gnu.org/20272>) |
| 1653 |
|
| 1654 |
** 'number->locale-string' and 'monetary-amount->locale-string' fixes |
| 1655 |
(<http://bugs.gnu.org/24990>) |
| 1656 |
|
| 1657 |
** (system base target) now recognizes "sh3" as a cross-compilation target |
| 1658 |
|
| 1659 |
** Fix race condition in '00-repl-server.test' |
| 1660 |
(<http://bugs.gnu.org/24769>) |
| 1661 |
|
| 1662 |
** 'scandir' from (ice-9 ftw) no longer calls 'stat' for each entry |
| 1663 |
|
| 1664 |
** Several documentation improvements |
| 1665 |
|
| 1666 |
|
| 1667 |
Changes in 2.0.13 (since 2.0.12): |
| 1668 |
|
| 1669 |
* Security fixes |
| 1670 |
|
| 1671 |
** CVE-2016-8606: REPL server now protects against HTTP inter-protocol |
| 1672 |
attacks |
| 1673 |
|
| 1674 |
Guile 2.x provides a "REPL server" started by the '--listen' |
| 1675 |
command-line option or equivalent API (see "REPL Servers" in the |
| 1676 |
manual). |
| 1677 |
|
| 1678 |
The REPL server is vulnerable to the HTTP inter-protocol attack as |
| 1679 |
described at |
| 1680 |
<https://en.wikipedia.org/wiki/Inter-protocol_exploitation>, notably the |
| 1681 |
HTML form protocol attack described at |
| 1682 |
<https://www.jochentopf.com/hfpa/hfpa.pdf>. A "DNS rebinding attack" |
| 1683 |
can be combined with this attack and allow an attacker to send arbitrary |
| 1684 |
Guile code to the REPL server through web pages accessed by the |
| 1685 |
developer, even though the REPL server is listening to a loopback device |
| 1686 |
("localhost"). This was demonstrated in an article entitled "How to |
| 1687 |
steal any developer's local database" available at |
| 1688 |
<http://bouk.co/blog/hacking-developers/>. |
| 1689 |
|
| 1690 |
The REPL server in Guile 2.0.13 now detects attempts to exploit this |
| 1691 |
vulnerability. It immediately closes the connection when it receives a |
| 1692 |
line that looks like an HTTP request. |
| 1693 |
|
| 1694 |
Nevertheless, we recommend binding the REPL server to a Unix-domain |
| 1695 |
socket, for instance by running: |
| 1696 |
|
| 1697 |
guile --listen=/tmp/guile-socket |
| 1698 |
|
| 1699 |
** CVE-2016-8605: 'mkdir' procedure no longer calls umask(2) |
| 1700 |
(<http://bugs.gnu.org/24659>) |
| 1701 |
|
| 1702 |
When the second argument to the 'mkdir' procedure was omitted, it would |
| 1703 |
call umask(0) followed by umask(previous_umask) and apply the umask to |
| 1704 |
mode #o777. |
| 1705 |
|
| 1706 |
This was unnecessary and a security issue for multi-threaded |
| 1707 |
applications: during a small window the process' umask was set to zero, |
| 1708 |
so other threads calling mkdir(2) or open(2) could end up creating |
| 1709 |
world-readable/writable/executable directories or files. |
| 1710 |
|
| 1711 |
* New interfaces |
| 1712 |
|
| 1713 |
** mkstemp! takes optional "mode" argument |
| 1714 |
|
| 1715 |
See "File System" in the manual, for more. |
| 1716 |
|
| 1717 |
** New 'scm_to_uintptr_t' and 'scm_from_uintptr_t' C functions |
| 1718 |
|
| 1719 |
* Bug fixes |
| 1720 |
|
| 1721 |
** Fix optimizer bug when compiling fixpoint operator |
| 1722 |
** Fix build error on MinGW |
| 1723 |
** Update 'uname' implementation on MinGW |
| 1724 |
** 'port-encoding' and 'set-port-encoding!' ensure they are passed an |
| 1725 |
open port |
| 1726 |
** (system base target) now recognizes Alpha as a cross-compilation target |
| 1727 |
|
| 1728 |
|
| 1729 |
Changes in 2.0.12 (since 2.0.11): |
| 1730 |
|
| 1731 |
* Notable changes |
| 1732 |
|
| 1733 |
** FFI: Add support for functions that set 'errno' |
| 1734 |
|
| 1735 |
When accessing POSIX functions from a system's libc via Guile's dynamic |
| 1736 |
FFI, you commonly want to access the 'errno' variable to be able to |
| 1737 |
produce useful diagnostic messages. |
| 1738 |
|
| 1739 |
This is now possible using 'pointer->procedure' or |
| 1740 |
'scm_pointer_to_procedure_with_errno'. See "Dynamic FFI" in the manual. |
| 1741 |
|
| 1742 |
** The #!r6rs directive now influences read syntax |
| 1743 |
|
| 1744 |
The #!r6rs directive now changes the per-port reader options to make |
| 1745 |
Guile's reader conform more closely to the R6RS syntax. In particular: |
| 1746 |
|
| 1747 |
- It makes the reader case sensitive. |
| 1748 |
- It disables the recognition of keyword syntax in conflict with the |
| 1749 |
R6RS (and R5RS). |
| 1750 |
- It enables the `square-brackets', `hungry-eol-escapes' and |
| 1751 |
`r6rs-hex-escapes' reader options. |
| 1752 |
|
| 1753 |
** 'read' now accepts "\(" as equivalent to "(" |
| 1754 |
|
| 1755 |
This is indented for use at the beginning of lines in multi-line strings |
| 1756 |
to avoid confusing Emacs' lisp modes. Previously "\(" was an error. |
| 1757 |
|
| 1758 |
** SRFI-14 character data set upgraded to Unicode 8.0.0 |
| 1759 |
|
| 1760 |
** SRFI-19 table of leap seconds updated |
| 1761 |
|
| 1762 |
** 'string-hash', 'read-string', and 'write' have been optimized |
| 1763 |
|
| 1764 |
** GOOPS bug fix for inherited accessor methods |
| 1765 |
|
| 1766 |
In the port of GOOPS to Guile 2.0, we introduced a bug related to |
| 1767 |
accessor methods. The bug resulted in GOOPS assuming that a slot S in |
| 1768 |
an object whose class is C would always be present in instances of all |
| 1769 |
subclasses C, and allocated to the same struct index. This is not the |
| 1770 |
case for multiple inheritance. This behavior has been fixed to be as it |
| 1771 |
was in 1.8. |
| 1772 |
|
| 1773 |
One aspect of this change may cause confusion among users. Previously |
| 1774 |
if you defined a class C: |
| 1775 |
|
| 1776 |
(use-modules (oop goops)) |
| 1777 |
(define-class C () |
| 1778 |
(a #:getter get-a)) |
| 1779 |
|
| 1780 |
And now you define a subclass, intending to provide an #:init-value for |
| 1781 |
the slot A: |
| 1782 |
|
| 1783 |
(define-class D (A) |
| 1784 |
(a #:init-value 42)) |
| 1785 |
|
| 1786 |
Really what you have done is define in D a new slot with the same name, |
| 1787 |
overriding the existing slot. The problem comes in that before fixing |
| 1788 |
this bug (but not in 1.8), the getter 'get-a' would succeed for |
| 1789 |
instances of D, even though 'get-a' should only work for the slot 'a' |
| 1790 |
that is defined on class C, not any other slot that happens to have the |
| 1791 |
same name and be in a class with C as a superclass. |
| 1792 |
|
| 1793 |
It would be possible to "merge" the slot definitions on C and D, but |
| 1794 |
that part of the meta-object protocol (`compute-slots' et al) is not |
| 1795 |
fully implemented. |
| 1796 |
|
| 1797 |
Somewhat relatedly, GOOPS also had a fix around #:init-value on |
| 1798 |
class-allocated slots. GOOPS was re-initializing the value of slots |
| 1799 |
with #:class or #:each-subclass allocation every time instances of that |
| 1800 |
class was allocated. This has been fixed. |
| 1801 |
|
| 1802 |
* New interfaces |
| 1803 |
|
| 1804 |
** New SRFI-28 string formatting implementation |
| 1805 |
|
| 1806 |
See "SRFI-28" in the manual. |
| 1807 |
|
| 1808 |
** New (ice-9 unicode) module |
| 1809 |
|
| 1810 |
See "Characters" in the manual. |
| 1811 |
|
| 1812 |
** Web server |
| 1813 |
|
| 1814 |
The (web server) module now exports 'make-server-impl', 'server-impl?', |
| 1815 |
and related procedures. Likewise, (web server http) exports 'http'. |
| 1816 |
|
| 1817 |
** New procedures: 'string-utf8-length' and 'scm_c_string_utf8_length' |
| 1818 |
|
| 1819 |
See "Bytevectors as Strings" in the manual, for more. |
| 1820 |
|
| 1821 |
** New 'EXIT_SUCCESS' and 'EXIT_FAILURE' Scheme variables |
| 1822 |
|
| 1823 |
See "Processes" in the manual. |
| 1824 |
|
| 1825 |
** New C functions to disable automatic SMOB finalization |
| 1826 |
|
| 1827 |
The new 'scm_set_automatic_finalization_enabled' C function allows you |
| 1828 |
to choose whether automatic object finalization should be enabled (as |
| 1829 |
was the case until now, and still is by default.) This is meant for |
| 1830 |
applications that are not thread-safe nor async-safe; such applications |
| 1831 |
can disable automatic finalization and call the new 'scm_run_finalizers' |
| 1832 |
function when appropriate. |
| 1833 |
|
| 1834 |
See the "Garbage Collecting Smobs" and "Smobs" sections in the manual. |
| 1835 |
|
| 1836 |
** Cross-compilation to ARM |
| 1837 |
|
| 1838 |
More ARM cross-compilation targets are supported: "arm.*eb", |
| 1839 |
"^aarch64.*be", and "aarch64". |
| 1840 |
|
| 1841 |
* New deprecation |
| 1842 |
|
| 1843 |
** The undocumented and unused C function 'scm_string_hash' is now deprecated |
| 1844 |
|
| 1845 |
* Bugs fixed |
| 1846 |
|
| 1847 |
** Compiler |
| 1848 |
*** 'call-with-prompt' does not truncate multiple-value returns |
| 1849 |
(<http://bugs.gnu.org/14347>) |
| 1850 |
*** Use permissions of source file for compiled file |
| 1851 |
(<http://bugs.gnu.org/18477>) |
| 1852 |
*** Fix bug when inlining some functions with optional arguments |
| 1853 |
(<http://bugs.gnu.org/17634>) |
| 1854 |
*** Avoid quadratic expansion time in 'and' and 'or' macros |
| 1855 |
(<http://bugs.gnu.org/17147>) |
| 1856 |
*** Fix expander bug introduced when adding support for tail patterns |
| 1857 |
(<http://lists.gnu.org/archive/html/guile-user/2015-09/msg00017.html>) |
| 1858 |
*** Handle ~p in 'format' warnings (<http://bugs.gnu.org/18299>) |
| 1859 |
*** Fix bug that exposed `list' invocations to CSE |
| 1860 |
(<http://bugs.gnu.org/21899>) |
| 1861 |
*** Reduce eq? and eqv? over constants using equal? |
| 1862 |
(<http://bugs.gnu.org/21855>) |
| 1863 |
*** Skip invalid .go files found in GUILE_LOAD_COMPILED_PATH |
| 1864 |
|
| 1865 |
** Threads |
| 1866 |
*** Fix data races leading to corruption (<http://bugs.gnu.org/22152>) |
| 1867 |
|
| 1868 |
** Memory management |
| 1869 |
*** Fix race between SMOB marking and finalization |
| 1870 |
(<http://bugs.gnu.org/19883>) |
| 1871 |
|
| 1872 |
** Ports |
| 1873 |
*** Fix port position handling on binary input ports |
| 1874 |
(<http://bugs.gnu.org/20302>) |
| 1875 |
*** Bytevector and custom binary ports to use ISO-8859-1 |
| 1876 |
(<http://bugs.gnu.org/20200>) |
| 1877 |
*** Fix buffer overrun with unbuffered custom binary input ports |
| 1878 |
(<http://bugs.gnu.org/19621>) |
| 1879 |
*** Fix memory corruption that arose when using 'get-bytevector-n' |
| 1880 |
(<http://bugs.gnu.org/17466>) |
| 1881 |
|
| 1882 |
** System |
| 1883 |
*** {get,set}sockopt now expect type 'int' for SO_SNDBUF/SO_RCVBUF |
| 1884 |
*** 'system*' now available on MS-Windows |
| 1885 |
*** 'open-pipe' now available on MS-Windows |
| 1886 |
*** Better support for file names containing backslashes on Windows |
| 1887 |
|
| 1888 |
** Web |
| 1889 |
*** 'split-and-decode-uri-path' no longer decodes "+" to space |
| 1890 |
*** HTTP: Support date strings with a leading space for hours |
| 1891 |
(<http://bugs.gnu.org/23421>) |
| 1892 |
*** HTTP: Accept empty reason phrases (<http://bugs.gnu.org/22273>) |
| 1893 |
*** HTTP: 'Location' header can now contain URI references, not just |
| 1894 |
absolute URIs |
| 1895 |
*** HTTP: Improve chunked-mode support (<http://bugs.gnu.org/19939>) |
| 1896 |
*** HTTP: 'open-socket-for-uri' now sets better OS buffering parameters |
| 1897 |
(<http://bugs.gnu.org/15368>) |
| 1898 |
|
| 1899 |
** Miscellaneous |
| 1900 |
*** Fix 'atan' procedure when applied to complex numbers |
| 1901 |
*** Fix Texinfo to HTML conversion for @itemize and @acronym |
| 1902 |
(<http://bugs.gnu.org/21772>) |
| 1903 |
*** 'bytevector-fill!' accepts fill arguments greater than 127 |
| 1904 |
(<http://bugs.gnu.org/19027>) |
| 1905 |
*** 'bytevector-copy' correctly copies SRFI-4 homogeneous vectors |
| 1906 |
(<http://bugs.gnu.org/18866>) |
| 1907 |
*** 'strerror' no longer hangs when passed a non-integer argument |
| 1908 |
(<http://bugs.gnu.org/18065>) |
| 1909 |
*** 'scm_boot_guile' now gracefully handles argc == 0 |
| 1910 |
(<http://bugs.gnu.org/18680>) |
| 1911 |
*** Fix 'SCM_SMOB_OBJECT_LOC' definition (<http://bugs.gnu.org/18495>) |
| 1912 |
*** Fix bug where 'bit-count*' was not using its second argument |
| 1913 |
*** SRFI-1 'length+' raises an error for non-lists and dotted lists |
| 1914 |
(<http://bugs.gnu.org/17296>) |
| 1915 |
*** Add documentation for SXPath (<http://bugs.gnu.org/19478>) |
| 1916 |
|
| 1917 |
|
| 1918 |
Changes in 2.0.11 (since 2.0.10): |
| 1919 |
|
| 1920 |
This release fixes an embarrassing regression introduced in the C |
| 1921 |
interface to SRFI-4 vectors. See |
| 1922 |
<https://lists.gnu.org/archive/html/guile-devel/2014-03/msg00047.html> |
| 1923 |
for details. |
| 1924 |
|
| 1925 |
|
| 1926 |
Changes in 2.0.10 (since 2.0.9): |
| 1927 |
|
| 1928 |
* Notable changes |
| 1929 |
|
| 1930 |
** New GDB extension to support Guile |
| 1931 |
|
| 1932 |
Guile now comes with an extension for GDB 7.8 or later (unreleased at |
| 1933 |
the time of writing) that simplifies debugging of C code that uses |
| 1934 |
Guile. See "GDB Support" in the manual. |
| 1935 |
|
| 1936 |
** Improved integration between R6RS and native Guile exceptions |
| 1937 |
|
| 1938 |
R6RS exception handlers, established using 'with-exception-handler' or |
| 1939 |
'guard', are now able to catch native Guile exceptions, which are |
| 1940 |
automatically converted into appropriate R6RS condition objects. |
| 1941 |
|
| 1942 |
** Support for HTTP proxies |
| 1943 |
|
| 1944 |
Guile's built-in web client now honors the 'http_proxy' environment |
| 1945 |
variable, as well as the new 'current-http-proxy' parameter. See |
| 1946 |
"Web Client" in the manual for details. |
| 1947 |
|
| 1948 |
** Lexical syntax improvements |
| 1949 |
|
| 1950 |
*** Support |...| symbol notation. |
| 1951 |
|
| 1952 |
Guile's core reader and printer now support the R7RS |...| notation |
| 1953 |
for writing symbols with arbitrary characters, as a more portable and |
| 1954 |
attractive alternative to Guile's native #{...}# notation. To enable |
| 1955 |
this notation by default, put one or both of the following in your |
| 1956 |
~/.guile: |
| 1957 |
|
| 1958 |
(read-enable 'r7rs-symbols) |
| 1959 |
(print-enable 'r7rs-symbols) |
| 1960 |
|
| 1961 |
*** Support '#true' and '#false' notation for booleans. |
| 1962 |
|
| 1963 |
The booleans '#t' and '#f' may now be written as '#true' and '#false' |
| 1964 |
for improved readability, per R7RS. |
| 1965 |
|
| 1966 |
*** Recognize '#\escape' character name. |
| 1967 |
|
| 1968 |
The escape character '#\esc' may now be written as '#\escape', per R7RS. |
| 1969 |
|
| 1970 |
*** Accept "\|" in string literals. |
| 1971 |
|
| 1972 |
The pipe character may now be preceded by a backslash, per R7RS. |
| 1973 |
|
| 1974 |
** Custom binary input ports now support 'setvbuf'. |
| 1975 |
|
| 1976 |
Until now, ports returned by 'make-custom-binary-input-port' were always |
| 1977 |
full-buffered. Now, their buffering mode can be changed using 'setvbuf'. |
| 1978 |
|
| 1979 |
** SRFI-4 predicates and length accessors no longer accept arrays. |
| 1980 |
|
| 1981 |
Given that the SRFI-4 accessors don't work for arrays, the fact that the |
| 1982 |
predicates and length accessors returned true for arrays was a bug. |
| 1983 |
|
| 1984 |
** GUILE_PROGS now supports specifying a minimum required version. |
| 1985 |
|
| 1986 |
The 'GUILE_PROGS' autoconf macro in guile.m4 now allows an optional |
| 1987 |
argument to specify a minimum required Guile version. By default, it |
| 1988 |
requires Guile >= 2.0. A micro version can also be specified, e.g.: |
| 1989 |
GUILE_PROGS([2.0.10]) |
| 1990 |
|
| 1991 |
** Error reporting improvements |
| 1992 |
|
| 1993 |
*** Improved run-time error reporting in (ice-9 match). |
| 1994 |
|
| 1995 |
If no pattern matches in a 'match' form, the datum that failed to match |
| 1996 |
is printed along with the location of the failed 'match' invocation. |
| 1997 |
|
| 1998 |
*** Print the faulty object upon invalid-keyword errors. |
| 1999 |
*** Improved error reporting of procedures defined by define-inlinable. |
| 2000 |
*** Improved error reporting for misplaced ellipses in macro definitions. |
| 2001 |
*** Improved error checking in 'define-public' and 'module-add!'. |
| 2002 |
*** Improved error when 'include' form with relative path is not in a file. |
| 2003 |
|
| 2004 |
** Speed improvements |
| 2005 |
|
| 2006 |
*** 'scm_c_read' on ISO-8859-1 (e.g. binary) unbuffered ports is faster. |
| 2007 |
*** New inline asm for VM fixnum multiply, for faster overflow checking. |
| 2008 |
*** New inline asm for VM fixnum operations on ARM and 32-bit x86. |
| 2009 |
*** 'positive?' and 'negative?' are now compiled to VM primitives. |
| 2010 |
*** Numerical comparisons with more than 2 arguments are compiled to VM code. |
| 2011 |
*** Several R6RS bitwise operators have been optimized. |
| 2012 |
|
| 2013 |
** Miscellaneous |
| 2014 |
|
| 2015 |
*** Web: 'content-disposition' headers are now supported. |
| 2016 |
*** Web: 'uri-encode' hexadecimal percent-encoding is now uppercase. |
| 2017 |
*** Size argument to 'make-doubly-weak-hash-table' is now optional. |
| 2018 |
*** Timeout for 'unlock-mutex' and SRFI-18 'mutex-unlock!' may now be #f. |
| 2019 |
|
| 2020 |
** Gnulib update |
| 2021 |
|
| 2022 |
Guile's copy of Gnulib was updated to v0.1-92-g546ff82. The following |
| 2023 |
modules were imported from Gnulib: copysign, fsync, isfinite, link, |
| 2024 |
lstat, mkdir, mkstemp, readlink, rename, rmdir, and unistd. |
| 2025 |
|
| 2026 |
* New interfaces |
| 2027 |
|
| 2028 |
** Cooperative REPL servers |
| 2029 |
|
| 2030 |
This new facility supports REPLs that run at specified times within an |
| 2031 |
existing thread, for example in programs utilizing an event loop or in |
| 2032 |
single-threaded programs. This allows for safe access and mutation of |
| 2033 |
a program's data structures from the REPL without concern for thread |
| 2034 |
synchronization. See "Cooperative REPL Servers" in the manual for |
| 2035 |
details. |
| 2036 |
|
| 2037 |
** SRFI-43 (Vector Library) |
| 2038 |
|
| 2039 |
Guile now includes SRFI-43, a comprehensive library of vector operations |
| 2040 |
analogous to the SRFI-1 list library. See "SRFI-43" in the manual for |
| 2041 |
details. |
| 2042 |
|
| 2043 |
** SRFI-64 (A Scheme API for test suites) |
| 2044 |
|
| 2045 |
Guile now includes SRFI-64, a flexible framework for creating test |
| 2046 |
suites. The reference implementation of SRFI-64 has also been updated |
| 2047 |
to fully support earlier versions of Guile. |
| 2048 |
|
| 2049 |
** SRFI-111 (Boxes) |
| 2050 |
|
| 2051 |
See "SRFI-111" in the manual. |
| 2052 |
|
| 2053 |
** 'define-values' |
| 2054 |
|
| 2055 |
See "Binding multiple return values" in the manual. |
| 2056 |
|
| 2057 |
** Custom ellipsis identifiers using 'with-ellipsis' or SRFI-46. |
| 2058 |
|
| 2059 |
Guile now allows macro definitions to use identifiers other than '...' |
| 2060 |
as the ellipsis. This is convenient when writing macros that generate |
| 2061 |
macro definitions. The desired ellipsis identifier can be given as the |
| 2062 |
first operand to 'syntax-rules', as specified in SRFI-46 and R7RS, or by |
| 2063 |
using the new 'with-ellipsis' special form in procedural macros. With |
| 2064 |
this addition, Guile now fully supports SRFI-46. |
| 2065 |
|
| 2066 |
See "Specifying a Custom Ellipsis Identifier" and "Custom Ellipsis |
| 2067 |
Identifiers for syntax-case Macros" in the manual for details. |
| 2068 |
|
| 2069 |
** R7RS 'syntax-error' |
| 2070 |
|
| 2071 |
Guile now supports 'syntax-error', as specified by R7RS, allowing for |
| 2072 |
improved compile-time error reporting from 'syntax-rules' macros. See |
| 2073 |
"Reporting Syntax Errors in Macros" in the manual for details. |
| 2074 |
|
| 2075 |
** New procedures to convert association lists into hash tables |
| 2076 |
|
| 2077 |
Guile now includes the convenience procedures 'alist->hash-table', |
| 2078 |
'alist->hashq-table', 'alist->hashv-table', and 'alist->hashx-table'. |
| 2079 |
See "Hash Table Reference" in the manual. |
| 2080 |
|
| 2081 |
** New predicates: 'exact-integer?' and 'scm_is_exact_integer' |
| 2082 |
|
| 2083 |
See "Integers" in the manual. |
| 2084 |
|
| 2085 |
** 'weak-vector-length', 'weak-vector-ref', and 'weak-vector-set!' |
| 2086 |
|
| 2087 |
These should now be used to access weak vectors, instead of |
| 2088 |
'vector-length', 'vector-ref', and 'vector-set!'. |
| 2089 |
|
| 2090 |
* Manual updates |
| 2091 |
|
| 2092 |
** Improve docs for 'eval-when'. |
| 2093 |
|
| 2094 |
Each 'eval-when' condition is now explained in detail, including |
| 2095 |
'expand' which was previously undocumented. (expand load eval) is now |
| 2096 |
the recommended set of conditions, instead of (compile load eval). |
| 2097 |
See "Eval When" in the manual, for details. |
| 2098 |
|
| 2099 |
** Update the section on SMOBs and memory management. |
| 2100 |
|
| 2101 |
See "Defining New Types (Smobs)" in the manual. |
| 2102 |
|
| 2103 |
** Fixes |
| 2104 |
|
| 2105 |
*** GOOPS: #:dsupers is the init keyword for the dsupers slot. |
| 2106 |
*** 'unfold-right' takes a tail, not a tail generator. |
| 2107 |
*** Clarify that 'append!' and 'reverse!' might not mutate. |
| 2108 |
*** Fix doc that incorrectly claimed (integer? +inf.0) => #t. |
| 2109 |
(http://bugs.gnu.org/16356) |
| 2110 |
*** Document that we support SRFI-62 (S-expression comments). |
| 2111 |
*** Document that we support SRFI-87 (=> in case clauses). |
| 2112 |
*** Document 'equal?' in the list of R6RS incompatibilities. |
| 2113 |
*** Remove outdated documentation of LTDL_LIBRARY_PATH. |
| 2114 |
*** Fix 'weak-vector?' doc: Weak hash tables are not weak vectors. |
| 2115 |
*** Fix 'my-or' examples to use let-bound variable. |
| 2116 |
(http://bugs.gnu.org/14203) |
| 2117 |
|
| 2118 |
* New deprecations |
| 2119 |
|
| 2120 |
** General 'uniform-vector' interface |
| 2121 |
|
| 2122 |
This interface lacked both generality and specificity. The general |
| 2123 |
replacements are 'array-length', 'array-ref', and friends on the scheme |
| 2124 |
side, and the array handle interface on the C side. On the specific |
| 2125 |
side of things, there are the specific bytevector, SRFI-4, and bitvector |
| 2126 |
interfaces. |
| 2127 |
|
| 2128 |
** Use of the vector interface on arrays |
| 2129 |
** 'vector-length', 'vector-ref', and 'vector-set!' on weak vectors |
| 2130 |
** 'vector-length', 'vector-ref', and 'vector-set!' as primitive-generics |
| 2131 |
|
| 2132 |
Making the vector interface operate only on a single representation will |
| 2133 |
allow future versions of Guile to compile loops involving vectors to |
| 2134 |
more efficient native code. |
| 2135 |
|
| 2136 |
** 'htons', 'htonl', 'ntohs', 'ntohl' |
| 2137 |
|
| 2138 |
These procedures, like their C counterpart, were used to convert numbers |
| 2139 |
to/from network byte order, typically in conjunction with the |
| 2140 |
now-deprecated uniform vector API. |
| 2141 |
|
| 2142 |
This functionality is now covered by the bytevector and binary I/O APIs. |
| 2143 |
See "Interpreting Bytevector Contents as Integers" in the manual. |
| 2144 |
|
| 2145 |
** 'gc-live-object-stats' |
| 2146 |
|
| 2147 |
It hasn't worked in the whole 2.0 series. There is no replacement, |
| 2148 |
unfortunately. |
| 2149 |
|
| 2150 |
** 'scm_c_program_source' |
| 2151 |
|
| 2152 |
This internal VM function was not meant to be public. Use |
| 2153 |
'scm_procedure_source' instead. |
| 2154 |
|
| 2155 |
* Build fixes |
| 2156 |
|
| 2157 |
** Fix build with Clang 3.4. |
| 2158 |
|
| 2159 |
** MinGW build fixes |
| 2160 |
*** Do not add $(EXEEXT) to guild or guile-tools. |
| 2161 |
*** tests: Use double quotes around shell arguments, for Windows. |
| 2162 |
*** tests: Don't rely on $TMPDIR and /tmp on Windows. |
| 2163 |
*** tests: Skip FFI tests that use `qsort' when it's not accessible. |
| 2164 |
*** tests: Remove symlink only when it exists. |
| 2165 |
*** tests: Don't rely on `scm_call_2' being visible. |
| 2166 |
|
| 2167 |
** Fix computation of LIBLOBJS so dependencies work properly. |
| 2168 |
(http://bugs.gnu.org/14193) |
| 2169 |
|
| 2170 |
* Bug fixes |
| 2171 |
|
| 2172 |
** Web: Fix web client with methods other than GET. |
| 2173 |
(http://bugs.gnu.org/15908) |
| 2174 |
** Web: Add Content-Length header for empty bodies. |
| 2175 |
** Web: Accept "UTC" as the zone offset in date headers. |
| 2176 |
(http://bugs.gnu.org/14128) |
| 2177 |
** Web: Don't throw if a response is longer than its Content-Length says. |
| 2178 |
** Web: Write out HTTP Basic auth headers correctly. |
| 2179 |
(http://bugs.gnu.org/14370) |
| 2180 |
** Web: Always print a path component in 'write-request-line'. |
| 2181 |
** Fix 'define-public' from (ice-9 curried-definitions). |
| 2182 |
** psyntax: toplevel variable definitions discard previous syntactic binding. |
| 2183 |
(http://bugs.gnu.org/11988) |
| 2184 |
** Fix thread-unsafe lazy initializations. |
| 2185 |
** Make (ice-9 popen) thread-safe. |
| 2186 |
(http://bugs.gnu.org/15683) |
| 2187 |
** Make guardians thread-safe. |
| 2188 |
** Make regexp_exec thread-safe. |
| 2189 |
(http://bugs.gnu.org/14404) |
| 2190 |
** vm: Gracefully handle stack overflows. |
| 2191 |
(http://bugs.gnu.org/15065) |
| 2192 |
** Fix 'rationalize'. |
| 2193 |
(http://bugs.gnu.org/14905) |
| 2194 |
** Fix inline asm for VM fixnum operations on x32. |
| 2195 |
** Fix 'SCM_SYSCALL' to really swallow EINTR. |
| 2196 |
** Hide EINTR returns from 'accept'. |
| 2197 |
** SRFI-19: Update the table of leap seconds. |
| 2198 |
** Add missing files to the test-suite Makefile. |
| 2199 |
** Make sure 'ftw' allows directory traversal when running as root. |
| 2200 |
** Fix 'hash-for-each' for weak hash tables. |
| 2201 |
** SRFI-18: Export 'current-thread'. |
| 2202 |
(http://bugs.gnu.org/16890) |
| 2203 |
** Fix inlining of tail list to apply. |
| 2204 |
(http://bugs.gnu.org/15533) |
| 2205 |
** Fix bug in remqueue in threads.c when removing last element. |
| 2206 |
** Fix build when '>>' on negative integers is not arithmetic. |
| 2207 |
** Fix 'bitwise-bit-count' for negative arguments. |
| 2208 |
(http://bugs.gnu.org/14864) |
| 2209 |
** Fix VM 'ash' for right shifts by large amounts. |
| 2210 |
(http://bugs.gnu.org/14864) |
| 2211 |
** Fix rounding in scm_i_divide2double for negative arguments. |
| 2212 |
** Avoid lossy conversion from inum to double in numerical comparisons. |
| 2213 |
** Fix numerical comparison of fractions to infinities. |
| 2214 |
** Allow fl+ and fl* to accept zero arguments. |
| 2215 |
(http://bugs.gnu.org/14869) |
| 2216 |
** flonum? returns false for complex number objects. |
| 2217 |
(http://bugs.gnu.org/14866) |
| 2218 |
** flfinite? applied to a NaN returns false. |
| 2219 |
(http://bugs.gnu.org/14868) |
| 2220 |
** Flonum operations always return flonums. |
| 2221 |
(http://bugs.gnu.org/14871) |
| 2222 |
** min and max: NaNs beat infinities, per R6RS errata. |
| 2223 |
(http://bugs.gnu.org/14865) |
| 2224 |
** Fix 'fxbit-count' for negative arguments. |
| 2225 |
** 'gcd' and 'lcm' support inexact integer arguments. |
| 2226 |
(http://bugs.gnu.org/14870) |
| 2227 |
** Fix R6RS 'fixnum-width'. |
| 2228 |
(http://bugs.gnu.org/14879) |
| 2229 |
** tests: Use shell constructs that /bin/sh on Solaris 10 can understand. |
| 2230 |
(http://bugs.gnu.org/14042) |
| 2231 |
** Fix display of symbols containing backslashes. |
| 2232 |
(http://bugs.gnu.org/15033) |
| 2233 |
** Fix truncated-print for uniform vectors. |
| 2234 |
** Define `AF_UNIX' only when Unix-domain sockets are supported. |
| 2235 |
** Decompiler: fix handling of empty 'case-lambda' expressions. |
| 2236 |
** Fix handling of signed zeroes and infinities in 'numerator' and 'denominator'. |
| 2237 |
** dereference-pointer: check for null pointer. |
| 2238 |
** Optimizer: Numerical comparisons are not negatable, for correct NaN handling. |
| 2239 |
** Compiler: Evaluate '-' and '/' in left-to-right order. |
| 2240 |
(for more robust floating-point arithmetic) |
| 2241 |
** snarf.h: Declare static const function name vars as SCM_UNUSED. |
| 2242 |
** chars.c: Remove duplicate 'const' specifiers. |
| 2243 |
** Modify SCM_UNPACK type check to avoid warnings in clang. |
| 2244 |
** Arrange so that 'file-encoding' does not truncate the encoding name. |
| 2245 |
(http://bugs.gnu.org/16463) |
| 2246 |
** Improve error checking in bytevector->uint-list and bytevector->sint-list. |
| 2247 |
(http://bugs.gnu.org/15100) |
| 2248 |
** Fix (ash -1 SCM_I_FIXNUM_BIT-1) to return a fixnum instead of a bignum. |
| 2249 |
** i18n: Fix null pointer dereference when locale info is missing. |
| 2250 |
** Fix 'string-copy!' to work properly with overlapping src/dest. |
| 2251 |
** Fix hashing of vectors to run in bounded time. |
| 2252 |
** 'port-position' works on CBIPs that do not support 'set-port-position!'. |
| 2253 |
** Custom binary input ports sanity-check the return value of 'read!'. |
| 2254 |
** bdw-gc.h: Check SCM_USE_PTHREAD_THREADS using #if not #ifdef. |
| 2255 |
** REPL Server: Don't establish a SIGINT handler. |
| 2256 |
** REPL Server: Redirect warnings to client socket. |
| 2257 |
** REPL Server: Improve robustness of 'stop-server-and-clients!'. |
| 2258 |
** Add srfi-16, srfi-30, srfi-46, srfi-62, srfi-87 to %cond-expand-features. |
| 2259 |
** Fix trap handlers to handle applicable structs. |
| 2260 |
(http://bugs.gnu.org/15691) |
| 2261 |
** Fix optional end argument in `uniform-vector-read!'. |
| 2262 |
(http://bugs.gnu.org/15370) |
| 2263 |
** Fix brainfuck->scheme compiler. |
| 2264 |
** texinfo: Fix newline preservation in @example with lines beginning with @ |
| 2265 |
|
| 2266 |
** C standards conformance improvements |
| 2267 |
|
| 2268 |
Improvements and bug fixes were made to the C part of Guile's run-time |
| 2269 |
support (libguile). |
| 2270 |
|
| 2271 |
*** Don't use the identifier 'noreturn'. |
| 2272 |
(http://bugs.gnu.org/15798) |
| 2273 |
*** Rewrite SCM_I_INUM to avoid unspecified behavior when not using GNU C. |
| 2274 |
*** Improve fallback implemention of SCM_SRS to avoid unspecified behavior. |
| 2275 |
*** SRFI-60: Reimplement 'rotate-bit-field' on inums to be more portable. |
| 2276 |
*** Improve compliance with C standards regarding signed integer shifts. |
| 2277 |
*** Avoid signed overflow in random.c. |
| 2278 |
*** VM: Avoid signed overflows in 'add1' and 'sub1'. |
| 2279 |
*** VM: Avoid overflow in ASM_ADD when the result is most-positive-fixnum. |
| 2280 |
*** read: Avoid signed integer overflow in 'read_decimal_integer'. |
| 2281 |
|
| 2282 |
|
| 2283 |
|
| 2284 |
Changes in 2.0.9 (since 2.0.7): |
| 2285 |
|
| 2286 |
Note: 2.0.8 was a brown paper bag release that was never announced, but |
| 2287 |
some mirrors may have picked it up. Please do not use it. |
| 2288 |
|
| 2289 |
* Notable changes |
| 2290 |
|
| 2291 |
** New keyword arguments for procedures that open files |
| 2292 |
|
| 2293 |
The following procedures that open files now support keyword arguments |
| 2294 |
to request binary I/O or to specify the character encoding for text |
| 2295 |
files: `open-file', `open-input-file', `open-output-file', |
| 2296 |
`call-with-input-file', `call-with-output-file', `with-input-from-file', |
| 2297 |
`with-output-to-file', and `with-error-to-file'. |
| 2298 |
|
| 2299 |
It is also now possible to specify whether Guile should scan files for |
| 2300 |
Emacs-style coding declarations. This scan was done by default in |
| 2301 |
versions 2.0.0 through 2.0.7, but now must be explicitly requested. |
| 2302 |
|
| 2303 |
See "File Ports" in the manual for details. |
| 2304 |
|
| 2305 |
** Rewritten guile.m4 |
| 2306 |
|
| 2307 |
The `guile.m4' autoconf macros have been rewritten to use `guild' and |
| 2308 |
`pkg-config' instead of the deprecated `guile-config' (which itself |
| 2309 |
calls pkg-config). |
| 2310 |
|
| 2311 |
There is also a new macro, `GUILE_PKG', which allows packages to select |
| 2312 |
the version of Guile that they want to compile against. See "Autoconf |
| 2313 |
Macros" in the manual, for more information. |
| 2314 |
|
| 2315 |
** Better Windows support |
| 2316 |
|
| 2317 |
Guile now correctly identifies absolute paths on Windows (MinGW), and |
| 2318 |
creates files on that platform according to its path conventions. See |
| 2319 |
"File System" in the manual, for all details. |
| 2320 |
|
| 2321 |
In addition, the new Gnulib imports provide `select' and `poll' on |
| 2322 |
Windows builds. |
| 2323 |
|
| 2324 |
As an incompatible change, systems that are missing <sys/select.h> were |
| 2325 |
previously provided a public `scm_std_select' C function that defined a |
| 2326 |
version of `select', but unhappily it also provided its own incompatible |
| 2327 |
definitions for FD_SET, FD_ZERO, and other system interfaces. Guile |
| 2328 |
should not be setting these macros in public API, so this interface was |
| 2329 |
removed on those plaforms (basically only MinGW). |
| 2330 |
|
| 2331 |
** Numerics improvements |
| 2332 |
|
| 2333 |
`number->string' now reliably outputs enough digits to produce the same |
| 2334 |
number when read back in. Previously, it mishandled subnormal numbers |
| 2335 |
(printing them as "#.#"), and failed to distinguish between some |
| 2336 |
distinct inexact numbers, e.g. 1.0 and (+ 1.0 (expt 2.0 -52)). These |
| 2337 |
problems had far-reaching implications, since the compiler uses |
| 2338 |
`number->string' to serialize numeric constants into .go files. |
| 2339 |
|
| 2340 |
`sqrt' now produces exact rational results when possible, and handles |
| 2341 |
very large or very small numbers more robustly. |
| 2342 |
|
| 2343 |
A number (ahem) of operations involving exact rationals have been |
| 2344 |
optimized, most notably `integer-expt' and `expt'. |
| 2345 |
|
| 2346 |
`exact->inexact' now performs correct IEEE rounding. |
| 2347 |
|
| 2348 |
** New optimizations |
| 2349 |
|
| 2350 |
There were a number of improvements to the partial evaluator, allowing |
| 2351 |
complete reduction of forms such as: |
| 2352 |
|
| 2353 |
((let ((_ 10)) (lambda () _))) |
| 2354 |
|
| 2355 |
((lambda _ _)) |
| 2356 |
|
| 2357 |
(apply (lambda _ _) 1 2 3 '(4)) |
| 2358 |
|
| 2359 |
(call-with-values (lambda () (values 1 2)) (lambda _ _)) |
| 2360 |
|
| 2361 |
`string-join' now handles huge lists efficiently. |
| 2362 |
|
| 2363 |
`get-bytevector-some' now uses buffered input, which is much faster. |
| 2364 |
|
| 2365 |
Finally, `array-ref', `array-set!' on arrays of rank 1 or 2 is now |
| 2366 |
faster, because it avoids building a rest list. Similarly, the |
| 2367 |
one-argument case of `array-for-each' and `array-map!' has been |
| 2368 |
optimized, and `array-copy!' and `array-fill!' are faster. |
| 2369 |
|
| 2370 |
** `peek-char' no longer consumes EOF |
| 2371 |
|
| 2372 |
As required by the R5RS, if `peek-char' returns EOF, then the next read |
| 2373 |
will also return EOF. Previously `peek-char' would consume the EOF. |
| 2374 |
This makes a difference for terminal devices where it is possible to |
| 2375 |
read past an EOF. |
| 2376 |
|
| 2377 |
** Gnulib update |
| 2378 |
|
| 2379 |
Guile's copy of Gnulib was updated to v0.0-7865-ga828bb2. The following |
| 2380 |
modules were imported from Gnulib: select, times, pipe-posix, fstat, |
| 2381 |
getlogin, poll, and c-strcase. |
| 2382 |
|
| 2383 |
** `include' resolves relative file names relative to including file |
| 2384 |
|
| 2385 |
Given a relative file name, `include' will look for it relative to the |
| 2386 |
directory of the including file. This harmonizes the behavior of |
| 2387 |
`include' with that of `load'. |
| 2388 |
|
| 2389 |
** SLIB compatibility restored |
| 2390 |
|
| 2391 |
Guile 2.0.8 is now compatible with SLIB. You will have to use a |
| 2392 |
development version of SLIB, however, until a new version of SLIB is |
| 2393 |
released. |
| 2394 |
|
| 2395 |
** Better ,trace REPL command |
| 2396 |
|
| 2397 |
Sometimes the ,trace output for nested function calls could overflow the |
| 2398 |
terminal width, which wasn't useful. Now there is a limit to the amount |
| 2399 |
of space the prefix will take. See the documentation for ",trace" for |
| 2400 |
more information. |
| 2401 |
|
| 2402 |
** Better docstring syntax supported for `case-lambda' |
| 2403 |
|
| 2404 |
Docstrings can now be placed immediately after the `case-lambda' or |
| 2405 |
`case-lambda*' keyword. See "Case-lambda" in the manual. |
| 2406 |
|
| 2407 |
** Improved handling of Unicode byte order marks |
| 2408 |
|
| 2409 |
See "BOM Handling" in the manual for details. |
| 2410 |
|
| 2411 |
** Update predefined character sets to Unicode 6.2 |
| 2412 |
|
| 2413 |
** GMP 4.2 or later required |
| 2414 |
|
| 2415 |
Guile used to require GMP at least version 4.1 (released in May 2002), |
| 2416 |
and now requires at least version 4.2 (released in March 2006). |
| 2417 |
|
| 2418 |
* Manual updates |
| 2419 |
|
| 2420 |
** Better SXML documentation |
| 2421 |
|
| 2422 |
The documentation for SXML modules was much improved, though there is |
| 2423 |
still far to go. See "SXML" in manual. |
| 2424 |
|
| 2425 |
** Style updates |
| 2426 |
|
| 2427 |
Use of "iff" was replaced with standard English. Keyword arguments are |
| 2428 |
now documented consistently, along with their default values. |
| 2429 |
|
| 2430 |
** An end to the generated-documentation experiment |
| 2431 |
|
| 2432 |
When Guile 2.0 imported some modules from Guile-Lib, they came with a |
| 2433 |
system that generated documentation from docstrings and module |
| 2434 |
commentaries. This produced terrible documentation. We finally bit the |
| 2435 |
bullet and incorporated these modules into the main text, and will be |
| 2436 |
improving them manually over time, as is the case with SXML. Help is |
| 2437 |
appreciated. |
| 2438 |
|
| 2439 |
** New documentation |
| 2440 |
|
| 2441 |
There is now documentation for `scm_array_type', and `scm_array_ref', as |
| 2442 |
well as for the new `array-length' / 'scm_c_array_length' / |
| 2443 |
`scm_array_length' functions. `array-in-bounds?' has better |
| 2444 |
documentation as well. The `program-arguments-alist' and |
| 2445 |
`program-lambda-list' functions are now documented, as well as `and=>', |
| 2446 |
`exit', and `quit'. The (system repl server) module is now documented |
| 2447 |
(see REPL Servers). Finally, the GOOPS class hierarchy diagram has been |
| 2448 |
regenerated for the web and print output formats. |
| 2449 |
|
| 2450 |
* New deprecations |
| 2451 |
|
| 2452 |
** Deprecate generalized vector interface |
| 2453 |
|
| 2454 |
The generalized vector interface, introduced in 1.8.0, is simply a |
| 2455 |
redundant, verbose interface to arrays of rank 1. `array-ref' and |
| 2456 |
similar functions are entirely sufficient. Thus, |
| 2457 |
`scm_generalized_vector_p', `scm_generalized_vector_length', |
| 2458 |
`scm_generalized_vector_ref', `scm_generalized_vector_set_x', and |
| 2459 |
`scm_generalized_vector_to_list' are now deprecated. |
| 2460 |
|
| 2461 |
** Deprecate SCM_CHAR_CODE_LIMIT and char-code-limit |
| 2462 |
|
| 2463 |
These constants were defined to 256, which is not the highest codepoint |
| 2464 |
supported by Guile. Given that they were useless and incorrect, they |
| 2465 |
have been deprecated. |
| 2466 |
|
| 2467 |
** Deprecate `http-get*' |
| 2468 |
|
| 2469 |
The new `#:streaming?' argument to `http-get' subsumes the functionality |
| 2470 |
of `http-get*' (introduced in 2.0.7). Also, the `#:extra-headers' |
| 2471 |
argument is deprecated in favor of `#:headers'. |
| 2472 |
|
| 2473 |
** Deprecate (ice-9 mapping) |
| 2474 |
|
| 2475 |
This module, present in Guile since 1996 but never used or documented, |
| 2476 |
has never worked in Guile 2.0. It has now been deprecated and will be |
| 2477 |
removed in Guile 2.2. |
| 2478 |
|
| 2479 |
** Deprecate undocumented array-related C functions |
| 2480 |
|
| 2481 |
These are `scm_array_fill_int', `scm_ra_eqp', `scm_ra_lessp', |
| 2482 |
`scm_ra_leqp', `scm_ra_grp', `scm_ra_greqp', `scm_ra_sum', |
| 2483 |
`scm_ra_product', `scm_ra_difference', `scm_ra_divide', and |
| 2484 |
`scm_array_identity'. |
| 2485 |
|
| 2486 |
* New interfaces |
| 2487 |
|
| 2488 |
** SRFI-41 Streams |
| 2489 |
|
| 2490 |
See "SRFI-41" in the manual. |
| 2491 |
|
| 2492 |
** SRFI-45 exports `promise?' |
| 2493 |
|
| 2494 |
SRFI-45 now exports a `promise?' procedure that works with its promises. |
| 2495 |
Also, its promises now print more nicely. |
| 2496 |
|
| 2497 |
** New HTTP client procedures |
| 2498 |
|
| 2499 |
See "Web Client" for documentation on the new `http-head', `http-post', |
| 2500 |
`http-put', `http-delete', `http-trace', and `http-options' procedures, |
| 2501 |
and also for more options to `http-get'. |
| 2502 |
|
| 2503 |
** Much more capable `xml->sxml' |
| 2504 |
|
| 2505 |
See "Reading and Writing XML" for information on how the `xml->sxml' |
| 2506 |
parser deals with namespaces, processed entities, doctypes, and literal |
| 2507 |
strings. Incidentally, `current-ssax-error-port' is now a parameter |
| 2508 |
object. |
| 2509 |
|
| 2510 |
** New procedures for converting strings to and from bytevectors |
| 2511 |
|
| 2512 |
See "Representing Strings as Bytes" for documention on the new `(ice-9 |
| 2513 |
iconv)' module and its `bytevector->string' and `string->bytevector' |
| 2514 |
procedures. |
| 2515 |
|
| 2516 |
** Escape continuations with `call/ec' and `let/ec' |
| 2517 |
|
| 2518 |
See "Prompt Primitives". |
| 2519 |
|
| 2520 |
** New procedures to read all characters from a port |
| 2521 |
|
| 2522 |
See "Line/Delimited" in the manual for documentation on `read-string' |
| 2523 |
and `read-string!'. |
| 2524 |
|
| 2525 |
** New procedure `sendfile' |
| 2526 |
|
| 2527 |
See "File System". |
| 2528 |
|
| 2529 |
** New procedure `unget-bytevector' |
| 2530 |
|
| 2531 |
See "R6RS Binary Input". |
| 2532 |
|
| 2533 |
** New C helper: `scm_c_bind_keyword_arguments' |
| 2534 |
|
| 2535 |
See "Keyword Procedures". |
| 2536 |
|
| 2537 |
** New command-line arguments: `--language' and `-C' |
| 2538 |
|
| 2539 |
See "Command-line Options" in the manual. |
| 2540 |
|
| 2541 |
** New environment variables: `GUILE_STACK_SIZE', `GUILE_INSTALL_LOCALE' |
| 2542 |
|
| 2543 |
See "Environment Variables". |
| 2544 |
|
| 2545 |
** New procedures for dealing with file names |
| 2546 |
|
| 2547 |
See "File System" for documentation on `system-file-name-convention', |
| 2548 |
`file-name-separator?', `absolute-file-name?', and |
| 2549 |
`file-name-separator-string'. |
| 2550 |
|
| 2551 |
** `array-length', an array's first dimension |
| 2552 |
|
| 2553 |
See "Array Procedures". |
| 2554 |
|
| 2555 |
** `hash-count', for hash tables |
| 2556 |
|
| 2557 |
See "Hash Tables". |
| 2558 |
|
| 2559 |
** `round-ash', a bit-shifting operator that rounds on right-shift |
| 2560 |
|
| 2561 |
See "Bitwise Operations". |
| 2562 |
|
| 2563 |
** New foreign types: `ssize_t', `ptrdiff_t' |
| 2564 |
|
| 2565 |
See "Foreign Types". |
| 2566 |
|
| 2567 |
** New C helpers: `scm_from_ptrdiff_t', `scm_to_ptrdiff_t' |
| 2568 |
|
| 2569 |
See "Integers". |
| 2570 |
|
| 2571 |
** Socket option `SO_REUSEPORT' now available from Scheme |
| 2572 |
|
| 2573 |
If supported on the platform, `SO_REUSEPORT' is now available from |
| 2574 |
Scheme as well. See "Network Sockets and Communication". |
| 2575 |
|
| 2576 |
** `current-language' in default environment |
| 2577 |
|
| 2578 |
Previously defined only in `(system base language)', `current-language' |
| 2579 |
is now defined in the default environment, and is used to determine the |
| 2580 |
language for the REPL, and for `compile-and-load'. |
| 2581 |
|
| 2582 |
** New procedure: `fluid->parameter' |
| 2583 |
|
| 2584 |
See "Parameters", for information on how to convert a fluid to a |
| 2585 |
parameter. |
| 2586 |
|
| 2587 |
** New `print' REPL option |
| 2588 |
|
| 2589 |
See "REPL Commands" in the manual for information on the new |
| 2590 |
user-customizable REPL printer. |
| 2591 |
|
| 2592 |
** New variable: %site-ccache-dir |
| 2593 |
|
| 2594 |
The "Installing Site Packages" and "Build Config" manual sections now |
| 2595 |
refer to this variable to describe where users should install their |
| 2596 |
`.go' files. |
| 2597 |
|
| 2598 |
* Build fixes |
| 2599 |
|
| 2600 |
** Fix compilation against libgc 7.3. |
| 2601 |
** Fix cross-compilation of `c-tokenize.o'. |
| 2602 |
** Fix warning when compiling against glibc 2.17. |
| 2603 |
** Fix documentation build against Texinfo 5.0. |
| 2604 |
** Fix building Guile from a directory with non-ASCII characters. |
| 2605 |
** Fix native MinGW build. |
| 2606 |
** Fix --disable-posix build. |
| 2607 |
** Fix MinGW builds with networking, POSIX, and thread support. |
| 2608 |
|
| 2609 |
* Bug fixes |
| 2610 |
|
| 2611 |
** Fix inexact number printer. |
| 2612 |
(http://bugs.gnu.org/13757) |
| 2613 |
** Fix infinite loop when parsing optional-argument short options (SRFI-37). |
| 2614 |
(http://bugs.gnu.org/13176) |
| 2615 |
** web: Support non-GMT date headers in the HTTP client. |
| 2616 |
(http://bugs.gnu.org/13544) |
| 2617 |
** web: support IP-literal (IPv6 address) in Host header. |
| 2618 |
** Avoid stack overflows with `par-map' and nested futures in general. |
| 2619 |
(http://bugs.gnu.org/13188) |
| 2620 |
** Peek-char no longer consumes EOF. |
| 2621 |
(http://bugs.gnu.org/12216) |
| 2622 |
** Avoid swallowing multiple EOFs in R6RS binary-input procedures. |
| 2623 |
** A fork when multiple threads are running will now print a warning. |
| 2624 |
** Allow for spurious wakeups from pthread_cond_wait. |
| 2625 |
(http://bugs.gnu.org/10641) |
| 2626 |
** Warn and ignore module autoload failures. |
| 2627 |
(http://bugs.gnu.org/12202) |
| 2628 |
** Use chmod portably in (system base compile). |
| 2629 |
(http://bugs.gnu.org/10474) |
| 2630 |
** Fix response-body-port for HTTP responses without content-length. |
| 2631 |
(http://bugs.gnu.org/13857) |
| 2632 |
** Allow case-lambda expressions with no clauses. |
| 2633 |
(http://bugs.gnu.org/9776) |
| 2634 |
** Improve standards conformance of string->number. |
| 2635 |
(http://bugs.gnu.org/11887) |
| 2636 |
** Support calls and tail-calls with more than 255 formals. |
| 2637 |
** ,option evaluates its right-hand-side. |
| 2638 |
(http://bugs.gnu.org/13076) |
| 2639 |
** Structs with tail arrays are not simple. |
| 2640 |
(http://bugs.gnu.org/12808) |
| 2641 |
** Make `SCM_LONG_BIT' usable in preprocessor conditionals. |
| 2642 |
(http://bugs.gnu.org/13848) |
| 2643 |
** Fix thread-unsafe lazy initializations. |
| 2644 |
** Allow SMOB mark procedures to be called from parallel markers. |
| 2645 |
(http://bugs.gnu.org/13611) |
| 2646 |
** Fix later-bindings-win logic in with-fluids. |
| 2647 |
(http://bugs.gnu.org/13843) |
| 2648 |
** Fix duplicate removal of with-fluids. |
| 2649 |
(http://bugs.gnu.org/13838) |
| 2650 |
** Support calling foreign functions of 10 arguments or more. |
| 2651 |
(http://bugs.gnu.org/13809) |
| 2652 |
** Let reverse! accept arbitrary types as second argument. |
| 2653 |
(http://bugs.gnu.org/13835) |
| 2654 |
** Recognize the `x86_64.*-gnux32' triplet. |
| 2655 |
** Check whether a triplet's OS part specifies an ABI. |
| 2656 |
** Recognize mips64* as having 32-bit pointers by default. |
| 2657 |
** Use portable sed constructs. |
| 2658 |
(http://bugs.gnu.org/14042) |
| 2659 |
** Remove language/glil/decompile-assembly.scm. |
| 2660 |
(http://bugs.gnu.org/10622) |
| 2661 |
** Use O_BINARY in `copy-file', `load-objcode', `mkstemp'. |
| 2662 |
** Use byte-oriented functions in `get-bytevector*'. |
| 2663 |
** Fix abort when iconv swallows BOM from UTF-16 or UTF-32 stream. |
| 2664 |
** Fix compilation of functions with more than 255 local variables. |
| 2665 |
** Fix `getgroups' for when zero supplementary group IDs exist. |
| 2666 |
** Allow (define-macro name (lambda ...)). |
| 2667 |
** Various fixes to the (texinfo) modules. |
| 2668 |
** guild: Gracefully handle failures to install the locale. |
| 2669 |
** Fix format string warnings for ~!, ~|, ~/, ~q, ~Q, and ~^. |
| 2670 |
(http://bugs.gnu.org/13485) |
| 2671 |
** Fix source annotation bug in psyntax 'expand-body'. |
| 2672 |
** Ecmascript: Fix conversion to boolean for non-numbers. |
| 2673 |
** Use case-insensitive comparisons for encoding names. |
| 2674 |
** Add missing cond-expand feature identifiers. |
| 2675 |
** A failure to find a module's file does not prevent future loading. |
| 2676 |
** Many (oop goops save) fixes. |
| 2677 |
** `http-get': don't shutdown write end of socket. |
| 2678 |
(http://bugs.gnu.org/13095) |
| 2679 |
** Avoid signed integer overflow in scm_product. |
| 2680 |
** http: read-response-body always returns bytevector or #f, never EOF. |
| 2681 |
** web: Correctly detect "No route to host" conditions. |
| 2682 |
** `system*': failure to execvp no longer leaks dangling processes. |
| 2683 |
(http://bugs.gnu.org/13166) |
| 2684 |
** More sensible case-lambda* dispatch. |
| 2685 |
(http://bugs.gnu.org/12929) |
| 2686 |
** Do not defer expansion of internal define-syntax forms. |
| 2687 |
(http://bugs.gnu.org/13509) |
| 2688 |
|
| 2689 |
|
| 2690 |
|
| 2691 |
Changes in 2.0.7 (since 2.0.6): |
| 2692 |
|
| 2693 |
* Notable changes |
| 2694 |
|
| 2695 |
** SRFI-105 curly infix expressions are supported |
| 2696 |
|
| 2697 |
Curly infix expressions as described at |
| 2698 |
http://srfi.schemers.org/srfi-105/srfi-105.html are now supported by |
| 2699 |
Guile's reader. This allows users to write things like {a * {b + c}} |
| 2700 |
instead of (* a (+ b c)). SRFI-105 support is enabled by using the |
| 2701 |
`#!curly-infix' directive in source code, or the `curly-infix' reader |
| 2702 |
option. See the manual for details. |
| 2703 |
|
| 2704 |
** Reader options may now be per-port |
| 2705 |
|
| 2706 |
Historically, `read-options' and related procedures would manipulate |
| 2707 |
global options, affecting the `read' procedure for all threads, and all |
| 2708 |
current uses of `read'. |
| 2709 |
|
| 2710 |
Guile can now associate `read' options with specific ports, allowing |
| 2711 |
different ports to use different options. For instance, the |
| 2712 |
`#!fold-case' and `#!no-fold-case' reader directives have been |
| 2713 |
implemented, and their effect is to modify the current read options of |
| 2714 |
the current port only; similarly for `#!curly-infix'. Thus, it is |
| 2715 |
possible, for instance, to have one port reading case-sensitive code, |
| 2716 |
while another port reads case-insensitive code. |
| 2717 |
|
| 2718 |
** Futures may now be nested |
| 2719 |
|
| 2720 |
Futures may now be nested: a future can itself spawn and then `touch' |
| 2721 |
other futures. In addition, any thread that touches a future that has |
| 2722 |
not completed now processes other futures while waiting for the touched |
| 2723 |
future to completed. This allows all threads to be kept busy, and was |
| 2724 |
made possible by the use of delimited continuations (see the manual for |
| 2725 |
details.) |
| 2726 |
|
| 2727 |
Consequently, `par-map' and `par-for-each' have been rewritten and can |
| 2728 |
now use all cores. |
| 2729 |
|
| 2730 |
** `GUILE_LOAD_PATH' et al can now add directories to the end of the path |
| 2731 |
|
| 2732 |
`GUILE_LOAD_PATH' and `GUILE_LOAD_COMPILED_PATH' can now be used to add |
| 2733 |
directories to both ends of the load path. If the special path |
| 2734 |
component `...' (ellipsis) is present in these environment variables, |
| 2735 |
then the default path is put in place of the ellipsis, otherwise the |
| 2736 |
default path is placed at the end. See "Environment Variables" in the |
| 2737 |
manual for details. |
| 2738 |
|
| 2739 |
** `load-in-vicinity' search for `.go' files in `%load-compiled-path' |
| 2740 |
|
| 2741 |
Previously, `load-in-vicinity' would look for compiled files in the |
| 2742 |
auto-compilation cache, but not in `%load-compiled-path'. This is now |
| 2743 |
fixed. This affects `load', and the `-l' command-line flag. See |
| 2744 |
<http://bugs.gnu.org/12519> for details. |
| 2745 |
|
| 2746 |
** Extension search order fixed, and LD_LIBRARY_PATH preserved |
| 2747 |
|
| 2748 |
Up to 2.0.6, Guile would modify the `LD_LIBRARY_PATH' environment |
| 2749 |
variable (or whichever is relevant for the host OS) to insert its own |
| 2750 |
default extension directories in the search path (using GNU libltdl |
| 2751 |
facilities was not possible here.) This approach was problematic in two |
| 2752 |
ways. |
| 2753 |
|
| 2754 |
First, the `LD_LIBRARY_PATH' modification would be visible to |
| 2755 |
sub-processes, and would also affect future calls to `dlopen', which |
| 2756 |
could lead to subtle bugs in the application or sub-processes. Second, |
| 2757 |
when the installation prefix is /usr, the `LD_LIBRARY_PATH' modification |
| 2758 |
would typically end up inserting /usr/lib before /usr/local/lib in the |
| 2759 |
search path, which is often the opposite of system-wide settings such as |
| 2760 |
`ld.so.conf'. |
| 2761 |
|
| 2762 |
Both issues have now been fixed. |
| 2763 |
|
| 2764 |
** `make-vtable-vtable' is now deprecated |
| 2765 |
|
| 2766 |
Programs should instead use `make-vtable' and `<standard-vtable>'. |
| 2767 |
|
| 2768 |
** The `-Wduplicate-case-datum' and `-Wbad-case-datum' are enabled |
| 2769 |
|
| 2770 |
These recently introduced warnings have been documented and are now |
| 2771 |
enabled by default when auto-compiling. |
| 2772 |
|
| 2773 |
** Optimize calls to `equal?' or `eqv?' with a constant argument |
| 2774 |
|
| 2775 |
The compiler simplifies calls to `equal?' or `eqv?' with a constant |
| 2776 |
argument to use `eq?' instead, when applicable. |
| 2777 |
|
| 2778 |
* Manual updates |
| 2779 |
|
| 2780 |
** SRFI-9 records now documented under "Compound Data Types" |
| 2781 |
|
| 2782 |
The documentation of SRFI-9 record types has been moved in the "Compound |
| 2783 |
Data Types", next to Guile's other record APIs. A new section |
| 2784 |
introduces the various record APIs, and describes the trade-offs they |
| 2785 |
make. These changes were made in an attempt to better guide users |
| 2786 |
through the maze of records API, and to recommend SRFI-9 as the main |
| 2787 |
API. |
| 2788 |
|
| 2789 |
The documentation of Guile's raw `struct' API has also been improved. |
| 2790 |
|
| 2791 |
** (ice-9 and-let-star) and (ice-9 curried-definitions) now documented |
| 2792 |
|
| 2793 |
These modules were missing from the manual. |
| 2794 |
|
| 2795 |
* New interfaces |
| 2796 |
|
| 2797 |
** New "functional record setters" as a GNU extension of SRFI-9 |
| 2798 |
|
| 2799 |
The (srfi srfi-9 gnu) module now provides three new macros to deal with |
| 2800 |
"updates" of immutable records: `define-immutable-record-type', |
| 2801 |
`set-field', and `set-fields'. |
| 2802 |
|
| 2803 |
The first one allows record type "functional setters" to be defined; |
| 2804 |
such setters keep the record unchanged, and instead return a new record |
| 2805 |
with only one different field. The remaining macros provide the same |
| 2806 |
functionality, and also optimize updates of multiple or nested fields. |
| 2807 |
See the manual for details. |
| 2808 |
|
| 2809 |
** web: New `http-get*', `response-body-port', and `text-content-type?' |
| 2810 |
procedures |
| 2811 |
|
| 2812 |
These procedures return a port from which to read the response's body. |
| 2813 |
Unlike `http-get' and `read-response-body', they allow the body to be |
| 2814 |
processed incrementally instead of being stored entirely in memory. |
| 2815 |
|
| 2816 |
The `text-content-type?' predicate allows users to determine whether the |
| 2817 |
content type of a response is textual. |
| 2818 |
|
| 2819 |
See the manual for details. |
| 2820 |
|
| 2821 |
** `string-split' accepts character sets and predicates |
| 2822 |
|
| 2823 |
The `string-split' procedure can now be given a SRFI-14 character set or |
| 2824 |
a predicate, instead of just a character. |
| 2825 |
|
| 2826 |
** R6RS SRFI support |
| 2827 |
|
| 2828 |
Previously, in R6RS modules, Guile incorrectly ignored components of |
| 2829 |
SRFI module names after the SRFI number, making it impossible to specify |
| 2830 |
sub-libraries. This release corrects this, bringing us into accordance |
| 2831 |
with SRFI 97. |
| 2832 |
|
| 2833 |
** `define-public' is no a longer curried definition by default |
| 2834 |
|
| 2835 |
The (ice-9 curried-definitions) should be used for such uses. See the |
| 2836 |
manual for details. |
| 2837 |
|
| 2838 |
* Build fixes |
| 2839 |
|
| 2840 |
** Remove reference to `scm_init_popen' when `fork' is unavailable |
| 2841 |
|
| 2842 |
This fixes a MinGW build issue (http://bugs.gnu.org/12477). |
| 2843 |
|
| 2844 |
** Fix race between installing `guild' and the `guile-tools' symlink |
| 2845 |
|
| 2846 |
* Bug fixes |
| 2847 |
|
| 2848 |
** Procedures returned by `eval' now have docstrings |
| 2849 |
(http://bugs.gnu.org/12173) |
| 2850 |
** web client: correctly handle uri-query, etc. in relative URI headers |
| 2851 |
(http://bugs.gnu.org/12827) |
| 2852 |
** Fix docs for R6RS `hashtable-copy' |
| 2853 |
** R6RS `string-for-each' now accepts multiple string arguments |
| 2854 |
** Fix out-of-range error in the compiler's CSE pass |
| 2855 |
(http://bugs.gnu.org/12883) |
| 2856 |
** Add missing R6RS `open-file-input/output-port' procedure |
| 2857 |
** Futures: Avoid creating the worker pool more than once |
| 2858 |
** Fix invalid assertion about mutex ownership in threads.c |
| 2859 |
(http://bugs.gnu.org/12719) |
| 2860 |
** Have `SCM_NUM2FLOAT' and `SCM_NUM2DOUBLE' use `scm_to_double' |
| 2861 |
** The `scandir' procedure now uses `lstat' instead of `stat' |
| 2862 |
** Fix `generalized-vector->list' indexing bug with shared arrays |
| 2863 |
(http://bugs.gnu.org/12465) |
| 2864 |
** web: Change `http-get' to try all the addresses for the given URI |
| 2865 |
** Implement `hash' for structs |
| 2866 |
(http://lists.gnu.org/archive/html/guile-devel/2012-10/msg00031.html) |
| 2867 |
** `read' now adds source properties for data types beyond pairs |
| 2868 |
** Improve error reporting in `append!' |
| 2869 |
** In fold-matches, set regexp/notbol unless matching string start |
| 2870 |
** Don't stat(2) and access(2) the .go location before using it |
| 2871 |
** SRFI-19: use zero padding for hours in ISO 8601 format, not blanks |
| 2872 |
** web: Fix uri-encoding for strings with no unreserved chars, and octets 0-15 |
| 2873 |
** More robust texinfo alias handling |
| 2874 |
** Optimize `format' and `simple-format' |
| 2875 |
(http://bugs.gnu.org/12033) |
| 2876 |
** Angle of -0.0 is pi, not zero |
| 2877 |
|
| 2878 |
|
| 2879 |
Changes in 2.0.6 (since 2.0.5): |
| 2880 |
|
| 2881 |
* Notable changes |
| 2882 |
|
| 2883 |
** New optimization pass: common subexpression elimination (CSE) |
| 2884 |
|
| 2885 |
Guile's optimizer will now run a CSE pass after partial evaluation. |
| 2886 |
This pass propagates static information about branches taken, bound |
| 2887 |
lexicals, and effects from an expression's dominators. It can replace |
| 2888 |
common subexpressions with their boolean values (potentially enabling |
| 2889 |
dead code elimination), equivalent bound lexicals, or it can elide them |
| 2890 |
entirely, depending on the context in which they are executed. This |
| 2891 |
pass is especially useful in removing duplicate type checks, such as |
| 2892 |
those produced by SRFI-9 record accessors. |
| 2893 |
|
| 2894 |
** Improvements to the partial evaluator |
| 2895 |
|
| 2896 |
Peval can now hoist tests that are common to both branches of a |
| 2897 |
conditional into the test. This can help with long chains of |
| 2898 |
conditionals, such as those generated by the `match' macro. Peval can |
| 2899 |
now do simple beta-reductions of procedures with rest arguments. It |
| 2900 |
also avoids residualizing degenerate lexical aliases, even when full |
| 2901 |
inlining is not possible. Finally, peval now uses the effects analysis |
| 2902 |
introduced for the CSE pass. More precise effects analysis allows peval |
| 2903 |
to move more code. |
| 2904 |
|
| 2905 |
** Run finalizers asynchronously in asyncs |
| 2906 |
|
| 2907 |
Finalizers are now run asynchronously, via an async. See Asyncs in the |
| 2908 |
manual. This allows Guile and user code to safely allocate memory while |
| 2909 |
holding a mutex. |
| 2910 |
|
| 2911 |
** Update SRFI-14 character sets to Unicode 6.1 |
| 2912 |
|
| 2913 |
Note that this update causes the Latin-1 characters `§' and `¶' to be |
| 2914 |
reclassified as punctuation. They were previously considered to be part |
| 2915 |
of `char-set:symbol'. |
| 2916 |
|
| 2917 |
** Better source information for datums |
| 2918 |
|
| 2919 |
When the `positions' reader option is on, as it is by default, Guile's |
| 2920 |
reader will record source information for more kinds of datums. |
| 2921 |
|
| 2922 |
** Improved error and warning messages |
| 2923 |
|
| 2924 |
`syntax-violation' errors now prefer `subform' for source info, with |
| 2925 |
`form' as fallback. Syntactic errors in `cond' and `case' now produce |
| 2926 |
better errors. `case' can now warn on duplicate datums, or datums that |
| 2927 |
cannot be usefully compared with `eqv?'. `-Warity-mismatch' now handles |
| 2928 |
applicable structs. `-Wformat' is more robust in the presence of |
| 2929 |
`gettext'. Finally, various exceptions thrown by the Web modules now |
| 2930 |
define appropriate exception printers. |
| 2931 |
|
| 2932 |
** A few important bug fixes in the HTTP modules. |
| 2933 |
|
| 2934 |
Guile's web server framework now checks if an application returns a body |
| 2935 |
where it is not permitted, for example in response to a HEAD request, |
| 2936 |
and warn or truncate the response as appropriate. Bad requests now |
| 2937 |
cause a 400 Bad Request response to be printed before closing the port. |
| 2938 |
Finally, some date-printing and URL-parsing bugs were fixed. |
| 2939 |
|
| 2940 |
** Pretty-print improvements |
| 2941 |
|
| 2942 |
When Guile needs to pretty-print Tree-IL, it will try to reconstruct |
| 2943 |
`cond', `or`, and other derived syntax forms from the primitive tree-IL |
| 2944 |
forms. It also uses the original names instead of the fresh unique |
| 2945 |
names, when it is unambiguous to do so. This can be seen in the output |
| 2946 |
of REPL commands like `,optimize'. |
| 2947 |
|
| 2948 |
Also, the `pretty-print' procedure has a new keyword argument, |
| 2949 |
`#:max-expr-width'. |
| 2950 |
|
| 2951 |
** Fix memory leak involving applicable SMOBs |
| 2952 |
|
| 2953 |
At some point in the 1.9.x series, Guile began leaking any applicable |
| 2954 |
SMOB that was actually applied. (There was a weak-key map from SMOB to |
| 2955 |
trampoline functions, where the value had a strong reference on the |
| 2956 |
key.) This has been fixed. There was much rejoicing! |
| 2957 |
|
| 2958 |
** Support for HTTP/1.1 chunked transfer coding |
| 2959 |
|
| 2960 |
See "Transfer Codings" in the manual, for more. |
| 2961 |
|
| 2962 |
** Micro-optimizations |
| 2963 |
|
| 2964 |
A pile of micro-optimizations: the `string-trim' function when called |
| 2965 |
with `char-set:whitespace'; the `(web http)' parsers; SMOB application; |
| 2966 |
conversion of raw UTF-8 and UTF-32 data to and from SCM strings; vlists |
| 2967 |
and vhashes; `read' when processing string literals. |
| 2968 |
|
| 2969 |
** Incompatible change to `scandir' |
| 2970 |
|
| 2971 |
As was the original intention, `scandir' now runs the `select?' |
| 2972 |
procedure on all items, including subdirectories and the `.' and `..' |
| 2973 |
entries. It receives the basename of the file in question instead of |
| 2974 |
the full name. We apologize for this incompatible change to this |
| 2975 |
function introduced in the 2.0.4 release. |
| 2976 |
|
| 2977 |
* Manual updates |
| 2978 |
|
| 2979 |
The manual has been made much more consistent in its naming conventions |
| 2980 |
with regards to formal parameters of functions. Thanks to Bake Timmons. |
| 2981 |
|
| 2982 |
* New interfaces |
| 2983 |
|
| 2984 |
** New C function: `scm_to_pointer' |
| 2985 |
** New C inline functions: `scm_new_smob', `scm_new_double_smob' |
| 2986 |
** (ice-9 format): Add ~h specifier for localized number output. |
| 2987 |
** (web response): New procedure: `response-must-not-include-body?' |
| 2988 |
** New predicate: 'supports-source-properties?' |
| 2989 |
** New C helpers: `scm_c_values', `scm_c_nvalues' |
| 2990 |
** Newly public inline C function: `scm_unget_byte' |
| 2991 |
** (language tree-il): New functions: `tree-il=?', `tree-il-hash' |
| 2992 |
** New fluid: `%default-port-conversion-strategy' |
| 2993 |
** New syntax: `=>' within `case' |
| 2994 |
** (web http): `make-chunked-input-port', `make-chunked-output-port' |
| 2995 |
** (web http): `declare-opaque-header!' |
| 2996 |
|
| 2997 |
Search the manual for these identifiers, for more information. |
| 2998 |
|
| 2999 |
* New deprecations |
| 3000 |
|
| 3001 |
** `close-io-port' deprecated |
| 3002 |
|
| 3003 |
Use `close-port'. |
| 3004 |
|
| 3005 |
** `scm_sym2var' deprecated |
| 3006 |
|
| 3007 |
In most cases, replace with `scm_lookup' or `scm_module_variable'. Use |
| 3008 |
`scm_define' or `scm_module_ensure_local_variable' if the second |
| 3009 |
argument is nonzero. See "Accessing Modules from C" in the manual, for |
| 3010 |
full details. |
| 3011 |
|
| 3012 |
** Lookup closures deprecated |
| 3013 |
|
| 3014 |
These were never documented. See "Module System Reflection" in the |
| 3015 |
manual for replacements. |
| 3016 |
|
| 3017 |
* Build fixes |
| 3018 |
|
| 3019 |
** Fix compilation against uninstalled Guile on non-GNU platforms. |
| 3020 |
** Fix `SCM_I_ERROR' definition for MinGW without networking. |
| 3021 |
** Fix compilation with the Sun C compiler. |
| 3022 |
** Fix check for `clock_gettime' on OpenBSD and some other systems. |
| 3023 |
** Fix build with --enable-debug-malloc. |
| 3024 |
** Honor $(program_transform_name) for the `guile-tools' symlink. |
| 3025 |
** Fix cross-compilation of GOOPS-using code. |
| 3026 |
|
| 3027 |
* Bug fixes |
| 3028 |
|
| 3029 |
** Fix use of unitialized stat buffer in search-path of absolute paths. |
| 3030 |
** Avoid calling `freelocale' with a NULL argument. |
| 3031 |
** Work around erroneous tr_TR locale in Darwin 8 in tests. |
| 3032 |
** Fix `getaddrinfo' test for Darwin 8. |
| 3033 |
** Use Gnulib's `regex' module for better regex portability. |
| 3034 |
** `source-properties' and friends work on any object |
| 3035 |
** Rewrite open-process in C, for robustness related to threads and fork |
| 3036 |
** Fix <TAG>vector-length when applied to other uniform vector types |
| 3037 |
** Fix escape-only prompt optimization (was disabled previously) |
| 3038 |
** Fix a segfault when /dev/urandom is not accessible |
| 3039 |
** Fix flush on soft ports, so that it actually runs. |
| 3040 |
** Better compatibility of SRFI-9 records with core records |
| 3041 |
** Fix and clarify documentation of `sorted?'. |
| 3042 |
** Fix IEEE-754 endianness conversion in bytevectors. |
| 3043 |
** Correct thunk check in the `wind' instruction. |
| 3044 |
** Add @acronym support to texinfo modules |
| 3045 |
** Fix docbook->texi for <ulink> without URL |
| 3046 |
** Fix `setvbuf' to leave the line/column number unchanged. |
| 3047 |
** Add missing public declaration for `scm_take_from_input_buffers'. |
| 3048 |
** Fix relative file name canonicalization with empty %LOAD-PATH entries. |
| 3049 |
** Import newer (ice-9 match) from Chibi-Scheme. |
| 3050 |
** Fix unbound variables and unbound values in ECMAScript runtime. |
| 3051 |
** Make SRFI-6 string ports Unicode-capable. |
| 3052 |
|
| 3053 |
|
| 3054 |
Changes in 2.0.5 (since 2.0.4): |
| 3055 |
|
| 3056 |
This release fixes the binary interface information (SONAME) of |
| 3057 |
libguile, which was incorrect in 2.0.4. It does not contain other |
| 3058 |
changes. |
| 3059 |
|
| 3060 |
|
| 3061 |
Changes in 2.0.4 (since 2.0.3): |
| 3062 |
|
| 3063 |
* Notable changes |
| 3064 |
|
| 3065 |
** Better debuggability for interpreted procedures. |
| 3066 |
|
| 3067 |
Guile 2.0 came with a great debugging experience for compiled |
| 3068 |
procedures, but the story for interpreted procedures was terrible. Now, |
| 3069 |
at least, interpreted procedures have names, and the `arity' procedure |
| 3070 |
property is always correct (or, as correct as it can be, in the presence |
| 3071 |
of `case-lambda'). |
| 3072 |
|
| 3073 |
** Support for cross-compilation. |
| 3074 |
|
| 3075 |
One can now use a native Guile to cross-compile `.go' files for a |
| 3076 |
different architecture. See the documentation for `--target' in the |
| 3077 |
"Compilation" section of the manual, for information on how to use the |
| 3078 |
cross-compiler. See the "Cross building Guile" section of the README, |
| 3079 |
for more on how to cross-compile Guile itself. |
| 3080 |
|
| 3081 |
** The return of `local-eval'. |
| 3082 |
|
| 3083 |
Back by popular demand, `the-environment' and `local-eval' allow the |
| 3084 |
user to capture a lexical environment, and then evaluate arbitrary |
| 3085 |
expressions in that context. There is also a new `local-compile' |
| 3086 |
command. See "Local Evaluation" in the manual, for more. Special |
| 3087 |
thanks to Mark Weaver for an initial implementation of this feature. |
| 3088 |
|
| 3089 |
** Fluids can now have default values. |
| 3090 |
|
| 3091 |
Fluids are used for dynamic and thread-local binding. They have always |
| 3092 |
inherited their values from the context or thread that created them. |
| 3093 |
However, there was a case in which a new thread would enter Guile, and |
| 3094 |
the default values of all the fluids would be `#f' for that thread. |
| 3095 |
|
| 3096 |
This has now been fixed so that `make-fluid' has an optional default |
| 3097 |
value for fluids in unrelated dynamic roots, which defaults to `#f'. |
| 3098 |
|
| 3099 |
** Garbage collector tuning. |
| 3100 |
|
| 3101 |
The garbage collector has now been tuned to run more often under some |
| 3102 |
circumstances. |
| 3103 |
|
| 3104 |
*** Unmanaged allocation |
| 3105 |
|
| 3106 |
The new `scm_gc_register_allocation' function will notify the collector |
| 3107 |
of unmanaged allocation. This will cause the collector to run sooner. |
| 3108 |
Guile's `scm_malloc', `scm_calloc', and `scm_realloc' unmanaged |
| 3109 |
allocators eventually call this function. This leads to better |
| 3110 |
performance under steady-state unmanaged allocation. |
| 3111 |
|
| 3112 |
*** Transient allocation |
| 3113 |
|
| 3114 |
When the collector runs, it will try to record the total memory |
| 3115 |
footprint of a process, if the platform supports this information. If |
| 3116 |
the memory footprint is growing, the collector will run more frequently. |
| 3117 |
This reduces the increase of the resident size of a process in response |
| 3118 |
to a transient increase in allocation. |
| 3119 |
|
| 3120 |
*** Management of threads, bignums |
| 3121 |
|
| 3122 |
Creating a thread will allocate a fair amount of memory. Guile now does |
| 3123 |
some GC work (using `GC_collect_a_little') when allocating a thread. |
| 3124 |
This leads to a better memory footprint when creating many short-lived |
| 3125 |
threads. |
| 3126 |
|
| 3127 |
Similarly, bignums can occupy a lot of memory. Guile now offers hooks |
| 3128 |
to enable custom GMP allocators that end up calling |
| 3129 |
`scm_gc_register_allocation'. These allocators are enabled by default |
| 3130 |
when running Guile from the command-line. To enable them in libraries, |
| 3131 |
set the `scm_install_gmp_memory_functions' variable to a nonzero value |
| 3132 |
before loading Guile. |
| 3133 |
|
| 3134 |
** SRFI-39 parameters are available by default. |
| 3135 |
|
| 3136 |
Guile now includes support for parameters, as defined by SRFI-39, in the |
| 3137 |
default environment. See "Parameters" in the manual, for more |
| 3138 |
information. `current-input-port', `current-output-port', and |
| 3139 |
`current-error-port' are now parameters. |
| 3140 |
|
| 3141 |
** Add `current-warning-port'. |
| 3142 |
|
| 3143 |
Guile now outputs warnings on a separate port, `current-warning-port', |
| 3144 |
initialized to the value that `current-error-port' has on startup. |
| 3145 |
|
| 3146 |
** Syntax parameters. |
| 3147 |
|
| 3148 |
Following Racket's lead, Guile now supports syntax parameters. See |
| 3149 |
"Syntax parameters" in the manual, for more. |
| 3150 |
|
| 3151 |
Also see Barzilay, Culpepper, and Flatt's 2011 SFP workshop paper, |
| 3152 |
"Keeping it Clean with syntax-parameterize". |
| 3153 |
|
| 3154 |
** Parse command-line arguments from the locale encoding. |
| 3155 |
|
| 3156 |
Guile now attempts to parse command-line arguments using the user's |
| 3157 |
locale. However for backwards compatibility with other 2.0.x releases, |
| 3158 |
it does so without actually calling `setlocale'. Please report any bugs |
| 3159 |
in this facility to [email protected]. |
| 3160 |
|
| 3161 |
** One-armed conditionals: `when' and `unless' |
| 3162 |
|
| 3163 |
Guile finally has `when' and `unless' in the default environment. Use |
| 3164 |
them whenever you would use an `if' with only one branch. See |
| 3165 |
"Conditionals" in the manual, for more. |
| 3166 |
|
| 3167 |
** `current-filename', `add-to-load-path' |
| 3168 |
|
| 3169 |
There is a new form, `(current-filename)', which expands out to the |
| 3170 |
source file in which it occurs. Combined with the new |
| 3171 |
`add-to-load-path', this allows simple scripts to easily add nearby |
| 3172 |
directories to the load path. See "Load Paths" in the manual, for more. |
| 3173 |
|
| 3174 |
** `random-state-from-platform' |
| 3175 |
|
| 3176 |
This procedure initializes a random seed using good random sources |
| 3177 |
available on your platform, such as /dev/urandom. See "Random Number |
| 3178 |
Generation" in the manual, for more. |
| 3179 |
|
| 3180 |
** Warn about unsupported `simple-format' options. |
| 3181 |
|
| 3182 |
The `-Wformat' compilation option now reports unsupported format options |
| 3183 |
passed to `simple-format'. |
| 3184 |
|
| 3185 |
** Manual updates |
| 3186 |
|
| 3187 |
Besides the sections already mentioned, the following manual sections |
| 3188 |
are new in this release: "Modules and the File System", "Module System |
| 3189 |
Reflection", "Syntax Transformer Helpers", and "Local Inclusion". |
| 3190 |
|
| 3191 |
* New interfaces |
| 3192 |
|
| 3193 |
** (ice-9 session): `apropos-hook' |
| 3194 |
** New print option: `escape-newlines', defaults to #t. |
| 3195 |
** (ice-9 ftw): `file-system-fold', `file-system-tree', `scandir' |
| 3196 |
** `scm_c_value_ref': access to multiple returned values from C |
| 3197 |
** scm_call (a varargs version), scm_call_7, scm_call_8, scm_call_9 |
| 3198 |
** Some new syntax helpers in (system syntax) |
| 3199 |
|
| 3200 |
Search the manual for these identifiers and modules, for more. |
| 3201 |
|
| 3202 |
* Build fixes |
| 3203 |
|
| 3204 |
** FreeBSD build fixes. |
| 3205 |
** OpenBSD compilation fixes. |
| 3206 |
** Solaris 2.10 test suite fixes. |
| 3207 |
** IA64 compilation fix. |
| 3208 |
** MinGW build fixes. |
| 3209 |
** Work around instruction reordering on SPARC and HPPA in the VM. |
| 3210 |
** Gnulib updates: added `dirfd', `setenv' modules. |
| 3211 |
|
| 3212 |
* Bug fixes |
| 3213 |
|
| 3214 |
** Add a deprecated alias for $expt. |
| 3215 |
** Add an exception printer for `getaddrinfo-error'. |
| 3216 |
** Add deprecated shim for `scm_display_error' with stack as first argument. |
| 3217 |
** Add warnings for unsupported `simple-format' options. |
| 3218 |
** Allow overlapping regions to be passed to `bytevector-copy!'. |
| 3219 |
** Better function prologue disassembly |
| 3220 |
** Compiler: fix miscompilation of (values foo ...) in some contexts. |
| 3221 |
** Compiler: fix serialization of #nil-terminated lists. |
| 3222 |
** Compiler: allow values bound in non-tail let expressions to be collected. |
| 3223 |
** Deprecate SCM_ASRTGO. |
| 3224 |
** Document invalidity of (begin) as expression; add back-compat shim. |
| 3225 |
** Don't leak file descriptors when mmaping objcode. |
| 3226 |
** Empty substrings no longer reference the original stringbuf. |
| 3227 |
** FFI: Fix `set-pointer-finalizer!' to leave the type cell unchanged. |
| 3228 |
** FFI: Hold a weak reference to the CIF made by `procedure->pointer'. |
| 3229 |
** FFI: Hold a weak reference to the procedure passed to `procedure->pointer'. |
| 3230 |
** FFI: Properly unpack small integer return values in closure call. |
| 3231 |
** Fix R6RS `fold-left' so the accumulator is the first argument. |
| 3232 |
** Fix bit-set*! bug from 2005. |
| 3233 |
** Fix bug in `make-repl' when `lang' is actually a <language>. |
| 3234 |
** Fix bugs related to mutation, the null string, and shared substrings. |
| 3235 |
** Fix <dynwind> serialization. |
| 3236 |
** Fix erroneous check in `set-procedure-properties!'. |
| 3237 |
** Fix generalized-vector-{ref,set!} for slices. |
| 3238 |
** Fix error messages involving definition forms. |
| 3239 |
** Fix primitive-eval to return #<unspecified> for definitions. |
| 3240 |
** HTTP: Extend handling of "Cache-Control" header. |
| 3241 |
** HTTP: Fix qstring writing of cache-extension values |
| 3242 |
** HTTP: Fix validators for various list-style headers. |
| 3243 |
** HTTP: Permit non-date values for Expires header. |
| 3244 |
** HTTP: `write-request-line' writes absolute paths, not absolute URIs. |
| 3245 |
** Hack the port-column of current-output-port after printing a prompt. |
| 3246 |
** Make sure `regexp-quote' tests use Unicode-capable string ports. |
| 3247 |
** Peval: Fix bugs in the new optimizer. |
| 3248 |
** Statistically unique marks and labels, for robust hygiene across sessions. |
| 3249 |
** Web: Allow URIs with empty authorities, like "file:///etc/hosts". |
| 3250 |
** `,language' at REPL sets the current-language fluid. |
| 3251 |
** `primitive-load' returns the value(s) of the last expression. |
| 3252 |
** `scm_from_stringn' always returns unique strings. |
| 3253 |
** `scm_i_substring_copy' tries to narrow the substring. |
| 3254 |
** i18n: Fix gc_malloc/free mismatch on non-GNU systems. |
| 3255 |
|
| 3256 |
|
| 3257 |
Changes in 2.0.3 (since 2.0.2): |
| 3258 |
|
| 3259 |
* Speed improvements |
| 3260 |
|
| 3261 |
** Guile has a new optimizer, `peval'. |
| 3262 |
|
| 3263 |
`Peval' is a partial evaluator that performs constant folding, dead code |
| 3264 |
elimination, copy propagation, and inlining. By default it runs on |
| 3265 |
every piece of code that Guile compiles, to fold computations that can |
| 3266 |
happen at compile-time, so they don't have to happen at runtime. |
| 3267 |
|
| 3268 |
If we did our job right, the only impact you would see would be your |
| 3269 |
programs getting faster. But if you notice slowdowns or bloated code, |
| 3270 |
please send a mail to [email protected] with details. |
| 3271 |
|
| 3272 |
Thanks to William R. Cook, Oscar Waddell, and Kent Dybvig for inspiring |
| 3273 |
peval and its implementation. |
| 3274 |
|
| 3275 |
You can see what peval does on a given piece of code by running the new |
| 3276 |
`,optimize' REPL meta-command, and comparing it to the output of |
| 3277 |
`,expand'. See "Compile Commands" in the manual, for more. |
| 3278 |
|
| 3279 |
** Fewer calls to `stat'. |
| 3280 |
|
| 3281 |
Guile now stats only the .go file and the .scm file when loading a fresh |
| 3282 |
compiled file. |
| 3283 |
|
| 3284 |
* Notable changes |
| 3285 |
|
| 3286 |
** New module: `(web client)', a simple synchronous web client. |
| 3287 |
|
| 3288 |
See "Web Client" in the manual, for more. |
| 3289 |
|
| 3290 |
** Users can now install compiled `.go' files. |
| 3291 |
|
| 3292 |
See "Installing Site Packages" in the manual. |
| 3293 |
|
| 3294 |
** Remove Front-Cover and Back-Cover text from the manual. |
| 3295 |
|
| 3296 |
The manual is still under the GNU Free Documentation License, but no |
| 3297 |
longer has any invariant sections. |
| 3298 |
|
| 3299 |
** More helpful `guild help'. |
| 3300 |
|
| 3301 |
`guild' is Guile's multi-tool, for use in shell scripting. Now it has a |
| 3302 |
nicer interface for querying the set of existing commands, and getting |
| 3303 |
help on those commands. Try it out and see! |
| 3304 |
|
| 3305 |
** New macro: `define-syntax-rule' |
| 3306 |
|
| 3307 |
`define-syntax-rule' is a shorthand to make a `syntax-rules' macro with |
| 3308 |
one clause. See "Syntax Rules" in the manual, for more. |
| 3309 |
|
| 3310 |
** The `,time' REPL meta-command now has more precision. |
| 3311 |
|
| 3312 |
The output of this command now has microsecond precision, instead of |
| 3313 |
10-millisecond precision. |
| 3314 |
|
| 3315 |
** `(ice-9 match)' can now match records. |
| 3316 |
|
| 3317 |
See "Pattern Matching" in the manual, for more on matching records. |
| 3318 |
|
| 3319 |
** New module: `(language tree-il debug)'. |
| 3320 |
|
| 3321 |
This module provides a tree-il verifier. This is useful for people that |
| 3322 |
generate tree-il, usually as part of a language compiler. |
| 3323 |
|
| 3324 |
** New functions: `scm_is_exact', `scm_is_inexact'. |
| 3325 |
|
| 3326 |
These provide a nice C interface for Scheme's `exact?' and `inexact?', |
| 3327 |
respectively. |
| 3328 |
|
| 3329 |
* Bugs fixed |
| 3330 |
|
| 3331 |
See the git log (or the ChangeLog) for more details on these bugs. |
| 3332 |
|
| 3333 |
** Fix order of importing modules and resolving duplicates handlers. |
| 3334 |
** Fix a number of bugs involving extended (merged) generics. |
| 3335 |
** Fix invocation of merge-generics duplicate handler. |
| 3336 |
** Fix write beyond array end in arrays.c. |
| 3337 |
** Fix read beyond end of hashtable size array in hashtab.c. |
| 3338 |
** (web http): Locale-independent parsing and serialization of dates. |
| 3339 |
** Ensure presence of Host header in HTTP/1.1 requests. |
| 3340 |
** Fix take-right and drop-right for improper lists. |
| 3341 |
** Fix leak in get_current_locale(). |
| 3342 |
** Fix recursive define-inlinable expansions. |
| 3343 |
** Check that srfi-1 procedure arguments are procedures. |
| 3344 |
** Fix r6rs `map' for multiple returns. |
| 3345 |
** Fix scm_tmpfile leak on POSIX platforms. |
| 3346 |
** Fix a couple of leaks (objcode->bytecode, make-boot-program). |
| 3347 |
** Fix guile-lib back-compatibility for module-stexi-documentation. |
| 3348 |
** Fix --listen option to allow other ports. |
| 3349 |
** Fix scm_to_latin1_stringn for substrings. |
| 3350 |
** Fix compilation of untyped arrays of rank not 1. |
| 3351 |
** Fix unparse-tree-il of <dynset>. |
| 3352 |
** Fix reading of #||||#. |
| 3353 |
** Fix segfault in GOOPS when class fields are redefined. |
| 3354 |
** Prefer poll(2) over select(2) to allow file descriptors above FD_SETSIZE. |
| 3355 |
|
| 3356 |
|
| 3357 |
Changes in 2.0.2 (since 2.0.1): |
| 3358 |
|
| 3359 |
* Notable changes |
| 3360 |
|
| 3361 |
** `guile-tools' renamed to `guild' |
| 3362 |
|
| 3363 |
The new name is shorter. Its intended future use is for a CPAN-like |
| 3364 |
system for Guile wizards and journeyfolk to band together to share code; |
| 3365 |
hence the name. `guile-tools' is provided as a backward-compatible |
| 3366 |
symbolic link. See "Using Guile Tools" in the manual, for more. |
| 3367 |
|
| 3368 |
** New control operators: `shift' and `reset' |
| 3369 |
|
| 3370 |
See "Shift and Reset" in the manual, for more information. |
| 3371 |
|
| 3372 |
** `while' as an expression |
| 3373 |
|
| 3374 |
Previously the return value of `while' was unspecified. Now its |
| 3375 |
values are specified both in the case of normal termination, and via |
| 3376 |
termination by invoking `break', possibly with arguments. See "while |
| 3377 |
do" in the manual for more. |
| 3378 |
|
| 3379 |
** Disallow access to handles of weak hash tables |
| 3380 |
|
| 3381 |
`hash-get-handle' and `hash-create-handle!' are no longer permitted to |
| 3382 |
be called on weak hash tables, because the fields in a weak handle could |
| 3383 |
be nulled out by the garbage collector at any time, but yet they are |
| 3384 |
otherwise indistinguishable from pairs. Use `hash-ref' and `hash-set!' |
| 3385 |
instead. |
| 3386 |
|
| 3387 |
** More precision for `get-internal-run-time', `get-internal-real-time' |
| 3388 |
|
| 3389 |
On 64-bit systems which support POSIX clocks, Guile's internal timing |
| 3390 |
procedures offer nanosecond resolution instead of the 10-millisecond |
| 3391 |
resolution previously available. 32-bit systems now use 1-millisecond |
| 3392 |
timers. |
| 3393 |
|
| 3394 |
** Guile now measures time spent in GC |
| 3395 |
|
| 3396 |
`gc-stats' now returns a meaningful value for `gc-time-taken'. |
| 3397 |
|
| 3398 |
** Add `gcprof' |
| 3399 |
|
| 3400 |
The statprof profiler now exports a `gcprof' procedure, driven by the |
| 3401 |
`after-gc-hook', to see which parts of your program are causing GC. Let |
| 3402 |
us know if you find it useful. |
| 3403 |
|
| 3404 |
** `map', `for-each' and some others now implemented in Scheme |
| 3405 |
|
| 3406 |
We would not mention this in NEWS, as it is not a user-visible change, |
| 3407 |
if it were not for one thing: `map' and `for-each' are no longer |
| 3408 |
primitive generics. Instead they are normal bindings, which can be |
| 3409 |
wrapped by normal generics. This fixes some modularity issues between |
| 3410 |
core `map', SRFI-1 `map', and GOOPS. |
| 3411 |
|
| 3412 |
Also it's pretty cool that we can do this without a performance impact. |
| 3413 |
|
| 3414 |
** Add `scm_peek_byte_or_eof'. |
| 3415 |
|
| 3416 |
This helper is like `scm_peek_char_or_eof', but for bytes instead of |
| 3417 |
full characters. |
| 3418 |
|
| 3419 |
** Implement #:stop-at-first-non-option option for getopt-long |
| 3420 |
|
| 3421 |
See "getopt-long Reference" in the manual, for more information. |
| 3422 |
|
| 3423 |
** Improve R6RS conformance for conditions in the I/O libraries |
| 3424 |
|
| 3425 |
The `(rnrs io simple)' module now raises the correct R6RS conditions in |
| 3426 |
error cases. `(rnrs io ports)' is also more correct now, though it is |
| 3427 |
still a work in progress. |
| 3428 |
|
| 3429 |
** All deprecated routines emit warnings |
| 3430 |
|
| 3431 |
A few deprecated routines were lacking deprecation warnings. This has |
| 3432 |
been fixed now. |
| 3433 |
|
| 3434 |
* Speed improvements |
| 3435 |
|
| 3436 |
** Constants in compiled code now share state better |
| 3437 |
|
| 3438 |
Constants with shared state, like `("foo")' and `"foo"', now share state |
| 3439 |
as much as possible, in the entire compilation unit. This cuts compiled |
| 3440 |
`.go' file sizes in half, generally, and speeds startup. |
| 3441 |
|
| 3442 |
** VLists: optimize `vlist-fold-right', and add `vhash-fold-right' |
| 3443 |
|
| 3444 |
These procedures are now twice as fast as they were. |
| 3445 |
|
| 3446 |
** UTF-8 ports to bypass `iconv' entirely |
| 3447 |
|
| 3448 |
This reduces memory usage in a very common case. |
| 3449 |
|
| 3450 |
** Compiler speedups |
| 3451 |
|
| 3452 |
The compiler is now about 40% faster. (Note that this is only the case |
| 3453 |
once the compiler is itself compiled, so the build still takes as long |
| 3454 |
as it did before.) |
| 3455 |
|
| 3456 |
** VM speed tuning |
| 3457 |
|
| 3458 |
Some assertions that were mostly useful for sanity-checks on the |
| 3459 |
bytecode compiler are now off for both "regular" and "debug" engines. |
| 3460 |
This together with a fix to cache a TLS access and some other tweaks |
| 3461 |
improve the VM's performance by about 20%. |
| 3462 |
|
| 3463 |
** SRFI-1 list-set optimizations |
| 3464 |
|
| 3465 |
lset-adjoin and lset-union now have fast paths for eq? sets. |
| 3466 |
|
| 3467 |
** `memq', `memv' optimizations |
| 3468 |
|
| 3469 |
These procedures are now at least twice as fast than in 2.0.1. |
| 3470 |
|
| 3471 |
* Deprecations |
| 3472 |
|
| 3473 |
** Deprecate scm_whash API |
| 3474 |
|
| 3475 |
`scm_whash_get_handle', `SCM_WHASHFOUNDP', `SCM_WHASHREF', |
| 3476 |
`SCM_WHASHSET', `scm_whash_create_handle', `scm_whash_lookup', and |
| 3477 |
`scm_whash_insert' are now deprecated. Use the normal hash table API |
| 3478 |
instead. |
| 3479 |
|
| 3480 |
** Deprecate scm_struct_table |
| 3481 |
|
| 3482 |
`SCM_STRUCT_TABLE_NAME', `SCM_SET_STRUCT_TABLE_NAME', |
| 3483 |
`SCM_STRUCT_TABLE_CLASS', `SCM_SET_STRUCT_TABLE_CLASS', |
| 3484 |
`scm_struct_table', and `scm_struct_create_handle' are now deprecated. |
| 3485 |
These routines formed part of the internals of the map between structs |
| 3486 |
and classes. |
| 3487 |
|
| 3488 |
** Deprecate scm_internal_dynamic_wind |
| 3489 |
|
| 3490 |
The `scm_t_inner' type and `scm_internal_dynamic_wind' are deprecated, |
| 3491 |
as the `scm_dynwind' API is better, and this API encourages users to |
| 3492 |
stuff SCM values into pointers. |
| 3493 |
|
| 3494 |
** Deprecate scm_immutable_cell, scm_immutable_double_cell |
| 3495 |
|
| 3496 |
These routines are deprecated, as the GC_STUBBORN API doesn't do |
| 3497 |
anything any more. |
| 3498 |
|
| 3499 |
* Manual updates |
| 3500 |
|
| 3501 |
Andreas Rottman kindly transcribed the missing parts of the `(rnrs io |
| 3502 |
ports)' documentation from the R6RS documentation. Thanks Andreas! |
| 3503 |
|
| 3504 |
* Bugs fixed |
| 3505 |
|
| 3506 |
** Fix double-loading of script in -ds case |
| 3507 |
** -x error message fix |
| 3508 |
** iconveh-related cross-compilation fixes |
| 3509 |
** Fix small integer return value packing on big endian machines. |
| 3510 |
** Fix hash-set! in weak-value table from non-immediate to immediate |
| 3511 |
** Fix call-with-input-file & relatives for multiple values |
| 3512 |
** Fix `hash' for inf and nan |
| 3513 |
** Fix libguile internal type errors caught by typing-strictness==2 |
| 3514 |
** Fix compile error in MinGW fstat socket detection |
| 3515 |
** Fix generation of auto-compiled file names on MinGW |
| 3516 |
** Fix multithreaded access to internal hash tables |
| 3517 |
** Emit a 1-based line number in error messages |
| 3518 |
** Fix define-module ordering |
| 3519 |
** Fix several POSIX functions to use the locale encoding |
| 3520 |
** Add type and range checks to the complex generalized vector accessors |
| 3521 |
** Fix unaligned accesses for bytevectors of complex numbers |
| 3522 |
** Fix '(a #{.} b) |
| 3523 |
** Fix erroneous VM stack overflow for canceled threads |
| 3524 |
|
| 3525 |
|
| 3526 |
Changes in 2.0.1 (since 2.0.0): |
| 3527 |
|
| 3528 |
* Notable changes |
| 3529 |
|
| 3530 |
** guile.m4 supports linking with rpath |
| 3531 |
|
| 3532 |
The GUILE_FLAGS macro now sets GUILE_LIBS and GUILE_LTLIBS, which |
| 3533 |
include appropriate directives to the linker to include libguile-2.0.so |
| 3534 |
in the runtime library lookup path. |
| 3535 |
|
| 3536 |
** `begin' expands macros in its body before other expressions |
| 3537 |
|
| 3538 |
This enables support for programs like the following: |
| 3539 |
|
| 3540 |
(begin |
| 3541 |
(define even? |
| 3542 |
(lambda (x) |
| 3543 |
(or (= x 0) (odd? (- x 1))))) |
| 3544 |
(define-syntax odd? |
| 3545 |
(syntax-rules () |
| 3546 |
((odd? x) (not (even? x))))) |
| 3547 |
(even? 10)) |
| 3548 |
|
| 3549 |
** REPL reader usability enhancements |
| 3550 |
|
| 3551 |
The REPL now flushes input after a read error, which should prevent one |
| 3552 |
error from causing other errors. The REPL also now interprets comments |
| 3553 |
as whitespace. |
| 3554 |
|
| 3555 |
** REPL output has configurable width |
| 3556 |
|
| 3557 |
The REPL now defaults to output with the current terminal's width, in |
| 3558 |
columns. See "Debug Commands" in the manual for more information on |
| 3559 |
the ,width command. |
| 3560 |
|
| 3561 |
** Better C access to the module system |
| 3562 |
|
| 3563 |
Guile now has convenient C accessors to look up variables or values in |
| 3564 |
modules and their public interfaces. See `scm_c_public_ref' and friends |
| 3565 |
in "Accessing Modules from C" in the manual. |
| 3566 |
|
| 3567 |
** Added `scm_call_5', `scm_call_6' |
| 3568 |
|
| 3569 |
See "Fly Evaluation" in the manual. |
| 3570 |
|
| 3571 |
** Added `scm_from_latin1_keyword', `scm_from_utf8_keyword' |
| 3572 |
|
| 3573 |
See "Keyword Procedures" in the manual, for more. Note that |
| 3574 |
`scm_from_locale_keyword' should not be used when the name is a C string |
| 3575 |
constant. |
| 3576 |
|
| 3577 |
** R6RS unicode and string I/O work |
| 3578 |
|
| 3579 |
Added efficient implementations of `get-string-n' and `get-string-n!' |
| 3580 |
for binary ports. Exported `current-input-port', `current-output-port' |
| 3581 |
and `current-error-port' from `(rnrs io ports)', and enhanced support |
| 3582 |
for transcoders. |
| 3583 |
|
| 3584 |
** Added `pointer->scm', `scm->pointer' to `(system foreign)' |
| 3585 |
|
| 3586 |
These procedure are useful if one needs to pass and receive SCM values |
| 3587 |
to and from foreign functions. See "Foreign Variables" in the manual, |
| 3588 |
for more. |
| 3589 |
|
| 3590 |
** Added `heap-allocated-since-gc' to `(gc-stats)' |
| 3591 |
|
| 3592 |
Also fixed the long-standing bug in the REPL `,stat' command. |
| 3593 |
|
| 3594 |
** Add `on-error' REPL option |
| 3595 |
|
| 3596 |
This option controls what happens when an error occurs at the REPL, and |
| 3597 |
defaults to `debug', indicating that Guile should enter the debugger. |
| 3598 |
Other values include `report', which will simply print a backtrace |
| 3599 |
without entering the debugger. See "System Commands" in the manual. |
| 3600 |
|
| 3601 |
** Enforce immutability of string literals |
| 3602 |
|
| 3603 |
Attempting to mutate a string literal now causes a runtime error. |
| 3604 |
|
| 3605 |
** Fix pthread redirection |
| 3606 |
|
| 3607 |
Guile 2.0.0 shipped with headers that, if configured with pthread |
| 3608 |
support, would re-define `pthread_create', `pthread_join', and other API |
| 3609 |
to redirect to the BDW-GC wrappers, `GC_pthread_create', etc. This was |
| 3610 |
unintended, and not necessary: because threads must enter Guile with |
| 3611 |
`scm_with_guile', Guile can handle thread registration itself, without |
| 3612 |
needing to make the GC aware of all threads. This oversight has been |
| 3613 |
fixed. |
| 3614 |
|
| 3615 |
** `with-continuation-barrier' now unwinds on `quit' |
| 3616 |
|
| 3617 |
A throw to `quit' in a continuation barrier will cause Guile to exit. |
| 3618 |
Before, it would do so before unwinding to the barrier, which would |
| 3619 |
prevent cleanup handlers from running. This has been fixed so that it |
| 3620 |
exits only after unwinding. |
| 3621 |
|
| 3622 |
** `string->pointer' and `pointer->string' have optional encoding arg |
| 3623 |
|
| 3624 |
This allows users of the FFI to more easily deal in strings with |
| 3625 |
particular (non-locale) encodings, like "utf-8". See "Void Pointers and |
| 3626 |
Byte Access" in the manual, for more. |
| 3627 |
|
| 3628 |
** R6RS fixnum arithmetic optimizations |
| 3629 |
|
| 3630 |
R6RS fixnum operations are are still slower than generic arithmetic, |
| 3631 |
however. |
| 3632 |
|
| 3633 |
** New procedure: `define-inlinable' |
| 3634 |
|
| 3635 |
See "Inlinable Procedures" in the manual, for more. |
| 3636 |
|
| 3637 |
** New procedure: `exact-integer-sqrt' |
| 3638 |
|
| 3639 |
See "Integer Operations" in the manual, for more. |
| 3640 |
|
| 3641 |
** "Extended read syntax" for symbols parses better |
| 3642 |
|
| 3643 |
In #{foo}# symbols, backslashes are now treated as escapes, as the |
| 3644 |
symbol-printing code intended. Additionally, "\x" within #{foo}# is now |
| 3645 |
interpreted as starting an R6RS hex escape. This is backward compatible |
| 3646 |
because the symbol printer would never produce a "\x" before. The |
| 3647 |
printer also works better too. |
| 3648 |
|
| 3649 |
** Added `--fresh-auto-compile' option |
| 3650 |
|
| 3651 |
This allows a user to invalidate the auto-compilation cache. It's |
| 3652 |
usually not needed. See "Compilation" in the manual, for a discussion. |
| 3653 |
|
| 3654 |
* Manual updates |
| 3655 |
|
| 3656 |
** GOOPS documentation updates |
| 3657 |
|
| 3658 |
** New man page |
| 3659 |
|
| 3660 |
Thanks to Mark Harig for improvements to guile.1. |
| 3661 |
|
| 3662 |
** SRFI-23 documented |
| 3663 |
|
| 3664 |
The humble `error' SRFI now has an entry in the manual. |
| 3665 |
|
| 3666 |
* New modules |
| 3667 |
|
| 3668 |
** `(ice-9 binary-ports)': "R6RS I/O Ports", in the manual |
| 3669 |
** `(ice-9 eval-string)': "Fly Evaluation", in the manual |
| 3670 |
** `(ice-9 command-line)', not documented yet |
| 3671 |
|
| 3672 |
* Bugs fixed |
| 3673 |
|
| 3674 |
** Fixed `iconv_t' memory leak on close-port |
| 3675 |
** Fixed some leaks with weak hash tables |
| 3676 |
** Export `vhash-delq' and `vhash-delv' from `(ice-9 vlist)' |
| 3677 |
** `after-gc-hook' works again |
| 3678 |
** `define-record-type' now allowed in nested contexts |
| 3679 |
** `exact-integer-sqrt' now handles large integers correctly |
| 3680 |
** Fixed C extension examples in manual |
| 3681 |
** `vhash-delete' honors HASH argument |
| 3682 |
** Make `locale-digit-grouping' more robust |
| 3683 |
** Default exception printer robustness fixes |
| 3684 |
** Fix presence of non-I CPPFLAGS in `guile-2.0.pc' |
| 3685 |
** `read' updates line/column numbers when reading SCSH block comments |
| 3686 |
** Fix imports of multiple custom interfaces of same module |
| 3687 |
** Fix encoding scanning for non-seekable ports |
| 3688 |
** Fix `setter' when called with a non-setter generic |
| 3689 |
** Fix f32 and f64 bytevectors to not accept rationals |
| 3690 |
** Fix description of the R6RS `finite?' in manual |
| 3691 |
** Quotient, remainder and modulo accept inexact integers again |
| 3692 |
** Fix `continue' within `while' to take zero arguments |
| 3693 |
** Fix alignment for structures in FFI |
| 3694 |
** Fix port-filename of stdin, stdout, stderr to match the docs |
| 3695 |
** Fix weak hash table-related bug in `define-wrapped-pointer-type' |
| 3696 |
** Fix partial continuation application with pending procedure calls |
| 3697 |
** scm_{to,from}_locale_string use current locale, not current ports |
| 3698 |
** Fix thread cleanup, by using a pthread_key destructor |
| 3699 |
** Fix `quit' at the REPL |
| 3700 |
** Fix a failure to sync regs in vm bytevector ops |
| 3701 |
** Fix (texinfo reflection) to handle nested structures like syntax patterns |
| 3702 |
** Fix stexi->html double translation |
| 3703 |
** Fix tree-il->scheme fix for <prompt> |
| 3704 |
** Fix compilation of <prompt> in <fix> in single-value context |
| 3705 |
** Fix race condition in ensure-writable-dir |
| 3706 |
** Fix error message on ,disassemble "non-procedure" |
| 3707 |
** Fix prompt and abort with the boot evaluator |
| 3708 |
** Fix `procedure->pointer' for functions returning `void' |
| 3709 |
** Fix error reporting in dynamic-pointer |
| 3710 |
** Fix problems detecting coding: in block comments |
| 3711 |
** Fix duplicate load-path and load-compiled-path in compilation environment |
| 3712 |
** Add fallback read(2) suppport for .go files if mmap(2) unavailable |
| 3713 |
** Fix c32vector-set!, c64vector-set! |
| 3714 |
** Fix mistakenly deprecated read syntax for uniform complex vectors |
| 3715 |
** Fix parsing of exact numbers with negative exponents |
| 3716 |
** Ignore SIGPIPE in (system repl server) |
| 3717 |
** Fix optional second arg to R6RS log function |
| 3718 |
** Fix R6RS `assert' to return true value. |
| 3719 |
** Fix fencepost error when seeking in bytevector input ports |
| 3720 |
** Gracefully handle `setlocale' errors when starting the REPL |
| 3721 |
** Improve support of the `--disable-posix' configure option |
| 3722 |
** Make sure R6RS binary ports pass `binary-port?' regardless of the locale |
| 3723 |
** Gracefully handle unterminated UTF-8 sequences instead of hitting an `assert' |
| 3724 |
|
| 3725 |
|
| 3726 |
|
| 3727 |
Changes in 2.0.0 (changes since the 1.8.x series): |
| 3728 |
|
| 3729 |
* New modules (see the manual for details) |
| 3730 |
|
| 3731 |
** `(srfi srfi-18)', more sophisticated multithreading support |
| 3732 |
** `(srfi srfi-27)', sources of random bits |
| 3733 |
** `(srfi srfi-38)', External Representation for Data With Shared Structure |
| 3734 |
** `(srfi srfi-42)', eager comprehensions |
| 3735 |
** `(srfi srfi-45)', primitives for expressing iterative lazy algorithms |
| 3736 |
** `(srfi srfi-67)', compare procedures |
| 3737 |
** `(ice-9 i18n)', internationalization support |
| 3738 |
** `(ice-9 futures)', fine-grain parallelism |
| 3739 |
** `(rnrs bytevectors)', the R6RS bytevector API |
| 3740 |
** `(rnrs io ports)', a subset of the R6RS I/O port API |
| 3741 |
** `(system xref)', a cross-referencing facility (FIXME undocumented) |
| 3742 |
** `(ice-9 vlist)', lists with constant-time random access; hash lists |
| 3743 |
** `(system foreign)', foreign function interface |
| 3744 |
** `(sxml match)', a pattern matcher for SXML |
| 3745 |
** `(srfi srfi-9 gnu)', extensions to the SRFI-9 record library |
| 3746 |
** `(system vm coverage)', a line-by-line code coverage library |
| 3747 |
** `(web uri)', URI data type, parser, and unparser |
| 3748 |
** `(web http)', HTTP header parsers and unparsers |
| 3749 |
** `(web request)', HTTP request data type, reader, and writer |
| 3750 |
** `(web response)', HTTP response data type, reader, and writer |
| 3751 |
** `(web server)', Generic HTTP server |
| 3752 |
** `(ice-9 poll)', a poll wrapper |
| 3753 |
** `(web server http)', HTTP-over-TCP web server implementation |
| 3754 |
|
| 3755 |
** Replaced `(ice-9 match)' with Alex Shinn's compatible, hygienic matcher. |
| 3756 |
|
| 3757 |
Guile's copy of Andrew K. Wright's `match' library has been replaced by |
| 3758 |
a compatible hygienic implementation by Alex Shinn. It is now |
| 3759 |
documented, see "Pattern Matching" in the manual. |
| 3760 |
|
| 3761 |
Compared to Andrew K. Wright's `match', the new `match' lacks |
| 3762 |
`match-define', `match:error-control', `match:set-error-control', |
| 3763 |
`match:error', `match:set-error', and all structure-related procedures. |
| 3764 |
|
| 3765 |
** Imported statprof, SSAX, and texinfo modules from Guile-Lib |
| 3766 |
|
| 3767 |
The statprof statistical profiler, the SSAX XML toolkit, and the texinfo |
| 3768 |
toolkit from Guile-Lib have been imported into Guile proper. See |
| 3769 |
"Standard Library" in the manual for more details. |
| 3770 |
|
| 3771 |
** Integration of lalr-scm, a parser generator |
| 3772 |
|
| 3773 |
Guile has included Dominique Boucher's fine `lalr-scm' parser generator |
| 3774 |
as `(system base lalr)'. See "LALR(1) Parsing" in the manual, for more |
| 3775 |
information. |
| 3776 |
|
| 3777 |
* Changes to the stand-alone interpreter |
| 3778 |
|
| 3779 |
** Guile now can compile Scheme to bytecode for a custom virtual machine. |
| 3780 |
|
| 3781 |
Compiled code loads much faster than Scheme source code, and runs around |
| 3782 |
3 or 4 times as fast, generating much less garbage in the process. |
| 3783 |
|
| 3784 |
** Evaluating Scheme code does not use the C stack. |
| 3785 |
|
| 3786 |
Besides when compiling Guile itself, Guile no longer uses a recursive C |
| 3787 |
function as an evaluator. This obviates the need to check the C stack |
| 3788 |
pointer for overflow. Continuations still capture the C stack, however. |
| 3789 |
|
| 3790 |
** New environment variables: GUILE_LOAD_COMPILED_PATH, |
| 3791 |
GUILE_SYSTEM_LOAD_COMPILED_PATH |
| 3792 |
|
| 3793 |
GUILE_LOAD_COMPILED_PATH is for compiled files what GUILE_LOAD_PATH is |
| 3794 |
for source files. It is a different path, however, because compiled |
| 3795 |
files are architecture-specific. GUILE_SYSTEM_LOAD_COMPILED_PATH is like |
| 3796 |
GUILE_SYSTEM_PATH. |
| 3797 |
|
| 3798 |
** New read-eval-print loop (REPL) implementation |
| 3799 |
|
| 3800 |
Running Guile with no arguments drops the user into the new REPL. See |
| 3801 |
"Using Guile Interactively" in the manual, for more information. |
| 3802 |
|
| 3803 |
** Remove old Emacs interface |
| 3804 |
|
| 3805 |
Guile had an unused `--emacs' command line argument that was supposed to |
| 3806 |
help when running Guile inside Emacs. This option has been removed, and |
| 3807 |
the helper functions `named-module-use!' and `load-emacs-interface' have |
| 3808 |
been deprecated. |
| 3809 |
|
| 3810 |
** Add `(system repl server)' module and `--listen' command-line argument |
| 3811 |
|
| 3812 |
The `(system repl server)' module exposes procedures to listen on |
| 3813 |
sockets for connections, and serve REPLs to those clients. The --listen |
| 3814 |
command-line argument allows any Guile program to thus be remotely |
| 3815 |
debuggable. |
| 3816 |
|
| 3817 |
See "Invoking Guile" for more information on `--listen'. |
| 3818 |
|
| 3819 |
** Command line additions |
| 3820 |
|
| 3821 |
The guile binary now supports a new switch "-x", which can be used to |
| 3822 |
extend the list of filename extensions tried when loading files |
| 3823 |
(%load-extensions). |
| 3824 |
|
| 3825 |
** New reader options: `square-brackets', `r6rs-hex-escapes', |
| 3826 |
`hungry-eol-escapes' |
| 3827 |
|
| 3828 |
The reader supports a new option (changeable via `read-options'), |
| 3829 |
`square-brackets', which instructs it to interpret square brackets as |
| 3830 |
parentheses. This option is on by default. |
| 3831 |
|
| 3832 |
When the new `r6rs-hex-escapes' reader option is enabled, the reader |
| 3833 |
will recognize string escape sequences as defined in R6RS. R6RS string |
| 3834 |
escape sequences are incompatible with Guile's existing escapes, though, |
| 3835 |
so this option is off by default. |
| 3836 |
|
| 3837 |
Additionally, Guile follows the R6RS newline escaping rules when the |
| 3838 |
`hungry-eol-escapes' option is enabled. |
| 3839 |
|
| 3840 |
See "String Syntax" in the manual, for more information. |
| 3841 |
|
| 3842 |
** Function profiling and tracing at the REPL |
| 3843 |
|
| 3844 |
The `,profile FORM' REPL meta-command can now be used to statistically |
| 3845 |
profile execution of a form, to see which functions are taking the most |
| 3846 |
time. See `,help profile' for more information. |
| 3847 |
|
| 3848 |
Similarly, `,trace FORM' traces all function applications that occur |
| 3849 |
during the execution of `FORM'. See `,help trace' for more information. |
| 3850 |
|
| 3851 |
** Recursive debugging REPL on error |
| 3852 |
|
| 3853 |
When Guile sees an error at the REPL, instead of saving the stack, Guile |
| 3854 |
will directly enter a recursive REPL in the dynamic context of the |
| 3855 |
error. See "Error Handling" in the manual, for more information. |
| 3856 |
|
| 3857 |
A recursive REPL is the same as any other REPL, except that it |
| 3858 |
has been augmented with debugging information, so that one can inspect |
| 3859 |
the context of the error. The debugger has been integrated with the REPL |
| 3860 |
via a set of debugging meta-commands. |
| 3861 |
|
| 3862 |
For example, one may access a backtrace with `,backtrace' (or |
| 3863 |
`,bt'). See "Interactive Debugging" in the manual, for more |
| 3864 |
information. |
| 3865 |
|
| 3866 |
** New `guile-tools' commands: `compile', `disassemble' |
| 3867 |
|
| 3868 |
Pass the `--help' command-line option to these commands for more |
| 3869 |
information. |
| 3870 |
|
| 3871 |
** Guile now adds its install prefix to the LTDL_LIBRARY_PATH |
| 3872 |
|
| 3873 |
Users may now install Guile to nonstandard prefixes and just run |
| 3874 |
`/path/to/bin/guile', instead of also having to set LTDL_LIBRARY_PATH to |
| 3875 |
include `/path/to/lib'. |
| 3876 |
|
| 3877 |
** Guile's Emacs integration is now more keyboard-friendly |
| 3878 |
|
| 3879 |
Backtraces may now be disclosed with the keyboard in addition to the |
| 3880 |
mouse. |
| 3881 |
|
| 3882 |
** Load path change: search in version-specific paths before site paths |
| 3883 |
|
| 3884 |
When looking for a module, Guile now searches first in Guile's |
| 3885 |
version-specific path (the library path), *then* in the site dir. This |
| 3886 |
allows Guile's copy of SSAX to override any Guile-Lib copy the user has |
| 3887 |
installed. Also it should cut the number of `stat' system calls by half, |
| 3888 |
in the common case. |
| 3889 |
|
| 3890 |
** Value history in the REPL on by default |
| 3891 |
|
| 3892 |
By default, the REPL will save computed values in variables like `$1', |
| 3893 |
`$2', and the like. There are programmatic and interactive interfaces to |
| 3894 |
control this. See "Value History" in the manual, for more information. |
| 3895 |
|
| 3896 |
** Readline tab completion for arguments |
| 3897 |
|
| 3898 |
When readline is enabled, tab completion works for arguments too, not |
| 3899 |
just for the operator position. |
| 3900 |
|
| 3901 |
** Expression-oriented readline history |
| 3902 |
|
| 3903 |
Guile's readline history now tries to operate on expressions instead of |
| 3904 |
input lines. Let us know what you think! |
| 3905 |
|
| 3906 |
** Interactive Guile follows GNU conventions |
| 3907 |
|
| 3908 |
As recommended by the GPL, Guile now shows a brief copyright and |
| 3909 |
warranty disclaimer on startup, along with pointers to more information. |
| 3910 |
|
| 3911 |
* Changes to Scheme functions and syntax |
| 3912 |
|
| 3913 |
** Support for R6RS libraries |
| 3914 |
|
| 3915 |
The `library' and `import' forms from the latest Scheme report have been |
| 3916 |
added to Guile, in such a way that R6RS libraries share a namespace with |
| 3917 |
Guile modules. R6RS modules may import Guile modules, and are available |
| 3918 |
for Guile modules to import via use-modules and all the rest. See "R6RS |
| 3919 |
Libraries" in the manual for more information. |
| 3920 |
|
| 3921 |
** Implementations of R6RS libraries |
| 3922 |
|
| 3923 |
Guile now has implementations for all of the libraries defined in the |
| 3924 |
R6RS. Thanks to Julian Graham for this excellent hack. See "R6RS |
| 3925 |
Standard Libraries" in the manual for a full list of libraries. |
| 3926 |
|
| 3927 |
** Partial R6RS compatibility |
| 3928 |
|
| 3929 |
Guile now has enough support for R6RS to run a reasonably large subset |
| 3930 |
of R6RS programs. |
| 3931 |
|
| 3932 |
Guile is not fully R6RS compatible. Many incompatibilities are simply |
| 3933 |
bugs, though some parts of Guile will remain R6RS-incompatible for the |
| 3934 |
foreseeable future. See "R6RS Incompatibilities" in the manual, for more |
| 3935 |
information. |
| 3936 |
|
| 3937 |
Please contact [email protected] if you have found an issue not |
| 3938 |
mentioned in that compatibility list. |
| 3939 |
|
| 3940 |
** New implementation of `primitive-eval' |
| 3941 |
|
| 3942 |
Guile's `primitive-eval' is now implemented in Scheme. Actually there is |
| 3943 |
still a C evaluator, used when building a fresh Guile to interpret the |
| 3944 |
compiler, so we can compile eval.scm. Thereafter all calls to |
| 3945 |
primitive-eval are implemented by VM-compiled code. |
| 3946 |
|
| 3947 |
This allows all of Guile's procedures, be they interpreted or compiled, |
| 3948 |
to execute on the same stack, unifying multiple-value return semantics, |
| 3949 |
providing for proper tail recursion between interpreted and compiled |
| 3950 |
code, and simplifying debugging. |
| 3951 |
|
| 3952 |
As part of this change, the evaluator no longer mutates the internal |
| 3953 |
representation of the code being evaluated in a thread-unsafe manner. |
| 3954 |
|
| 3955 |
There are two negative aspects of this change, however. First, Guile |
| 3956 |
takes a lot longer to compile now. Also, there is less debugging |
| 3957 |
information available for debugging interpreted code. We hope to improve |
| 3958 |
both of these situations. |
| 3959 |
|
| 3960 |
There are many changes to the internal C evalator interface, but all |
| 3961 |
public interfaces should be the same. See the ChangeLog for details. If |
| 3962 |
we have inadvertantly changed an interface that you were using, please |
| 3963 |
contact [email protected]. |
| 3964 |
|
| 3965 |
** Procedure removed: `the-environment' |
| 3966 |
|
| 3967 |
This procedure was part of the interpreter's execution model, and does |
| 3968 |
not apply to the compiler. |
| 3969 |
|
| 3970 |
** No more `local-eval' |
| 3971 |
|
| 3972 |
`local-eval' used to exist so that one could evaluate code in the |
| 3973 |
lexical context of a function. Since there is no way to get the lexical |
| 3974 |
environment any more, as that concept has no meaning for the compiler, |
| 3975 |
and a different meaning for the interpreter, we have removed the |
| 3976 |
function. |
| 3977 |
|
| 3978 |
If you think you need `local-eval', you should probably implement your |
| 3979 |
own metacircular evaluator. It will probably be as fast as Guile's |
| 3980 |
anyway. |
| 3981 |
|
| 3982 |
** Scheme source files will now be compiled automatically. |
| 3983 |
|
| 3984 |
If a compiled .go file corresponding to a .scm file is not found or is |
| 3985 |
not fresh, the .scm file will be compiled on the fly, and the resulting |
| 3986 |
.go file stored away. An advisory note will be printed on the console. |
| 3987 |
|
| 3988 |
Note that this mechanism depends on the timestamp of the .go file being |
| 3989 |
newer than that of the .scm file; if the .scm or .go files are moved |
| 3990 |
after installation, care should be taken to preserve their original |
| 3991 |
timestamps. |
| 3992 |
|
| 3993 |
Auto-compiled files will be stored in the $XDG_CACHE_HOME/guile/ccache |
| 3994 |
directory, where $XDG_CACHE_HOME defaults to ~/.cache. This directory |
| 3995 |
will be created if needed. |
| 3996 |
|
| 3997 |
To inhibit automatic compilation, set the GUILE_AUTO_COMPILE environment |
| 3998 |
variable to 0, or pass --no-auto-compile on the Guile command line. |
| 3999 |
|
| 4000 |
** New POSIX procedures: `getrlimit' and `setrlimit' |
| 4001 |
|
| 4002 |
Note however that the interface of these functions is likely to change |
| 4003 |
in the next prerelease. |
| 4004 |
|
| 4005 |
** New POSIX procedure: `getsid' |
| 4006 |
|
| 4007 |
Scheme binding for the `getsid' C library call. |
| 4008 |
|
| 4009 |
** New POSIX procedure: `getaddrinfo' |
| 4010 |
|
| 4011 |
Scheme binding for the `getaddrinfo' C library function. |
| 4012 |
|
| 4013 |
** Multicast socket options |
| 4014 |
|
| 4015 |
Support was added for the IP_MULTICAST_TTL and IP_MULTICAST_IF socket |
| 4016 |
options. See "Network Sockets and Communication" in the manual, for |
| 4017 |
more information. |
| 4018 |
|
| 4019 |
** `recv!', `recvfrom!', `send', `sendto' now deal in bytevectors |
| 4020 |
|
| 4021 |
These socket procedures now take bytevectors as arguments, instead of |
| 4022 |
strings. There is some deprecated string support, however. |
| 4023 |
|
| 4024 |
** New GNU procedures: `setaffinity' and `getaffinity'. |
| 4025 |
|
| 4026 |
See "Processes" in the manual, for more information. |
| 4027 |
|
| 4028 |
** New procedures: `compose', `negate', and `const' |
| 4029 |
|
| 4030 |
See "Higher-Order Functions" in the manual, for more information. |
| 4031 |
|
| 4032 |
** New procedure in `(oops goops)': `method-formals' |
| 4033 |
|
| 4034 |
** New procedures in (ice-9 session): `add-value-help-handler!', |
| 4035 |
`remove-value-help-handler!', `add-name-help-handler!' |
| 4036 |
`remove-name-help-handler!', `procedure-arguments' |
| 4037 |
|
| 4038 |
The value and name help handlers provide some minimal extensibility to |
| 4039 |
the help interface. Guile-lib's `(texinfo reflection)' uses them, for |
| 4040 |
example, to make stexinfo help documentation available. See those |
| 4041 |
procedures' docstrings for more information. |
| 4042 |
|
| 4043 |
`procedure-arguments' describes the arguments that a procedure can take, |
| 4044 |
combining arity and formals. For example: |
| 4045 |
|
| 4046 |
(procedure-arguments resolve-interface) |
| 4047 |
=> ((required . (name)) (rest . args)) |
| 4048 |
|
| 4049 |
Additionally, `module-commentary' is now publically exported from |
| 4050 |
`(ice-9 session). |
| 4051 |
|
| 4052 |
** Removed: `procedure->memoizing-macro', `procedure->syntax' |
| 4053 |
|
| 4054 |
These procedures created primitive fexprs for the old evaluator, and are |
| 4055 |
no longer supported. If you feel that you need these functions, you |
| 4056 |
probably need to write your own metacircular evaluator (which will |
| 4057 |
probably be as fast as Guile's, anyway). |
| 4058 |
|
| 4059 |
** New language: ECMAScript |
| 4060 |
|
| 4061 |
Guile now ships with one other high-level language supported, |
| 4062 |
ECMAScript. The goal is to support all of version 3.1 of the standard, |
| 4063 |
but not all of the libraries are there yet. This support is not yet |
| 4064 |
documented; ask on the mailing list if you are interested. |
| 4065 |
|
| 4066 |
** New language: Brainfuck |
| 4067 |
|
| 4068 |
Brainfuck is a toy language that closely models Turing machines. Guile's |
| 4069 |
brainfuck compiler is meant to be an example of implementing other |
| 4070 |
languages. See the manual for details, or |
| 4071 |
http://en.wikipedia.org/wiki/Brainfuck for more information about the |
| 4072 |
Brainfuck language itself. |
| 4073 |
|
| 4074 |
** New language: Elisp |
| 4075 |
|
| 4076 |
Guile now has an experimental Emacs Lisp compiler and runtime. You can |
| 4077 |
now switch to Elisp at the repl: `,language elisp'. All kudos to Daniel |
| 4078 |
Kraft and Brian Templeton, and all bugs to [email protected]. |
| 4079 |
|
| 4080 |
** Better documentation infrastructure for macros |
| 4081 |
|
| 4082 |
It is now possible to introspect on the type of a macro, e.g. |
| 4083 |
syntax-rules, identifier-syntax, etc, and extract information about that |
| 4084 |
macro, such as the syntax-rules patterns or the defmacro arguments. |
| 4085 |
`(texinfo reflection)' takes advantage of this to give better macro |
| 4086 |
documentation. |
| 4087 |
|
| 4088 |
** Support for arbitrary procedure metadata |
| 4089 |
|
| 4090 |
Building on its support for docstrings, Guile now supports multiple |
| 4091 |
docstrings, adding them to the tail of a compiled procedure's |
| 4092 |
properties. For example: |
| 4093 |
|
| 4094 |
(define (foo) |
| 4095 |
"one" |
| 4096 |
"two" |
| 4097 |
3) |
| 4098 |
(procedure-properties foo) |
| 4099 |
=> ((name . foo) (documentation . "one") (documentation . "two")) |
| 4100 |
|
| 4101 |
Also, vectors of pairs are now treated as additional metadata entries: |
| 4102 |
|
| 4103 |
(define (bar) |
| 4104 |
#((quz . #f) (docstring . "xyzzy")) |
| 4105 |
3) |
| 4106 |
(procedure-properties bar) |
| 4107 |
=> ((name . bar) (quz . #f) (docstring . "xyzzy")) |
| 4108 |
|
| 4109 |
This allows arbitrary literals to be embedded as metadata in a compiled |
| 4110 |
procedure. |
| 4111 |
|
| 4112 |
** The psyntax expander now knows how to interpret the @ and @@ special |
| 4113 |
forms. |
| 4114 |
|
| 4115 |
** The psyntax expander is now hygienic with respect to modules. |
| 4116 |
|
| 4117 |
Free variables in a macro are scoped in the module that the macro was |
| 4118 |
defined in, not in the module the macro is used in. For example, code |
| 4119 |
like this works now: |
| 4120 |
|
| 4121 |
(define-module (foo) #:export (bar)) |
| 4122 |
(define (helper x) ...) |
| 4123 |
(define-syntax bar |
| 4124 |
(syntax-rules () ((_ x) (helper x)))) |
| 4125 |
|
| 4126 |
(define-module (baz) #:use-module (foo)) |
| 4127 |
(bar qux) |
| 4128 |
|
| 4129 |
It used to be you had to export `helper' from `(foo)' as well. |
| 4130 |
Thankfully, this has been fixed. |
| 4131 |
|
| 4132 |
** Support for version information in Guile's `module' form |
| 4133 |
|
| 4134 |
Guile modules now have a `#:version' field. See "R6RS Version |
| 4135 |
References", "General Information about Modules", "Using Guile Modules", |
| 4136 |
and "Creating Guile Modules" in the manual for more information. |
| 4137 |
|
| 4138 |
** Support for renaming bindings on module export |
| 4139 |
|
| 4140 |
Wherever Guile accepts a symbol as an argument to specify a binding to |
| 4141 |
export, it now also accepts a pair of symbols, indicating that a binding |
| 4142 |
should be renamed on export. See "Creating Guile Modules" in the manual |
| 4143 |
for more information. |
| 4144 |
|
| 4145 |
** New procedure: `module-export-all!' |
| 4146 |
|
| 4147 |
This procedure exports all current and future bindings from a module. |
| 4148 |
Use as `(module-export-all! (current-module))'. |
| 4149 |
|
| 4150 |
** New procedure `reload-module', and `,reload' REPL command |
| 4151 |
|
| 4152 |
See "Module System Reflection" and "Module Commands" in the manual, for |
| 4153 |
more information. |
| 4154 |
|
| 4155 |
** `eval-case' has been deprecated, and replaced by `eval-when'. |
| 4156 |
|
| 4157 |
The semantics of `eval-when' are easier to understand. See "Eval When" |
| 4158 |
in the manual, for more information. |
| 4159 |
|
| 4160 |
** Guile is now more strict about prohibiting definitions in expression |
| 4161 |
contexts. |
| 4162 |
|
| 4163 |
Although previous versions of Guile accepted it, the following |
| 4164 |
expression is not valid, in R5RS or R6RS: |
| 4165 |
|
| 4166 |
(if test (define foo 'bar) (define foo 'baz)) |
| 4167 |
|
| 4168 |
In this specific case, it would be better to do: |
| 4169 |
|
| 4170 |
(define foo (if test 'bar 'baz)) |
| 4171 |
|
| 4172 |
It is possible to circumvent this restriction with e.g. |
| 4173 |
`(module-define! (current-module) 'foo 'baz)'. Contact the list if you |
| 4174 |
have any questions. |
| 4175 |
|
| 4176 |
** Support for `letrec*' |
| 4177 |
|
| 4178 |
Guile now supports `letrec*', a recursive lexical binding operator in |
| 4179 |
which the identifiers are bound in order. See "Local Bindings" in the |
| 4180 |
manual, for more details. |
| 4181 |
|
| 4182 |
** Internal definitions now expand to `letrec*' |
| 4183 |
|
| 4184 |
Following the R6RS, internal definitions now expand to letrec* instead |
| 4185 |
of letrec. The following program is invalid for R5RS, but valid for |
| 4186 |
R6RS: |
| 4187 |
|
| 4188 |
(define (foo) |
| 4189 |
(define bar 10) |
| 4190 |
(define baz (+ bar 20)) |
| 4191 |
baz) |
| 4192 |
|
| 4193 |
;; R5RS and Guile <= 1.8: |
| 4194 |
(foo) => Unbound variable: bar |
| 4195 |
;; R6RS and Guile >= 2.0: |
| 4196 |
(foo) => 30 |
| 4197 |
|
| 4198 |
This change should not affect correct R5RS programs, or programs written |
| 4199 |
in earlier Guile dialects. |
| 4200 |
|
| 4201 |
** Macro expansion produces structures instead of s-expressions |
| 4202 |
|
| 4203 |
In the olden days, macroexpanding an s-expression would yield another |
| 4204 |
s-expression. Though the lexical variables were renamed, expansions of |
| 4205 |
core forms like `if' and `begin' were still non-hygienic, as they relied |
| 4206 |
on the toplevel definitions of `if' et al being the conventional ones. |
| 4207 |
|
| 4208 |
The solution is to expand to structures instead of s-expressions. There |
| 4209 |
is an `if' structure, a `begin' structure, a `toplevel-ref' structure, |
| 4210 |
etc. The expander already did this for compilation, producing Tree-IL |
| 4211 |
directly; it has been changed now to do so when expanding for the |
| 4212 |
evaluator as well. |
| 4213 |
|
| 4214 |
** Defmacros must now produce valid Scheme expressions. |
| 4215 |
|
| 4216 |
It used to be that defmacros could unquote in Scheme values, as a way of |
| 4217 |
supporting partial evaluation, and avoiding some hygiene issues. For |
| 4218 |
example: |
| 4219 |
|
| 4220 |
(define (helper x) ...) |
| 4221 |
(define-macro (foo bar) |
| 4222 |
`(,helper ,bar)) |
| 4223 |
|
| 4224 |
Assuming this macro is in the `(baz)' module, the direct translation of |
| 4225 |
this code would be: |
| 4226 |
|
| 4227 |
(define (helper x) ...) |
| 4228 |
(define-macro (foo bar) |
| 4229 |
`((@@ (baz) helper) ,bar)) |
| 4230 |
|
| 4231 |
Of course, one could just use a hygienic macro instead: |
| 4232 |
|
| 4233 |
(define-syntax foo |
| 4234 |
(syntax-rules () |
| 4235 |
((_ bar) (helper bar)))) |
| 4236 |
|
| 4237 |
** Guile's psyntax now supports docstrings and internal definitions. |
| 4238 |
|
| 4239 |
The following Scheme is not strictly legal: |
| 4240 |
|
| 4241 |
(define (foo) |
| 4242 |
"bar" |
| 4243 |
(define (baz) ...) |
| 4244 |
(baz)) |
| 4245 |
|
| 4246 |
However its intent is fairly clear. Guile interprets "bar" to be the |
| 4247 |
docstring of `foo', and the definition of `baz' is still in definition |
| 4248 |
context. |
| 4249 |
|
| 4250 |
** Support for settable identifier syntax |
| 4251 |
|
| 4252 |
Following the R6RS, "variable transformers" are settable |
| 4253 |
identifier-syntax. See "Identifier macros" in the manual, for more |
| 4254 |
information. |
| 4255 |
|
| 4256 |
** syntax-case treats `_' as a placeholder |
| 4257 |
|
| 4258 |
Following R6RS, a `_' in a syntax-rules or syntax-case pattern matches |
| 4259 |
anything, and binds no pattern variables. Unlike the R6RS, Guile also |
| 4260 |
permits `_' to be in the literals list for a pattern. |
| 4261 |
|
| 4262 |
** Macros need to be defined before their first use. |
| 4263 |
|
| 4264 |
It used to be that with lazy memoization, this might work: |
| 4265 |
|
| 4266 |
(define (foo x) |
| 4267 |
(ref x)) |
| 4268 |
(define-macro (ref x) x) |
| 4269 |
(foo 1) => 1 |
| 4270 |
|
| 4271 |
But now, the body of `foo' is interpreted to mean a call to the toplevel |
| 4272 |
`ref' function, instead of a macro expansion. The solution is to define |
| 4273 |
macros before code that uses them. |
| 4274 |
|
| 4275 |
** Functions needed by macros at expand-time need to be present at |
| 4276 |
expand-time. |
| 4277 |
|
| 4278 |
For example, this code will work at the REPL: |
| 4279 |
|
| 4280 |
(define (double-helper x) (* x x)) |
| 4281 |
(define-macro (double-literal x) (double-helper x)) |
| 4282 |
(double-literal 2) => 4 |
| 4283 |
|
| 4284 |
But it will not work when a file is compiled, because the definition of |
| 4285 |
`double-helper' is not present at expand-time. The solution is to wrap |
| 4286 |
the definition of `double-helper' in `eval-when': |
| 4287 |
|
| 4288 |
(eval-when (load compile eval) |
| 4289 |
(define (double-helper x) (* x x))) |
| 4290 |
(define-macro (double-literal x) (double-helper x)) |
| 4291 |
(double-literal 2) => 4 |
| 4292 |
|
| 4293 |
See the documentation for eval-when for more information. |
| 4294 |
|
| 4295 |
** `macroexpand' produces structures, not S-expressions. |
| 4296 |
|
| 4297 |
Given the need to maintain referential transparency, both lexically and |
| 4298 |
modular, the result of expanding Scheme expressions is no longer itself |
| 4299 |
an s-expression. If you want a human-readable approximation of the |
| 4300 |
result of `macroexpand', call `tree-il->scheme' from `(language |
| 4301 |
tree-il)'. |
| 4302 |
|
| 4303 |
** Removed function: `macroexpand-1' |
| 4304 |
|
| 4305 |
It is unclear how to implement `macroexpand-1' with syntax-case, though |
| 4306 |
PLT Scheme does prove that it is possible. |
| 4307 |
|
| 4308 |
** New reader macros: #' #` #, #,@ |
| 4309 |
|
| 4310 |
These macros translate, respectively, to `syntax', `quasisyntax', |
| 4311 |
`unsyntax', and `unsyntax-splicing'. See the R6RS for more information. |
| 4312 |
These reader macros may be overridden by `read-hash-extend'. |
| 4313 |
|
| 4314 |
** Incompatible change to #' |
| 4315 |
|
| 4316 |
Guile did have a #' hash-extension, by default, which just returned the |
| 4317 |
subsequent datum: #'foo => foo. In the unlikely event that anyone |
| 4318 |
actually used this, this behavior may be reinstated via the |
| 4319 |
`read-hash-extend' mechanism. |
| 4320 |
|
| 4321 |
** `unquote' and `unquote-splicing' accept multiple expressions |
| 4322 |
|
| 4323 |
As per the R6RS, these syntax operators can now accept any number of |
| 4324 |
expressions to unquote. |
| 4325 |
|
| 4326 |
** Scheme expresssions may be commented out with #; |
| 4327 |
|
| 4328 |
#; comments out an entire expression. See SRFI-62 or the R6RS for more |
| 4329 |
information. |
| 4330 |
|
| 4331 |
** Prompts: Delimited, composable continuations |
| 4332 |
|
| 4333 |
Guile now has prompts as part of its primitive language. See "Prompts" |
| 4334 |
in the manual, for more information. |
| 4335 |
|
| 4336 |
Expressions entered in at the REPL, or from the command line, are |
| 4337 |
surrounded by a prompt with the default prompt tag. |
| 4338 |
|
| 4339 |
** `make-stack' with a tail-called procedural narrowing argument no longer |
| 4340 |
works (with compiled procedures) |
| 4341 |
|
| 4342 |
It used to be the case that a captured stack could be narrowed to select |
| 4343 |
calls only up to or from a certain procedure, even if that procedure |
| 4344 |
already tail-called another procedure. This was because the debug |
| 4345 |
information from the original procedure was kept on the stack. |
| 4346 |
|
| 4347 |
Now with the new compiler, the stack only contains active frames from |
| 4348 |
the current continuation. A narrow to a procedure that is not in the |
| 4349 |
stack will result in an empty stack. To fix this, narrow to a procedure |
| 4350 |
that is active in the current continuation, or narrow to a specific |
| 4351 |
number of stack frames. |
| 4352 |
|
| 4353 |
** Backtraces through compiled procedures only show procedures that are |
| 4354 |
active in the current continuation |
| 4355 |
|
| 4356 |
Similarly to the previous issue, backtraces in compiled code may be |
| 4357 |
different from backtraces in interpreted code. There are no semantic |
| 4358 |
differences, however. Please mail [email protected] if you see any |
| 4359 |
deficiencies with Guile's backtraces. |
| 4360 |
|
| 4361 |
** `positions' reader option enabled by default |
| 4362 |
|
| 4363 |
This change allows primitive-load without --auto-compile to also |
| 4364 |
propagate source information through the expander, for better errors and |
| 4365 |
to let macros know their source locations. The compiler was already |
| 4366 |
turning it on anyway. |
| 4367 |
|
| 4368 |
** New macro: `current-source-location' |
| 4369 |
|
| 4370 |
The macro returns the current source location (to be documented). |
| 4371 |
|
| 4372 |
** syntax-rules and syntax-case macros now propagate source information |
| 4373 |
through to the expanded code |
| 4374 |
|
| 4375 |
This should result in better backtraces. |
| 4376 |
|
| 4377 |
** The currying behavior of `define' has been removed. |
| 4378 |
|
| 4379 |
Before, `(define ((f a) b) (* a b))' would translate to |
| 4380 |
|
| 4381 |
(define f (lambda (a) (lambda (b) (* a b)))) |
| 4382 |
|
| 4383 |
Now a syntax error is signaled, as this syntax is not supported by |
| 4384 |
default. Use the `(ice-9 curried-definitions)' module to get back the |
| 4385 |
old behavior. |
| 4386 |
|
| 4387 |
** New procedure, `define!' |
| 4388 |
|
| 4389 |
`define!' is a procedure that takes two arguments, a symbol and a value, |
| 4390 |
and binds the value to the symbol in the current module. It's useful to |
| 4391 |
programmatically make definitions in the current module, and is slightly |
| 4392 |
less verbose than `module-define!'. |
| 4393 |
|
| 4394 |
** All modules have names now |
| 4395 |
|
| 4396 |
Before, you could have anonymous modules: modules without names. Now, |
| 4397 |
because of hygiene and macros, all modules have names. If a module was |
| 4398 |
created without a name, the first time `module-name' is called on it, a |
| 4399 |
fresh name will be lazily generated for it. |
| 4400 |
|
| 4401 |
** The module namespace is now separate from the value namespace |
| 4402 |
|
| 4403 |
It was a little-known implementation detail of Guile's module system |
| 4404 |
that it was built on a single hierarchical namespace of values -- that |
| 4405 |
if there was a module named `(foo bar)', then in the module named |
| 4406 |
`(foo)' there was a binding from `bar' to the `(foo bar)' module. |
| 4407 |
|
| 4408 |
This was a neat trick, but presented a number of problems. One problem |
| 4409 |
was that the bindings in a module were not apparent from the module |
| 4410 |
itself; perhaps the `(foo)' module had a private binding for `bar', and |
| 4411 |
then an external contributor defined `(foo bar)'. In the end there can |
| 4412 |
be only one binding, so one of the two will see the wrong thing, and |
| 4413 |
produce an obtuse error of unclear provenance. |
| 4414 |
|
| 4415 |
Also, the public interface of a module was also bound in the value |
| 4416 |
namespace, as `%module-public-interface'. This was a hack from the early |
| 4417 |
days of Guile's modules. |
| 4418 |
|
| 4419 |
Both of these warts have been fixed by the addition of fields in the |
| 4420 |
`module' data type. Access to modules and their interfaces from the |
| 4421 |
value namespace has been deprecated, and all accessors use the new |
| 4422 |
record accessors appropriately. |
| 4423 |
|
| 4424 |
When Guile is built with support for deprecated code, as is the default, |
| 4425 |
the value namespace is still searched for modules and public interfaces, |
| 4426 |
and a deprecation warning is raised as appropriate. |
| 4427 |
|
| 4428 |
Finally, to support lazy loading of modules as one used to be able to do |
| 4429 |
with module binder procedures, Guile now has submodule binders, called |
| 4430 |
if a given submodule is not found. See boot-9.scm for more information. |
| 4431 |
|
| 4432 |
** New procedures: module-ref-submodule, module-define-submodule, |
| 4433 |
nested-ref-module, nested-define-module!, local-ref-module, |
| 4434 |
local-define-module |
| 4435 |
|
| 4436 |
These new accessors are like their bare variants, but operate on |
| 4437 |
namespaces instead of values. |
| 4438 |
|
| 4439 |
** The (app modules) module tree is officially deprecated |
| 4440 |
|
| 4441 |
It used to be that one could access a module named `(foo bar)' via |
| 4442 |
`(nested-ref the-root-module '(app modules foo bar))'. The `(app |
| 4443 |
modules)' bit was a never-used and never-documented abstraction, and has |
| 4444 |
been deprecated. See the following mail for a full discussion: |
| 4445 |
|
| 4446 |
http://lists.gnu.org/archive/html/guile-devel/2010-04/msg00168.html |
| 4447 |
|
| 4448 |
The `%app' binding is also deprecated. |
| 4449 |
|
| 4450 |
** `module-filename' field and accessor |
| 4451 |
|
| 4452 |
Modules now record the file in which they are defined. This field may be |
| 4453 |
accessed with the new `module-filename' procedure. |
| 4454 |
|
| 4455 |
** Modules load within a known environment |
| 4456 |
|
| 4457 |
It takes a few procedure calls to define a module, and those procedure |
| 4458 |
calls need to be in scope. Now we ensure that the current module when |
| 4459 |
loading a module is one that has the needed bindings, instead of relying |
| 4460 |
on chance. |
| 4461 |
|
| 4462 |
** `load' is a macro (!) that resolves paths relative to source file dir |
| 4463 |
|
| 4464 |
The familiar Schem `load' procedure is now a macro that captures the |
| 4465 |
name of the source file being expanded, and dispatches to the new |
| 4466 |
`load-in-vicinity'. Referencing `load' by bare name returns a closure |
| 4467 |
that embeds the current source file name. |
| 4468 |
|
| 4469 |
This fix allows `load' of relative paths to be resolved with respect to |
| 4470 |
the location of the file that calls `load'. |
| 4471 |
|
| 4472 |
** Many syntax errors have different texts now |
| 4473 |
|
| 4474 |
Syntax errors still throw to the `syntax-error' key, but the arguments |
| 4475 |
are often different now. Perhaps in the future, Guile will switch to |
| 4476 |
using standard SRFI-35 conditions. |
| 4477 |
|
| 4478 |
** Returning multiple values to compiled code will silently truncate the |
| 4479 |
values to the expected number |
| 4480 |
|
| 4481 |
For example, the interpreter would raise an error evaluating the form, |
| 4482 |
`(+ (values 1 2) (values 3 4))', because it would see the operands as |
| 4483 |
being two compound "values" objects, to which `+' does not apply. |
| 4484 |
|
| 4485 |
The compiler, on the other hand, receives multiple values on the stack, |
| 4486 |
not as a compound object. Given that it must check the number of values |
| 4487 |
anyway, if too many values are provided for a continuation, it chooses |
| 4488 |
to truncate those values, effectively evaluating `(+ 1 3)' instead. |
| 4489 |
|
| 4490 |
The idea is that the semantics that the compiler implements is more |
| 4491 |
intuitive, and the use of the interpreter will fade out with time. |
| 4492 |
This behavior is allowed both by the R5RS and the R6RS. |
| 4493 |
|
| 4494 |
** Multiple values in compiled code are not represented by compound |
| 4495 |
objects |
| 4496 |
|
| 4497 |
This change may manifest itself in the following situation: |
| 4498 |
|
| 4499 |
(let ((val (foo))) (do-something) val) |
| 4500 |
|
| 4501 |
In the interpreter, if `foo' returns multiple values, multiple values |
| 4502 |
are produced from the `let' expression. In the compiler, those values |
| 4503 |
are truncated to the first value, and that first value is returned. In |
| 4504 |
the compiler, if `foo' returns no values, an error will be raised, while |
| 4505 |
the interpreter would proceed. |
| 4506 |
|
| 4507 |
Both of these behaviors are allowed by R5RS and R6RS. The compiler's |
| 4508 |
behavior is more correct, however. If you wish to preserve a potentially |
| 4509 |
multiply-valued return, you will need to set up a multiple-value |
| 4510 |
continuation, using `call-with-values'. |
| 4511 |
|
| 4512 |
** Defmacros are now implemented in terms of syntax-case. |
| 4513 |
|
| 4514 |
The practical ramification of this is that the `defmacro?' predicate has |
| 4515 |
been removed, along with `defmacro-transformer', `macro-table', |
| 4516 |
`xformer-table', `assert-defmacro?!', `set-defmacro-transformer!' and |
| 4517 |
`defmacro:transformer'. This is because defmacros are simply macros. If |
| 4518 |
any of these procedures provided useful facilities to you, we encourage |
| 4519 |
you to contact the Guile developers. |
| 4520 |
|
| 4521 |
** Hygienic macros documented as the primary syntactic extension mechanism. |
| 4522 |
|
| 4523 |
The macro documentation was finally fleshed out with some documentation |
| 4524 |
on `syntax-rules' and `syntax-case' macros, and other parts of the macro |
| 4525 |
expansion process. See "Macros" in the manual, for details. |
| 4526 |
|
| 4527 |
** psyntax is now the default expander |
| 4528 |
|
| 4529 |
Scheme code is now expanded by default by the psyntax hygienic macro |
| 4530 |
expander. Expansion is performed completely before compilation or |
| 4531 |
interpretation. |
| 4532 |
|
| 4533 |
Notably, syntax errors will be signalled before interpretation begins. |
| 4534 |
In the past, many syntax errors were only detected at runtime if the |
| 4535 |
code in question was memoized. |
| 4536 |
|
| 4537 |
As part of its expansion, psyntax renames all lexically-bound |
| 4538 |
identifiers. Original identifier names are preserved and given to the |
| 4539 |
compiler, but the interpreter will see the renamed variables, e.g., |
| 4540 |
`x432' instead of `x'. |
| 4541 |
|
| 4542 |
Note that the psyntax that Guile uses is a fork, as Guile already had |
| 4543 |
modules before incompatible modules were added to psyntax -- about 10 |
| 4544 |
years ago! Thus there are surely a number of bugs that have been fixed |
| 4545 |
in psyntax since then. If you find one, please notify [email protected]. |
| 4546 |
|
| 4547 |
** syntax-rules and syntax-case are available by default. |
| 4548 |
|
| 4549 |
There is no longer any need to import the `(ice-9 syncase)' module |
| 4550 |
(which is now deprecated). The expander may be invoked directly via |
| 4551 |
`macroexpand', though it is normally searched for via the current module |
| 4552 |
transformer. |
| 4553 |
|
| 4554 |
Also, the helper routines for syntax-case are available in the default |
| 4555 |
environment as well: `syntax->datum', `datum->syntax', |
| 4556 |
`bound-identifier=?', `free-identifier=?', `generate-temporaries', |
| 4557 |
`identifier?', and `syntax-violation'. See the R6RS for documentation. |
| 4558 |
|
| 4559 |
** Tail patterns in syntax-case |
| 4560 |
|
| 4561 |
Guile has pulled in some more recent changes from the psyntax portable |
| 4562 |
syntax expander, to implement support for "tail patterns". Such patterns |
| 4563 |
are supported by syntax-rules and syntax-case. This allows a syntax-case |
| 4564 |
match clause to have ellipses, then a pattern at the end. For example: |
| 4565 |
|
| 4566 |
(define-syntax case |
| 4567 |
(syntax-rules (else) |
| 4568 |
((_ val match-clause ... (else e e* ...)) |
| 4569 |
[...]))) |
| 4570 |
|
| 4571 |
Note how there is MATCH-CLAUSE, which is ellipsized, then there is a |
| 4572 |
tail pattern for the else clause. Thanks to Andreas Rottmann for the |
| 4573 |
patch, and Kent Dybvig for the code. |
| 4574 |
|
| 4575 |
** Lexical bindings introduced by hygienic macros may not be referenced |
| 4576 |
by nonhygienic macros. |
| 4577 |
|
| 4578 |
If a lexical binding is introduced by a hygienic macro, it may not be |
| 4579 |
referenced by a nonhygienic macro. For example, this works: |
| 4580 |
|
| 4581 |
(let () |
| 4582 |
(define-macro (bind-x val body) |
| 4583 |
`(let ((x ,val)) ,body)) |
| 4584 |
(define-macro (ref x) |
| 4585 |
x) |
| 4586 |
(bind-x 10 (ref x))) |
| 4587 |
|
| 4588 |
But this does not: |
| 4589 |
|
| 4590 |
(let () |
| 4591 |
(define-syntax bind-x |
| 4592 |
(syntax-rules () |
| 4593 |
((_ val body) (let ((x val)) body)))) |
| 4594 |
(define-macro (ref x) |
| 4595 |
x) |
| 4596 |
(bind-x 10 (ref x))) |
| 4597 |
|
| 4598 |
It is not normal to run into this situation with existing code. However, |
| 4599 |
if you have defmacros that expand to hygienic macros, it is possible to |
| 4600 |
run into situations like this. For example, if you have a defmacro that |
| 4601 |
generates a `while' expression, the `break' bound by the `while' may not |
| 4602 |
be visible within other parts of your defmacro. The solution is to port |
| 4603 |
from defmacros to syntax-rules or syntax-case. |
| 4604 |
|
| 4605 |
** Macros may no longer be referenced as first-class values. |
| 4606 |
|
| 4607 |
In the past, you could evaluate e.g. `if', and get its macro value. Now, |
| 4608 |
expanding this form raises a syntax error. |
| 4609 |
|
| 4610 |
Macros still /exist/ as first-class values, but they must be |
| 4611 |
/referenced/ via the module system, e.g. `(module-ref (current-module) |
| 4612 |
'if)'. |
| 4613 |
|
| 4614 |
** Macros may now have docstrings. |
| 4615 |
|
| 4616 |
`object-documentation' from `(ice-9 documentation)' may be used to |
| 4617 |
retrieve the docstring, once you have a macro value -- but see the above |
| 4618 |
note about first-class macros. Docstrings are associated with the syntax |
| 4619 |
transformer procedures. |
| 4620 |
|
| 4621 |
** `case-lambda' is now available in the default environment. |
| 4622 |
|
| 4623 |
The binding in the default environment is equivalent to the one from the |
| 4624 |
`(srfi srfi-16)' module. Use the srfi-16 module explicitly if you wish |
| 4625 |
to maintain compatibility with Guile 1.8 and earlier. |
| 4626 |
|
| 4627 |
** Procedures may now have more than one arity. |
| 4628 |
|
| 4629 |
This can be the case, for example, in case-lambda procedures. The |
| 4630 |
arities of compiled procedures may be accessed via procedures from the |
| 4631 |
`(system vm program)' module; see "Compiled Procedures", "Optional |
| 4632 |
Arguments", and "Case-lambda" in the manual. |
| 4633 |
|
| 4634 |
** Deprecate arity access via (procedure-properties proc 'arity) |
| 4635 |
|
| 4636 |
Instead of accessing a procedure's arity as a property, use the new |
| 4637 |
`procedure-minimum-arity' function, which gives the most permissive |
| 4638 |
arity that the function has, in the same format as the old arity |
| 4639 |
accessor. |
| 4640 |
|
| 4641 |
** `lambda*' and `define*' are now available in the default environment |
| 4642 |
|
| 4643 |
As with `case-lambda', `(ice-9 optargs)' continues to be supported, for |
| 4644 |
compatibility purposes. No semantic change has been made (we hope). |
| 4645 |
Optional and keyword arguments now dispatch via special VM operations, |
| 4646 |
without the need to cons rest arguments, making them very fast. |
| 4647 |
|
| 4648 |
** New syntax: define-once |
| 4649 |
|
| 4650 |
`define-once' is like Lisp's `defvar': it creates a toplevel binding, |
| 4651 |
but only if one does not exist already. |
| 4652 |
|
| 4653 |
** New function, `truncated-print', with `format' support |
| 4654 |
|
| 4655 |
`(ice-9 pretty-print)' now exports `truncated-print', a printer that |
| 4656 |
will ensure that the output stays within a certain width, truncating the |
| 4657 |
output in what is hopefully an intelligent manner. See the manual for |
| 4658 |
more details. |
| 4659 |
|
| 4660 |
There is a new `format' specifier, `~@y', for doing a truncated |
| 4661 |
print (as opposed to `~y', which does a pretty-print). See the `format' |
| 4662 |
documentation for more details. |
| 4663 |
|
| 4664 |
** Better pretty-printing |
| 4665 |
|
| 4666 |
Indentation recognizes more special forms, like `syntax-case', and read |
| 4667 |
macros like `quote' are printed better. |
| 4668 |
|
| 4669 |
** Passing a number as the destination of `format' is deprecated |
| 4670 |
|
| 4671 |
The `format' procedure in `(ice-9 format)' now emits a deprecation |
| 4672 |
warning if a number is passed as its first argument. |
| 4673 |
|
| 4674 |
Also, it used to be that you could omit passing a port to `format', in |
| 4675 |
some cases. This still works, but has been formally deprecated. |
| 4676 |
|
| 4677 |
** SRFI-4 vectors reimplemented in terms of R6RS bytevectors |
| 4678 |
|
| 4679 |
Guile now implements SRFI-4 vectors using bytevectors. Often when you |
| 4680 |
have a numeric vector, you end up wanting to write its bytes somewhere, |
| 4681 |
or have access to the underlying bytes, or read in bytes from somewhere |
| 4682 |
else. Bytevectors are very good at this sort of thing. But the SRFI-4 |
| 4683 |
APIs are nicer to use when doing number-crunching, because they are |
| 4684 |
addressed by element and not by byte. |
| 4685 |
|
| 4686 |
So as a compromise, Guile allows all bytevector functions to operate on |
| 4687 |
numeric vectors. They address the underlying bytes in the native |
| 4688 |
endianness, as one would expect. |
| 4689 |
|
| 4690 |
Following the same reasoning, that it's just bytes underneath, Guile |
| 4691 |
also allows uniform vectors of a given type to be accessed as if they |
| 4692 |
were of any type. One can fill a u32vector, and access its elements with |
| 4693 |
u8vector-ref. One can use f64vector-ref on bytevectors. It's all the |
| 4694 |
same to Guile. |
| 4695 |
|
| 4696 |
In this way, uniform numeric vectors may be written to and read from |
| 4697 |
input/output ports using the procedures that operate on bytevectors. |
| 4698 |
|
| 4699 |
Calls to SRFI-4 accessors (ref and set functions) from Scheme are now |
| 4700 |
inlined to the VM instructions for bytevector access. |
| 4701 |
|
| 4702 |
See "SRFI-4" in the manual, for more information. |
| 4703 |
|
| 4704 |
** Nonstandard SRFI-4 procedures now available from `(srfi srfi-4 gnu)' |
| 4705 |
|
| 4706 |
Guile's `(srfi srfi-4)' now only exports those srfi-4 procedures that |
| 4707 |
are part of the standard. Complex uniform vectors and the |
| 4708 |
`any->FOOvector' family are now available only from `(srfi srfi-4 gnu)'. |
| 4709 |
|
| 4710 |
Guile's default environment imports `(srfi srfi-4)', and probably should |
| 4711 |
import `(srfi srfi-4 gnu)' as well. |
| 4712 |
|
| 4713 |
See "SRFI-4 Extensions" in the manual, for more information. |
| 4714 |
|
| 4715 |
** New syntax: include-from-path. |
| 4716 |
|
| 4717 |
`include-from-path' is like `include', except it looks for its file in |
| 4718 |
the load path. It can be used to compile other files into a file. |
| 4719 |
|
| 4720 |
** New syntax: quasisyntax. |
| 4721 |
|
| 4722 |
`quasisyntax' is to `syntax' as `quasiquote' is to `quote'. See the R6RS |
| 4723 |
documentation for more information. Thanks to Andre van Tonder for the |
| 4724 |
implementation. |
| 4725 |
|
| 4726 |
** `*unspecified*' is identifier syntax |
| 4727 |
|
| 4728 |
`*unspecified*' is no longer a variable, so it is optimized properly by |
| 4729 |
the compiler, and is not `set!'-able. |
| 4730 |
|
| 4731 |
** Changes and bugfixes in numerics code |
| 4732 |
|
| 4733 |
*** Added six new sets of fast quotient and remainder operators |
| 4734 |
|
| 4735 |
Added six new sets of fast quotient and remainder operator pairs with |
| 4736 |
different semantics than the R5RS operators. They support not only |
| 4737 |
integers, but all reals, including exact rationals and inexact |
| 4738 |
floating point numbers. |
| 4739 |
|
| 4740 |
These procedures accept two real numbers N and D, where the divisor D |
| 4741 |
must be non-zero. Each set of operators computes an integer quotient |
| 4742 |
Q and a real remainder R such that N = Q*D + R and |R| < |D|. They |
| 4743 |
differ only in how N/D is rounded to produce Q. |
| 4744 |
|
| 4745 |
`euclidean-quotient' returns the integer Q and `euclidean-remainder' |
| 4746 |
returns the real R such that N = Q*D + R and 0 <= R < |D|. `euclidean/' |
| 4747 |
returns both Q and R, and is more efficient than computing each |
| 4748 |
separately. Note that when D > 0, `euclidean-quotient' returns |
| 4749 |
floor(N/D), and when D < 0 it returns ceiling(N/D). |
| 4750 |
|
| 4751 |
`centered-quotient', `centered-remainder', and `centered/' are similar |
| 4752 |
except that the range of remainders is -abs(D/2) <= R < abs(D/2), and |
| 4753 |
`centered-quotient' rounds N/D to the nearest integer. Note that these |
| 4754 |
operators are equivalent to the R6RS integer division operators `div', |
| 4755 |
`mod', `div-and-mod', `div0', `mod0', and `div0-and-mod0'. |
| 4756 |
|
| 4757 |
`floor-quotient' and `floor-remainder' compute Q and R, respectively, |
| 4758 |
where Q has been rounded toward negative infinity. `floor/' returns |
| 4759 |
both Q and R, and is more efficient than computing each separately. |
| 4760 |
Note that when applied to integers, `floor-remainder' is equivalent to |
| 4761 |
the R5RS integer-only `modulo' operator. `ceiling-quotient', |
| 4762 |
`ceiling-remainder', and `ceiling/' are similar except that Q is |
| 4763 |
rounded toward positive infinity. |
| 4764 |
|
| 4765 |
For `truncate-quotient', `truncate-remainder', and `truncate/', Q is |
| 4766 |
rounded toward zero. Note that when applied to integers, |
| 4767 |
`truncate-quotient' and `truncate-remainder' are equivalent to the |
| 4768 |
R5RS integer-only operators `quotient' and `remainder'. |
| 4769 |
|
| 4770 |
For `round-quotient', `round-remainder', and `round/', Q is rounded to |
| 4771 |
the nearest integer, with ties going to the nearest even integer. |
| 4772 |
|
| 4773 |
*** Complex number changes |
| 4774 |
|
| 4775 |
Guile is now able to represent non-real complex numbers whose |
| 4776 |
imaginary part is an _inexact_ zero (0.0 or -0.0), per R6RS. |
| 4777 |
Previously, such numbers were immediately changed into inexact reals. |
| 4778 |
|
| 4779 |
(real? 0.0+0.0i) now returns #f, per R6RS, although (zero? 0.0+0.0i) |
| 4780 |
still returns #t, per R6RS. (= 0 0.0+0.0i) and (= 0.0 0.0+0.0i) are |
| 4781 |
#t, but the same comparisons using `eqv?' or `equal?' are #f. |
| 4782 |
|
| 4783 |
Like other non-real numbers, these complex numbers with inexact zero |
| 4784 |
imaginary part will raise exceptions is passed to procedures requiring |
| 4785 |
reals, such as `<', `>', `<=', `>=', `min', `max', `positive?', |
| 4786 |
`negative?', `inf?', `nan?', `finite?', etc. |
| 4787 |
|
| 4788 |
**** `make-rectangular' changes |
| 4789 |
|
| 4790 |
scm_make_rectangular `make-rectangular' now returns a real number only |
| 4791 |
if the imaginary part is an _exact_ 0. Previously, it would return a |
| 4792 |
real number if the imaginary part was an inexact zero. |
| 4793 |
|
| 4794 |
scm_c_make_rectangular now always returns a non-real complex number, |
| 4795 |
even if the imaginary part is zero. Previously, it would return a |
| 4796 |
real number if the imaginary part was zero. |
| 4797 |
|
| 4798 |
**** `make-polar' changes |
| 4799 |
|
| 4800 |
scm_make_polar `make-polar' now returns a real number only if the |
| 4801 |
angle or magnitude is an _exact_ 0. If the magnitude is an exact 0, |
| 4802 |
it now returns an exact 0. Previously, it would return a real |
| 4803 |
number if the imaginary part was an inexact zero. |
| 4804 |
|
| 4805 |
scm_c_make_polar now always returns a non-real complex number, even if |
| 4806 |
the imaginary part is 0.0. Previously, it would return a real number |
| 4807 |
if the imaginary part was 0.0. |
| 4808 |
|
| 4809 |
**** `imag-part' changes |
| 4810 |
|
| 4811 |
scm_imag_part `imag-part' now returns an exact 0 if applied to an |
| 4812 |
inexact real number. Previously it returned an inexact zero in this |
| 4813 |
case. |
| 4814 |
|
| 4815 |
*** `eqv?' and `equal?' now compare numbers equivalently |
| 4816 |
|
| 4817 |
scm_equal_p `equal?' now behaves equivalently to scm_eqv_p `eqv?' for |
| 4818 |
numeric values, per R5RS. Previously, equal? worked differently, |
| 4819 |
e.g. `(equal? 0.0 -0.0)' returned #t but `(eqv? 0.0 -0.0)' returned #f, |
| 4820 |
and `(equal? +nan.0 +nan.0)' returned #f but `(eqv? +nan.0 +nan.0)' |
| 4821 |
returned #t. |
| 4822 |
|
| 4823 |
*** `(equal? +nan.0 +nan.0)' now returns #t |
| 4824 |
|
| 4825 |
Previously, `(equal? +nan.0 +nan.0)' returned #f, although |
| 4826 |
`(let ((x +nan.0)) (equal? x x))' and `(eqv? +nan.0 +nan.0)' |
| 4827 |
both returned #t. R5RS requires that `equal?' behave like |
| 4828 |
`eqv?' when comparing numbers. |
| 4829 |
|
| 4830 |
*** Change in handling products `*' involving exact 0 |
| 4831 |
|
| 4832 |
scm_product `*' now handles exact 0 differently. A product containing |
| 4833 |
an exact 0 now returns an exact 0 if and only if the other arguments |
| 4834 |
are all exact. An inexact zero is returned if and only if the other |
| 4835 |
arguments are all finite but not all exact. If an infinite or NaN |
| 4836 |
value is present, a NaN value is returned. Previously, any product |
| 4837 |
containing an exact 0 yielded an exact 0, regardless of the other |
| 4838 |
arguments. |
| 4839 |
|
| 4840 |
*** `expt' and `integer-expt' changes when the base is 0 |
| 4841 |
|
| 4842 |
While `(expt 0 0)' is still 1, and `(expt 0 N)' for N > 0 is still |
| 4843 |
zero, `(expt 0 N)' for N < 0 is now a NaN value, and likewise for |
| 4844 |
integer-expt. This is more correct, and conforming to R6RS, but seems |
| 4845 |
to be incompatible with R5RS, which would return 0 for all non-zero |
| 4846 |
values of N. |
| 4847 |
|
| 4848 |
*** `expt' and `integer-expt' are more generic, less strict |
| 4849 |
|
| 4850 |
When raising to an exact non-negative integer exponent, `expt' and |
| 4851 |
`integer-expt' are now able to exponentiate any object that can be |
| 4852 |
multiplied using `*'. They can also raise an object to an exact |
| 4853 |
negative integer power if its reciprocal can be taken using `/'. |
| 4854 |
In order to allow this, the type of the first argument is no longer |
| 4855 |
checked when raising to an exact integer power. If the exponent is 0 |
| 4856 |
or 1, the first parameter is not manipulated at all, and need not |
| 4857 |
even support multiplication. |
| 4858 |
|
| 4859 |
*** Infinities are no longer integers, nor rationals |
| 4860 |
|
| 4861 |
scm_integer_p `integer?' and scm_rational_p `rational?' now return #f |
| 4862 |
for infinities, per R6RS. Previously they returned #t for real |
| 4863 |
infinities. The real infinities and NaNs are still considered real by |
| 4864 |
scm_real `real?' however, per R6RS. |
| 4865 |
|
| 4866 |
*** NaNs are no longer rationals |
| 4867 |
|
| 4868 |
scm_rational_p `rational?' now returns #f for NaN values, per R6RS. |
| 4869 |
Previously it returned #t for real NaN values. They are still |
| 4870 |
considered real by scm_real `real?' however, per R6RS. |
| 4871 |
|
| 4872 |
*** `inf?' and `nan?' now throw exceptions for non-reals |
| 4873 |
|
| 4874 |
The domain of `inf?' and `nan?' is the real numbers. Guile now signals |
| 4875 |
an error when a non-real number or non-number is passed to these |
| 4876 |
procedures. (Note that NaNs _are_ considered numbers by scheme, despite |
| 4877 |
their name). |
| 4878 |
|
| 4879 |
*** `rationalize' bugfixes and changes |
| 4880 |
|
| 4881 |
Fixed bugs in scm_rationalize `rationalize'. Previously, it returned |
| 4882 |
exact integers unmodified, although that was incorrect if the epsilon |
| 4883 |
was at least 1 or inexact, e.g. (rationalize 4 1) should return 3 per |
| 4884 |
R5RS and R6RS, but previously it returned 4. It also now handles |
| 4885 |
cases involving infinities and NaNs properly, per R6RS. |
| 4886 |
|
| 4887 |
*** Trigonometric functions now return exact numbers in some cases |
| 4888 |
|
| 4889 |
scm_sin `sin', scm_cos `cos', scm_tan `tan', scm_asin `asin', scm_acos |
| 4890 |
`acos', scm_atan `atan', scm_sinh `sinh', scm_cosh `cosh', scm_tanh |
| 4891 |
`tanh', scm_sys_asinh `asinh', scm_sys_acosh `acosh', and |
| 4892 |
scm_sys_atanh `atanh' now return exact results in some cases. |
| 4893 |
|
| 4894 |
*** New procedure: `finite?' |
| 4895 |
|
| 4896 |
Add scm_finite_p `finite?' from R6RS to guile core, which returns #t |
| 4897 |
if and only if its argument is neither infinite nor a NaN. Note that |
| 4898 |
this is not the same as (not (inf? x)) or (not (infinite? x)), since |
| 4899 |
NaNs are neither finite nor infinite. |
| 4900 |
|
| 4901 |
*** Improved exactness handling for complex number parsing |
| 4902 |
|
| 4903 |
When parsing non-real complex numbers, exactness specifiers are now |
| 4904 |
applied to each component, as is done in PLT Scheme. For complex |
| 4905 |
numbers written in rectangular form, exactness specifiers are applied |
| 4906 |
to the real and imaginary parts before calling scm_make_rectangular. |
| 4907 |
For complex numbers written in polar form, exactness specifiers are |
| 4908 |
applied to the magnitude and angle before calling scm_make_polar. |
| 4909 |
|
| 4910 |
Previously, exactness specifiers were applied to the number as a whole |
| 4911 |
_after_ calling scm_make_rectangular or scm_make_polar. |
| 4912 |
|
| 4913 |
For example, (string->number "#i5.0+0i") now does the equivalent of: |
| 4914 |
|
| 4915 |
(make-rectangular (exact->inexact 5.0) (exact->inexact 0)) |
| 4916 |
|
| 4917 |
which yields 5.0+0.0i. Previously it did the equivalent of: |
| 4918 |
|
| 4919 |
(exact->inexact (make-rectangular 5.0 0)) |
| 4920 |
|
| 4921 |
which yielded 5.0. |
| 4922 |
|
| 4923 |
** Unicode characters |
| 4924 |
|
| 4925 |
Unicode characters may be entered in octal format via e.g. `#\454', or |
| 4926 |
created via (integer->char 300). A hex external representation will |
| 4927 |
probably be introduced at some point. |
| 4928 |
|
| 4929 |
** Unicode strings |
| 4930 |
|
| 4931 |
Internally, strings are now represented either in the `latin-1' |
| 4932 |
encoding, one byte per character, or in UTF-32, with four bytes per |
| 4933 |
character. Strings manage their own allocation, switching if needed. |
| 4934 |
|
| 4935 |
Extended characters may be written in a literal string using the |
| 4936 |
hexadecimal escapes `\xXX', `\uXXXX', or `\UXXXXXX', for 8-bit, 16-bit, |
| 4937 |
or 24-bit codepoints, respectively, or entered directly in the native |
| 4938 |
encoding of the port on which the string is read. |
| 4939 |
|
| 4940 |
** Unicode symbols |
| 4941 |
|
| 4942 |
One may now use U+03BB (GREEK SMALL LETTER LAMBDA) as an identifier. |
| 4943 |
|
| 4944 |
** Support for non-ASCII source code files |
| 4945 |
|
| 4946 |
The default reader now handles source code files for some of the |
| 4947 |
non-ASCII character encodings, such as UTF-8. A non-ASCII source file |
| 4948 |
should have an encoding declaration near the top of the file. Also, |
| 4949 |
there is a new function, `file-encoding', that scans a port for a coding |
| 4950 |
declaration. See the section of the manual entitled, "Character Encoding |
| 4951 |
of Source Files". |
| 4952 |
|
| 4953 |
The pre-1.9.3 reader handled 8-bit clean but otherwise unspecified source |
| 4954 |
code. This use is now discouraged. Binary input and output is |
| 4955 |
currently supported by opening ports in the ISO-8859-1 locale. |
| 4956 |
|
| 4957 |
** Source files default to UTF-8. |
| 4958 |
|
| 4959 |
If source files do not specify their encoding via a `coding:' block, |
| 4960 |
the default encoding is UTF-8, instead of being taken from the current |
| 4961 |
locale. |
| 4962 |
|
| 4963 |
** Interactive Guile installs the current locale. |
| 4964 |
|
| 4965 |
Instead of leaving the user in the "C" locale, running the Guile REPL |
| 4966 |
installs the current locale. [FIXME xref?] |
| 4967 |
|
| 4968 |
** Support for locale transcoding when reading from and writing to ports |
| 4969 |
|
| 4970 |
Ports now have an associated character encoding, and port read and write |
| 4971 |
operations do conversion to and from locales automatically. Ports also |
| 4972 |
have an associated strategy for how to deal with locale conversion |
| 4973 |
failures. |
| 4974 |
|
| 4975 |
See the documentation in the manual for the four new support functions, |
| 4976 |
`set-port-encoding!', `port-encoding', `set-port-conversion-strategy!', |
| 4977 |
and `port-conversion-strategy'. |
| 4978 |
|
| 4979 |
** String and SRFI-13 functions can operate on Unicode strings |
| 4980 |
|
| 4981 |
** Unicode support for SRFI-14 character sets |
| 4982 |
|
| 4983 |
The default character sets are no longer locale dependent and contain |
| 4984 |
characters from the whole Unicode range. There is a new predefined |
| 4985 |
character set, `char-set:designated', which contains all assigned |
| 4986 |
Unicode characters. There is a new debugging function, `%char-set-dump'. |
| 4987 |
|
| 4988 |
** Character functions operate on Unicode characters |
| 4989 |
|
| 4990 |
`char-upcase' and `char-downcase' use default Unicode casing rules. |
| 4991 |
Character comparisons such as `char<?' and `char-ci<?' now sort based on |
| 4992 |
Unicode code points. |
| 4993 |
|
| 4994 |
** Global variables `scm_charnames' and `scm_charnums' are removed |
| 4995 |
|
| 4996 |
These variables contained the names of control characters and were |
| 4997 |
used when writing characters. While these were global, they were |
| 4998 |
never intended to be public API. They have been replaced with private |
| 4999 |
functions. |
| 5000 |
|
| 5001 |
** EBCDIC support is removed |
| 5002 |
|
| 5003 |
There was an EBCDIC compile flag that altered some of the character |
| 5004 |
processing. It appeared that full EBCDIC support was never completed |
| 5005 |
and was unmaintained. |
| 5006 |
|
| 5007 |
** Compile-time warnings |
| 5008 |
|
| 5009 |
Guile can warn about potentially unbound free variables. Pass the |
| 5010 |
-Wunbound-variable on the `guile-tools compile' command line, or add |
| 5011 |
`#:warnings '(unbound-variable)' to your `compile' or `compile-file' |
| 5012 |
invocation. Warnings are also enabled by default for expressions entered |
| 5013 |
at the REPL. |
| 5014 |
|
| 5015 |
Guile can also warn when you pass the wrong number of arguments to a |
| 5016 |
procedure, with -Warity-mismatch, or `arity-mismatch' in the |
| 5017 |
`#:warnings' as above. |
| 5018 |
|
| 5019 |
Other warnings include `-Wunused-variable' and `-Wunused-toplevel', to |
| 5020 |
warn about unused local or global (top-level) variables, and `-Wformat', |
| 5021 |
to check for various errors related to the `format' procedure. |
| 5022 |
|
| 5023 |
** A new `memoize-symbol' evaluator trap has been added. |
| 5024 |
|
| 5025 |
This trap can be used for efficiently implementing a Scheme code |
| 5026 |
coverage. |
| 5027 |
|
| 5028 |
** Duplicate bindings among used modules are resolved lazily. |
| 5029 |
|
| 5030 |
This slightly improves program startup times. |
| 5031 |
|
| 5032 |
** New thread cancellation and thread cleanup API |
| 5033 |
|
| 5034 |
See `cancel-thread', `set-thread-cleanup!', and `thread-cleanup'. |
| 5035 |
|
| 5036 |
** New threads are in `(guile-user)' by default, not `(guile)' |
| 5037 |
|
| 5038 |
It used to be that a new thread entering Guile would do so in the |
| 5039 |
`(guile)' module, unless this was the first time Guile was initialized, |
| 5040 |
in which case it was `(guile-user)'. This has been fixed to have all |
| 5041 |
new threads unknown to Guile default to `(guile-user)'. |
| 5042 |
|
| 5043 |
** New helpers: `print-exception', `set-exception-printer!' |
| 5044 |
|
| 5045 |
These functions implement an extensible exception printer. Guile |
| 5046 |
registers printers for all of the exceptions it throws. Users may add |
| 5047 |
their own printers. There is also `scm_print_exception', for use by C |
| 5048 |
programs. Pleasantly, this allows SRFI-35 and R6RS exceptions to be |
| 5049 |
printed appropriately. |
| 5050 |
|
| 5051 |
** GOOPS dispatch in scheme |
| 5052 |
|
| 5053 |
As an implementation detail, GOOPS dispatch is no longer implemented by |
| 5054 |
special evaluator bytecodes, but rather directly via a Scheme function |
| 5055 |
associated with an applicable struct. There is some VM support for the |
| 5056 |
underlying primitives, like `class-of'. |
| 5057 |
|
| 5058 |
This change will in the future allow users to customize generic function |
| 5059 |
dispatch without incurring a performance penalty, and allow us to |
| 5060 |
implement method combinations. |
| 5061 |
|
| 5062 |
** Applicable struct support |
| 5063 |
|
| 5064 |
One may now make structs from Scheme that may be applied as procedures. |
| 5065 |
To do so, make a struct whose vtable is `<applicable-struct-vtable>'. |
| 5066 |
That struct will be the vtable of your applicable structs; instances of |
| 5067 |
that new struct are assumed to have the procedure in their first slot. |
| 5068 |
`<applicable-struct-vtable>' is like Common Lisp's |
| 5069 |
`funcallable-standard-class'. Likewise there is |
| 5070 |
`<applicable-struct-with-setter-vtable>', which looks for the setter in |
| 5071 |
the second slot. This needs to be better documented. |
| 5072 |
|
| 5073 |
** GOOPS cleanups. |
| 5074 |
|
| 5075 |
GOOPS had a number of concepts that were relevant to the days of Tcl, |
| 5076 |
but not any more: operators and entities, mainly. These objects were |
| 5077 |
never documented, and it is unlikely that they were ever used. Operators |
| 5078 |
were a kind of generic specific to the Tcl support. Entities were |
| 5079 |
replaced by applicable structs, mentioned above. |
| 5080 |
|
| 5081 |
** New struct slot allocation: "hidden" |
| 5082 |
|
| 5083 |
A hidden slot is readable and writable, but will not be initialized by a |
| 5084 |
call to make-struct. For example in your layout you would say "ph" |
| 5085 |
instead of "pw". Hidden slots are useful for adding new slots to a |
| 5086 |
vtable without breaking existing invocations to make-struct. |
| 5087 |
|
| 5088 |
** eqv? not a generic |
| 5089 |
|
| 5090 |
One used to be able to extend `eqv?' as a primitive-generic, but no |
| 5091 |
more. Because `eqv?' is in the expansion of `case' (via `memv'), which |
| 5092 |
should be able to compile to static dispatch tables, it doesn't make |
| 5093 |
sense to allow extensions that would subvert this optimization. |
| 5094 |
|
| 5095 |
** `inet-ntop' and `inet-pton' are always available. |
| 5096 |
|
| 5097 |
Guile now use a portable implementation of `inet_pton'/`inet_ntop', so |
| 5098 |
there is no more need to use `inet-aton'/`inet-ntoa'. The latter |
| 5099 |
functions are deprecated. |
| 5100 |
|
| 5101 |
** `getopt-long' parsing errors throw to `quit', not `misc-error' |
| 5102 |
|
| 5103 |
This change should inhibit backtraces on argument parsing errors. |
| 5104 |
`getopt-long' has been modified to print out the error that it throws |
| 5105 |
itself. |
| 5106 |
|
| 5107 |
** New primitive: `tmpfile'. |
| 5108 |
|
| 5109 |
See "File System" in the manual. |
| 5110 |
|
| 5111 |
** Random generator state may be serialized to a datum |
| 5112 |
|
| 5113 |
`random-state->datum' will serialize a random state to a datum, which |
| 5114 |
may be written out, read back in later, and revivified using |
| 5115 |
`datum->random-state'. See "Random" in the manual, for more details. |
| 5116 |
|
| 5117 |
** Fix random number generator on 64-bit platforms |
| 5118 |
|
| 5119 |
There was a nasty bug on 64-bit platforms in which asking for a random |
| 5120 |
integer with a range between 2**32 and 2**64 caused a segfault. After |
| 5121 |
many embarrassing iterations, this was fixed. |
| 5122 |
|
| 5123 |
** Fast bit operations. |
| 5124 |
|
| 5125 |
The bit-twiddling operations `ash', `logand', `logior', and `logxor' now |
| 5126 |
have dedicated bytecodes. Guile is not just for symbolic computation, |
| 5127 |
it's for number crunching too. |
| 5128 |
|
| 5129 |
** Faster SRFI-9 record access |
| 5130 |
|
| 5131 |
SRFI-9 records are now implemented directly on top of Guile's structs, |
| 5132 |
and their accessors are defined in such a way that normal call-sites |
| 5133 |
inline to special VM opcodes, while still allowing for the general case |
| 5134 |
(e.g. passing a record accessor to `apply'). |
| 5135 |
|
| 5136 |
** R6RS block comment support |
| 5137 |
|
| 5138 |
Guile now supports R6RS nested block comments. The start of a comment is |
| 5139 |
marked with `#|', and the end with `|#'. |
| 5140 |
|
| 5141 |
** `guile-2' cond-expand feature |
| 5142 |
|
| 5143 |
To test if your code is running under Guile 2.0 (or its alpha releases), |
| 5144 |
test for the `guile-2' cond-expand feature. Like this: |
| 5145 |
|
| 5146 |
(cond-expand (guile-2 (eval-when (compile) |
| 5147 |
;; This must be evaluated at compile time. |
| 5148 |
(fluid-set! current-reader my-reader))) |
| 5149 |
(guile |
| 5150 |
;; Earlier versions of Guile do not have a |
| 5151 |
;; separate compilation phase. |
| 5152 |
(fluid-set! current-reader my-reader))) |
| 5153 |
|
| 5154 |
** New global variables: %load-compiled-path, %load-compiled-extensions |
| 5155 |
|
| 5156 |
These are analogous to %load-path and %load-extensions. |
| 5157 |
|
| 5158 |
** New fluid: `%file-port-name-canonicalization' |
| 5159 |
|
| 5160 |
This fluid parameterizes the file names that are associated with file |
| 5161 |
ports. If %file-port-name-canonicalization is 'absolute, then file names |
| 5162 |
are canonicalized to be absolute paths. If it is 'relative, then the |
| 5163 |
name is canonicalized, but any prefix corresponding to a member of |
| 5164 |
`%load-path' is stripped off. Otherwise the names are passed through |
| 5165 |
unchanged. |
| 5166 |
|
| 5167 |
In addition, the `compile-file' and `compile-and-load' procedures bind |
| 5168 |
%file-port-name-canonicalization to their `#:canonicalization' keyword |
| 5169 |
argument, which defaults to 'relative. In this way, one might compile |
| 5170 |
"../module/ice-9/boot-9.scm", but the path that gets residualized into |
| 5171 |
the .go is "ice-9/boot-9.scm". |
| 5172 |
|
| 5173 |
** New procedure, `make-promise' |
| 5174 |
|
| 5175 |
`(make-promise (lambda () foo))' is equivalent to `(delay foo)'. |
| 5176 |
|
| 5177 |
** `defined?' may accept a module as its second argument |
| 5178 |
|
| 5179 |
Previously it only accepted internal structures from the evaluator. |
| 5180 |
|
| 5181 |
** New entry into %guile-build-info: `ccachedir' |
| 5182 |
|
| 5183 |
** Fix bug in `module-bound?'. |
| 5184 |
|
| 5185 |
`module-bound?' was returning true if a module did have a local |
| 5186 |
variable, but one that was unbound, but another imported module bound |
| 5187 |
the variable. This was an error, and was fixed. |
| 5188 |
|
| 5189 |
** `(ice-9 syncase)' has been deprecated. |
| 5190 |
|
| 5191 |
As syntax-case is available by default, importing `(ice-9 syncase)' has |
| 5192 |
no effect, and will trigger a deprecation warning. |
| 5193 |
|
| 5194 |
** New readline history functions |
| 5195 |
|
| 5196 |
The (ice-9 readline) module now provides add-history, read-history, |
| 5197 |
write-history and clear-history, which wrap the corresponding GNU |
| 5198 |
History library functions. |
| 5199 |
|
| 5200 |
** Removed deprecated uniform array procedures: |
| 5201 |
dimensions->uniform-array, list->uniform-array, array-prototype |
| 5202 |
|
| 5203 |
Instead, use make-typed-array, list->typed-array, or array-type, |
| 5204 |
respectively. |
| 5205 |
|
| 5206 |
** Deprecate the old `scm-style-repl' |
| 5207 |
|
| 5208 |
The following bindings from boot-9 are now found in `(ice-9 |
| 5209 |
scm-style-repl)': `scm-style-repl', `error-catching-loop', |
| 5210 |
`error-catching-repl', `bad-throw', `scm-repl-silent' |
| 5211 |
`assert-repl-silence', `repl-print-unspecified', |
| 5212 |
`assert-repl-print-unspecified', `scm-repl-verbose', |
| 5213 |
`assert-repl-verbosity', `scm-repl-prompt', `set-repl-prompt!', `repl', |
| 5214 |
`default-pre-unwind-handler', `handle-system-error', |
| 5215 |
|
| 5216 |
The following bindings have been deprecated, with no replacement: |
| 5217 |
`pre-unwind-handler-dispatch'. |
| 5218 |
|
| 5219 |
The following bindings have been totally removed: |
| 5220 |
`before-signal-stack'. |
| 5221 |
|
| 5222 |
Deprecated forwarding shims have been installed so that users that |
| 5223 |
expect these bindings in the main namespace will still work, but receive |
| 5224 |
a deprecation warning. |
| 5225 |
|
| 5226 |
** `set-batch-mode?!' replaced by `ensure-batch-mode!' |
| 5227 |
|
| 5228 |
"Batch mode" is a flag used to tell a program that it is not running |
| 5229 |
interactively. One usually turns it on after a fork. It may not be |
| 5230 |
turned off. `ensure-batch-mode!' deprecates the old `set-batch-mode?!', |
| 5231 |
because it is a better interface, as it can only turn on batch mode, not |
| 5232 |
turn it off. |
| 5233 |
|
| 5234 |
** Deprecate `save-stack', `the-last-stack' |
| 5235 |
|
| 5236 |
It used to be that the way to debug programs in Guile was to capture the |
| 5237 |
stack at the time of error, drop back to the REPL, then debug that |
| 5238 |
stack. But this approach didn't compose, was tricky to get right in the |
| 5239 |
presence of threads, and was not very powerful. |
| 5240 |
|
| 5241 |
So `save-stack', `stack-saved?', and `the-last-stack' have been moved to |
| 5242 |
`(ice-9 save-stack)', with deprecated bindings left in the root module. |
| 5243 |
|
| 5244 |
** `top-repl' has its own module |
| 5245 |
|
| 5246 |
The `top-repl' binding, called with Guile is run interactively, is now |
| 5247 |
is its own module, `(ice-9 top-repl)'. A deprecated forwarding shim was |
| 5248 |
left in the default environment. |
| 5249 |
|
| 5250 |
** `display-error' takes a frame |
| 5251 |
|
| 5252 |
The `display-error' / `scm_display_error' helper now takes a frame as an |
| 5253 |
argument instead of a stack. Stacks are still supported in deprecated |
| 5254 |
builds. Additionally, `display-error' will again source location |
| 5255 |
information for the error. |
| 5256 |
|
| 5257 |
** No more `(ice-9 debug)' |
| 5258 |
|
| 5259 |
This module had some debugging helpers that are no longer applicable to |
| 5260 |
the current debugging model. Importing this module will produce a |
| 5261 |
deprecation warning. Users should contact bug-guile for support. |
| 5262 |
|
| 5263 |
** Remove obsolete debug-options |
| 5264 |
|
| 5265 |
Removed `breakpoints', `trace', `procnames', `indent', `frames', |
| 5266 |
`maxdepth', and `debug' debug-options. |
| 5267 |
|
| 5268 |
** `backtrace' debug option on by default |
| 5269 |
|
| 5270 |
Given that Guile 2.0 can always give you a backtrace, backtraces are now |
| 5271 |
on by default. |
| 5272 |
|
| 5273 |
** `turn-on-debugging' deprecated |
| 5274 |
|
| 5275 |
** Remove obsolete print-options |
| 5276 |
|
| 5277 |
The `source' and `closure-hook' print options are obsolete, and have |
| 5278 |
been removed. |
| 5279 |
|
| 5280 |
** Remove obsolete read-options |
| 5281 |
|
| 5282 |
The "elisp-strings" and "elisp-vectors" read options were unused and |
| 5283 |
obsolete, so they have been removed. |
| 5284 |
|
| 5285 |
** Remove eval-options and trap-options |
| 5286 |
|
| 5287 |
Eval-options and trap-options are obsolete with the new VM and |
| 5288 |
evaluator. |
| 5289 |
|
| 5290 |
** Remove (ice-9 debugger) and (ice-9 debugging) |
| 5291 |
|
| 5292 |
See "Traps" and "Interactive Debugging" in the manual, for information |
| 5293 |
on their replacements. |
| 5294 |
|
| 5295 |
** Remove the GDS Emacs integration |
| 5296 |
|
| 5297 |
See "Using Guile in Emacs" in the manual, for info on how we think you |
| 5298 |
should use Guile with Emacs. |
| 5299 |
|
| 5300 |
** Deprecated: `lazy-catch' |
| 5301 |
|
| 5302 |
`lazy-catch' was a form that captured the stack at the point of a |
| 5303 |
`throw', but the dynamic state at the point of the `catch'. It was a bit |
| 5304 |
crazy. Please change to use `catch', possibly with a throw-handler, or |
| 5305 |
`with-throw-handler'. |
| 5306 |
|
| 5307 |
** Deprecated: primitive properties |
| 5308 |
|
| 5309 |
The `primitive-make-property', `primitive-property-set!', |
| 5310 |
`primitive-property-ref', and `primitive-property-del!' procedures were |
| 5311 |
crufty and only used to implement object properties, which has a new, |
| 5312 |
threadsafe implementation. Use object properties or weak hash tables |
| 5313 |
instead. |
| 5314 |
|
| 5315 |
** Deprecated `@bind' syntax |
| 5316 |
|
| 5317 |
`@bind' was part of an older implementation of the Emacs Lisp language, |
| 5318 |
and is no longer used. |
| 5319 |
|
| 5320 |
** Miscellaneous other deprecations |
| 5321 |
|
| 5322 |
`cuserid' has been deprecated, as it only returns 8 bytes of a user's |
| 5323 |
login. Use `(passwd:name (getpwuid (geteuid)))' instead. |
| 5324 |
|
| 5325 |
Additionally, the procedures `apply-to-args', `has-suffix?', `scheme-file-suffix' |
| 5326 |
`get-option', `for-next-option', `display-usage-report', |
| 5327 |
`transform-usage-lambda', `collect', and `set-batch-mode?!' have all |
| 5328 |
been deprecated. |
| 5329 |
|
| 5330 |
** Add support for unbound fluids |
| 5331 |
|
| 5332 |
See `make-unbound-fluid', `fluid-unset!', and `fluid-bound?' in the |
| 5333 |
manual. |
| 5334 |
|
| 5335 |
** Add `variable-unset!' |
| 5336 |
|
| 5337 |
See "Variables" in the manual, for more details. |
| 5338 |
|
| 5339 |
** Last but not least, the `λ' macro can be used in lieu of `lambda' |
| 5340 |
|
| 5341 |
* Changes to the C interface |
| 5342 |
|
| 5343 |
** Guile now uses libgc, the Boehm-Demers-Weiser garbage collector |
| 5344 |
|
| 5345 |
The semantics of `scm_gc_malloc ()' have been changed, in a |
| 5346 |
backward-compatible way. A new allocation routine, |
| 5347 |
`scm_gc_malloc_pointerless ()', was added. |
| 5348 |
|
| 5349 |
Libgc is a conservative GC, which we hope will make interaction with C |
| 5350 |
code easier and less error-prone. |
| 5351 |
|
| 5352 |
** New procedures: `scm_to_stringn', `scm_from_stringn' |
| 5353 |
** New procedures: scm_{to,from}_{utf8,latin1}_symbol{n,} |
| 5354 |
** New procedures: scm_{to,from}_{utf8,utf32,latin1}_string{n,} |
| 5355 |
|
| 5356 |
These new procedures convert to and from string representations in |
| 5357 |
particular encodings. |
| 5358 |
|
| 5359 |
Users should continue to use locale encoding for user input, user |
| 5360 |
output, or interacting with the C library. |
| 5361 |
|
| 5362 |
Use the Latin-1 functions for ASCII, and for literals in source code. |
| 5363 |
|
| 5364 |
Use UTF-8 functions for interaction with modern libraries which deal in |
| 5365 |
UTF-8, and UTF-32 for interaction with utf32-using libraries. |
| 5366 |
|
| 5367 |
Otherwise, use scm_to_stringn or scm_from_stringn with a specific |
| 5368 |
encoding. |
| 5369 |
|
| 5370 |
** New type definitions for `scm_t_intptr' and friends. |
| 5371 |
|
| 5372 |
`SCM_T_UINTPTR_MAX', `SCM_T_INTPTR_MIN', `SCM_T_INTPTR_MAX', |
| 5373 |
`SIZEOF_SCM_T_BITS', `scm_t_intptr' and `scm_t_uintptr' are now |
| 5374 |
available to C. Have fun! |
| 5375 |
|
| 5376 |
** The GH interface (deprecated in version 1.6, 2001) was removed. |
| 5377 |
|
| 5378 |
** Internal `scm_i_' functions now have "hidden" linkage with GCC/ELF |
| 5379 |
|
| 5380 |
This makes these internal functions technically not callable from |
| 5381 |
application code. |
| 5382 |
|
| 5383 |
** Functions for handling `scm_option' now no longer require an argument |
| 5384 |
indicating length of the `scm_t_option' array. |
| 5385 |
|
| 5386 |
** Procedures-with-setters are now implemented using applicable structs |
| 5387 |
|
| 5388 |
From a user's perspective this doesn't mean very much. But if, for some |
| 5389 |
odd reason, you used the SCM_PROCEDURE_WITH_SETTER_P, SCM_PROCEDURE, or |
| 5390 |
SCM_SETTER macros, know that they're deprecated now. Also, scm_tc7_pws |
| 5391 |
is gone. |
| 5392 |
|
| 5393 |
** Remove old evaluator closures |
| 5394 |
|
| 5395 |
There used to be ranges of typecodes allocated to interpreted data |
| 5396 |
structures, but that it no longer the case, given that interpreted |
| 5397 |
procedure are now just regular VM closures. As a result, there is a |
| 5398 |
newly free tc3, and a number of removed macros. See the ChangeLog for |
| 5399 |
details. |
| 5400 |
|
| 5401 |
** Primitive procedures are now VM trampoline procedures |
| 5402 |
|
| 5403 |
It used to be that there were something like 12 different typecodes |
| 5404 |
allocated to primitive procedures, each with its own calling convention. |
| 5405 |
Now there is only one, the gsubr. This may affect user code if you were |
| 5406 |
defining a procedure using scm_c_make_subr rather scm_c_make_gsubr. The |
| 5407 |
solution is to switch to use scm_c_make_gsubr. This solution works well |
| 5408 |
both with the old 1.8 and with the current 1.9 branch. |
| 5409 |
|
| 5410 |
Guile's old evaluator used to have special cases for applying "gsubrs", |
| 5411 |
primitive procedures with specified numbers of required, optional, and |
| 5412 |
rest arguments. Now, however, Guile represents gsubrs as normal VM |
| 5413 |
procedures, with appropriate bytecode to parse out the correct number of |
| 5414 |
arguments, including optional and rest arguments, and then with a |
| 5415 |
special bytecode to apply the gsubr. |
| 5416 |
|
| 5417 |
This allows primitive procedures to appear on the VM stack, allowing |
| 5418 |
them to be accurately counted in profiles. Also they now have more |
| 5419 |
debugging information attached to them -- their number of arguments, for |
| 5420 |
example. In addition, the VM can completely inline the application |
| 5421 |
mechanics, allowing for faster primitive calls. |
| 5422 |
|
| 5423 |
However there are some changes on the C level. There is no more |
| 5424 |
`scm_tc7_gsubr' or `scm_tcs_subrs' typecode for primitive procedures, as |
| 5425 |
they are just VM procedures. Likewise the macros `SCM_GSUBR_TYPE', |
| 5426 |
`SCM_GSUBR_MAKTYPE', `SCM_GSUBR_REQ', `SCM_GSUBR_OPT', and |
| 5427 |
`SCM_GSUBR_REST' are gone, as are `SCM_SUBR_META_INFO', `SCM_SUBR_PROPS' |
| 5428 |
`SCM_SET_SUBR_GENERIC_LOC', and `SCM_SUBR_ARITY_TO_TYPE'. |
| 5429 |
|
| 5430 |
Perhaps more significantly, `scm_c_make_subr', |
| 5431 |
`scm_c_make_subr_with_generic', `scm_c_define_subr', and |
| 5432 |
`scm_c_define_subr_with_generic'. They all operated on subr typecodes, |
| 5433 |
and there are no more subr typecodes. Use the scm_c_make_gsubr family |
| 5434 |
instead. |
| 5435 |
|
| 5436 |
Normal users of gsubrs should not be affected, though, as the |
| 5437 |
scm_c_make_gsubr family still is the correct way to create primitive |
| 5438 |
procedures. |
| 5439 |
|
| 5440 |
** Remove deprecated array C interfaces |
| 5441 |
|
| 5442 |
Removed the deprecated array functions `scm_i_arrayp', |
| 5443 |
`scm_i_array_ndim', `scm_i_array_mem', `scm_i_array_v', |
| 5444 |
`scm_i_array_base', `scm_i_array_dims', and the deprecated macros |
| 5445 |
`SCM_ARRAYP', `SCM_ARRAY_NDIM', `SCM_ARRAY_CONTP', `SCM_ARRAY_MEM', |
| 5446 |
`SCM_ARRAY_V', `SCM_ARRAY_BASE', and `SCM_ARRAY_DIMS'. |
| 5447 |
|
| 5448 |
** Remove unused snarf macros |
| 5449 |
|
| 5450 |
`SCM_DEFINE1', `SCM_PRIMITIVE_GENERIC_1', `SCM_PROC1, and `SCM_GPROC1' |
| 5451 |
are no more. Use SCM_DEFINE or SCM_PRIMITIVE_GENERIC instead. |
| 5452 |
|
| 5453 |
** New functions: `scm_call_n', `scm_c_run_hookn' |
| 5454 |
|
| 5455 |
`scm_call_n' applies to apply a function to an array of arguments. |
| 5456 |
`scm_c_run_hookn' runs a hook with an array of arguments. |
| 5457 |
|
| 5458 |
** Some SMOB types changed to have static typecodes |
| 5459 |
|
| 5460 |
Fluids, dynamic states, and hash tables used to be SMOB objects, but now |
| 5461 |
they have statically allocated tc7 typecodes. |
| 5462 |
|
| 5463 |
** Preparations for changing SMOB representation |
| 5464 |
|
| 5465 |
If things go right, we'll be changing the SMOB representation soon. To |
| 5466 |
that end, we did a lot of cleanups to calls to e.g. SCM_CELL_WORD_2(x) when |
| 5467 |
the code meant SCM_SMOB_DATA_2(x); user code will need similar changes |
| 5468 |
in the future. Code accessing SMOBs using SCM_CELL macros was never |
| 5469 |
correct, but until now things still worked. Users should be aware of |
| 5470 |
such changes. |
| 5471 |
|
| 5472 |
** Changed invocation mechanics of applicable SMOBs |
| 5473 |
|
| 5474 |
Guile's old evaluator used to have special cases for applying SMOB |
| 5475 |
objects. Now, with the VM, when Guile sees a SMOB, it looks up a VM |
| 5476 |
trampoline procedure for it, and use the normal mechanics to apply the |
| 5477 |
trampoline. This simplifies procedure application in the normal, |
| 5478 |
non-SMOB case. |
| 5479 |
|
| 5480 |
The upshot is that the mechanics used to apply a SMOB are different from |
| 5481 |
1.8. Descriptors no longer have `apply_0', `apply_1', `apply_2', and |
| 5482 |
`apply_3' functions, and the macros SCM_SMOB_APPLY_0 and friends are now |
| 5483 |
deprecated. Just use the scm_call_0 family of procedures. |
| 5484 |
|
| 5485 |
** Removed support shlibs for SRFIs 1, 4, 13, 14, and 60 |
| 5486 |
|
| 5487 |
Though these SRFI support libraries did expose API, they encoded a |
| 5488 |
strange version string into their library names. That version was never |
| 5489 |
programmatically exported, so there was no way people could use the |
| 5490 |
libs. |
| 5491 |
|
| 5492 |
This was a fortunate oversight, as it allows us to remove the need for |
| 5493 |
extra, needless shared libraries --- the C support code for SRFIs 4, 13, |
| 5494 |
and 14 was already in core --- and allow us to incrementally return the |
| 5495 |
SRFI implementation to Scheme. |
| 5496 |
|
| 5497 |
** New C function: scm_module_public_interface |
| 5498 |
|
| 5499 |
This procedure corresponds to Scheme's `module-public-interface'. |
| 5500 |
|
| 5501 |
** Undeprecate `scm_the_root_module ()' |
| 5502 |
|
| 5503 |
It's useful to be able to get the root module from C without doing a |
| 5504 |
full module lookup. |
| 5505 |
|
| 5506 |
** Inline vector allocation |
| 5507 |
|
| 5508 |
Instead of having vectors point out into the heap for their data, their |
| 5509 |
data is now allocated inline to the vector object itself. The same is |
| 5510 |
true for bytevectors, by default, though there is an indirection |
| 5511 |
available which should allow for making a bytevector from an existing |
| 5512 |
memory region. |
| 5513 |
|
| 5514 |
** New struct constructors that don't involve making lists |
| 5515 |
|
| 5516 |
`scm_c_make_struct' and `scm_c_make_structv' are new varargs and array |
| 5517 |
constructors, respectively, for structs. You might find them useful. |
| 5518 |
|
| 5519 |
** Stack refactor |
| 5520 |
|
| 5521 |
In Guile 1.8, there were debugging frames on the C stack. Now there is |
| 5522 |
no more need to explicitly mark the stack in this way, because Guile has |
| 5523 |
a VM stack that it knows how to walk, which simplifies the C API |
| 5524 |
considerably. See the ChangeLog for details; the relevant interface is |
| 5525 |
in libguile/stacks.h. The Scheme API has not been changed significantly. |
| 5526 |
|
| 5527 |
** Removal of Guile's primitive object system. |
| 5528 |
|
| 5529 |
There were a number of pieces in `objects.[ch]' that tried to be a |
| 5530 |
minimal object system, but were never documented, and were quickly |
| 5531 |
obseleted by GOOPS' merge into Guile proper. So `scm_make_class_object', |
| 5532 |
`scm_make_subclass_object', `scm_metaclass_standard', and like symbols |
| 5533 |
from objects.h are no more. In the very unlikely case in which these |
| 5534 |
were useful to you, we urge you to contact guile-devel. |
| 5535 |
|
| 5536 |
** No future. |
| 5537 |
|
| 5538 |
Actually the future is still in the state that it was, is, and ever |
| 5539 |
shall be, Amen, except that `futures.c' and `futures.h' are no longer a |
| 5540 |
part of it. These files were experimental, never compiled, and would be |
| 5541 |
better implemented in Scheme anyway. In the future, that is. |
| 5542 |
|
| 5543 |
** Deprecate trampolines |
| 5544 |
|
| 5545 |
There used to be C functions `scm_trampoline_0', `scm_trampoline_1', and |
| 5546 |
so on. The point was to do some precomputation on the type of the |
| 5547 |
procedure, then return a specialized "call" procedure. However this |
| 5548 |
optimization wasn't actually an optimization, so it is now deprecated. |
| 5549 |
Just use `scm_call_0', etc instead. |
| 5550 |
|
| 5551 |
** Deprecated `scm_badargsp' |
| 5552 |
|
| 5553 |
This function is unused in Guile, but was part of its API. |
| 5554 |
|
| 5555 |
** Better support for Lisp `nil'. |
| 5556 |
|
| 5557 |
The bit representation of `nil' has been tweaked so that it is now very |
| 5558 |
efficient to check e.g. if a value is equal to Scheme's end-of-list or |
| 5559 |
Lisp's nil. Additionally there are a heap of new, specific predicates |
| 5560 |
like scm_is_null_or_nil. |
| 5561 |
|
| 5562 |
** Better integration of Lisp `nil'. |
| 5563 |
|
| 5564 |
`scm_is_boolean', `scm_is_false', and `scm_is_null' all return true now |
| 5565 |
for Lisp's `nil'. This shouldn't affect any Scheme code at this point, |
| 5566 |
but when we start to integrate more with Emacs, it is possible that we |
| 5567 |
break code that assumes that, for example, `(not x)' implies that `x' is |
| 5568 |
`eq?' to `#f'. This is not a common assumption. Refactoring affected |
| 5569 |
code to rely on properties instead of identities will improve code |
| 5570 |
correctness. See "Nil" in the manual, for more details. |
| 5571 |
|
| 5572 |
** Support for static allocation of strings, symbols, and subrs. |
| 5573 |
|
| 5574 |
Calls to snarfing CPP macros like SCM_DEFINE macro will now allocate |
| 5575 |
much of their associated data as static variables, reducing Guile's |
| 5576 |
memory footprint. |
| 5577 |
|
| 5578 |
** `scm_stat' has an additional argument, `exception_on_error' |
| 5579 |
** `scm_primitive_load_path' has an additional argument `exception_on_not_found' |
| 5580 |
|
| 5581 |
** `scm_set_port_seek' and `scm_set_port_truncate' use the `scm_t_off' type |
| 5582 |
|
| 5583 |
Previously they would use the `off_t' type, which is fragile since its |
| 5584 |
definition depends on the application's value for `_FILE_OFFSET_BITS'. |
| 5585 |
|
| 5586 |
** The `long_long' C type, deprecated in 1.8, has been removed |
| 5587 |
|
| 5588 |
** Removed deprecated uniform array procedures: scm_make_uve, |
| 5589 |
scm_array_prototype, scm_list_to_uniform_array, |
| 5590 |
scm_dimensions_to_uniform_array, scm_make_ra, scm_shap2ra, scm_cvref, |
| 5591 |
scm_ra_set_contp, scm_aind, scm_raprin1 |
| 5592 |
|
| 5593 |
These functions have been deprecated since early 2005. |
| 5594 |
|
| 5595 |
* Changes to the distribution |
| 5596 |
|
| 5597 |
** Guile's license is now LGPLv3+ |
| 5598 |
|
| 5599 |
In other words the GNU Lesser General Public License, version 3 or |
| 5600 |
later (at the discretion of each person that chooses to redistribute |
| 5601 |
part of Guile). |
| 5602 |
|
| 5603 |
** AM_SILENT_RULES |
| 5604 |
|
| 5605 |
Guile's build is visually quieter, due to the use of Automake 1.11's |
| 5606 |
AM_SILENT_RULES. Build as `make V=1' to see all of the output. |
| 5607 |
|
| 5608 |
** GOOPS documentation folded into Guile reference manual |
| 5609 |
|
| 5610 |
GOOPS, Guile's object system, used to be documented in separate manuals. |
| 5611 |
This content is now included in Guile's manual directly. |
| 5612 |
|
| 5613 |
** `guile-config' will be deprecated in favor of `pkg-config' |
| 5614 |
|
| 5615 |
`guile-config' has been rewritten to get its information from |
| 5616 |
`pkg-config', so this should be a transparent change. Note however that |
| 5617 |
guile.m4 has yet to be modified to call pkg-config instead of |
| 5618 |
guile-config. |
| 5619 |
|
| 5620 |
** Guile now provides `guile-2.0.pc' instead of `guile-1.8.pc' |
| 5621 |
|
| 5622 |
Programs that use `pkg-config' to find Guile or one of its Autoconf |
| 5623 |
macros should now require `guile-2.0' instead of `guile-1.8'. |
| 5624 |
|
| 5625 |
** New installation directory: $(pkglibdir)/1.9/ccache |
| 5626 |
|
| 5627 |
If $(libdir) is /usr/lib, for example, Guile will install its .go files |
| 5628 |
to /usr/lib/guile/1.9/ccache. These files are architecture-specific. |
| 5629 |
|
| 5630 |
** Parallel installability fixes |
| 5631 |
|
| 5632 |
Guile now installs its header files to a effective-version-specific |
| 5633 |
directory, and includes the effective version (e.g. 2.0) in the library |
| 5634 |
name (e.g. libguile-2.0.so). |
| 5635 |
|
| 5636 |
This change should be transparent to users, who should detect Guile via |
| 5637 |
the guile.m4 macro, or the guile-2.0.pc pkg-config file. It will allow |
| 5638 |
parallel installs for multiple versions of Guile development |
| 5639 |
environments. |
| 5640 |
|
| 5641 |
** Dynamically loadable extensions may be placed in a Guile-specific path |
| 5642 |
|
| 5643 |
Before, Guile only searched the system library paths for extensions |
| 5644 |
(e.g. /usr/lib), which meant that the names of Guile extensions had to |
| 5645 |
be globally unique. Installing them to a Guile-specific extensions |
| 5646 |
directory is cleaner. Use `pkg-config --variable=extensiondir |
| 5647 |
guile-2.0' to get the location of the extensions directory. |
| 5648 |
|
| 5649 |
** User Scheme code may be placed in a version-specific path |
| 5650 |
|
| 5651 |
Before, there was only one way to install user Scheme code to a |
| 5652 |
version-specific Guile directory: install to Guile's own path, |
| 5653 |
e.g. /usr/share/guile/2.0. The site directory, |
| 5654 |
e.g. /usr/share/guile/site, was unversioned. This has been changed to |
| 5655 |
add a version-specific site directory, e.g. /usr/share/guile/site/2.0, |
| 5656 |
searched before the global site directory. |
| 5657 |
|
| 5658 |
** New dependency: libgc |
| 5659 |
|
| 5660 |
See http://www.hpl.hp.com/personal/Hans_Boehm/gc/, for more information. |
| 5661 |
|
| 5662 |
** New dependency: GNU libunistring |
| 5663 |
|
| 5664 |
See http://www.gnu.org/software/libunistring/, for more information. Our |
| 5665 |
Unicode support uses routines from libunistring. |
| 5666 |
|
| 5667 |
** New dependency: libffi |
| 5668 |
|
| 5669 |
See http://sourceware.org/libffi/, for more information. |
| 5670 |
|
| 5671 |
|
| 5672 |
|
| 5673 |
Changes in 1.8.8 (since 1.8.7) |
| 5674 |
|
| 5675 |
* Bugs fixed |
| 5676 |
|
| 5677 |
** Fix possible buffer overruns when parsing numbers |
| 5678 |
** Avoid clash with system setjmp/longjmp on IA64 |
| 5679 |
** Fix `wrong type arg' exceptions with IPv6 addresses |
| 5680 |
|
| 5681 |
|
| 5682 |
Changes in 1.8.7 (since 1.8.6) |
| 5683 |
|
| 5684 |
* New modules (see the manual for details) |
| 5685 |
|
| 5686 |
** `(srfi srfi-98)', an interface to access environment variables |
| 5687 |
|
| 5688 |
* Bugs fixed |
| 5689 |
|
| 5690 |
** Fix compilation with `--disable-deprecated' |
| 5691 |
** Fix %fast-slot-ref/set!, to avoid possible segmentation fault |
| 5692 |
** Fix MinGW build problem caused by HAVE_STRUCT_TIMESPEC confusion |
| 5693 |
** Fix build problem when scm_t_timespec is different from struct timespec |
| 5694 |
** Fix build when compiled with -Wundef -Werror |
| 5695 |
** More build fixes for `alphaev56-dec-osf5.1b' (Tru64) |
| 5696 |
** Build fixes for `powerpc-ibm-aix5.3.0.0' (AIX 5.3) |
| 5697 |
** With GCC, always compile with `-mieee' on `alpha*' and `sh*' |
| 5698 |
** Better diagnose broken `(strftime "%z" ...)' in `time.test' (bug #24130) |
| 5699 |
** Fix parsing of SRFI-88/postfix keywords longer than 128 characters |
| 5700 |
** Fix reading of complex numbers where both parts are inexact decimals |
| 5701 |
|
| 5702 |
** Allow @ macro to work with (ice-9 syncase) |
| 5703 |
|
| 5704 |
Previously, use of the @ macro in a module whose code is being |
| 5705 |
transformed by (ice-9 syncase) would cause an "Invalid syntax" error. |
| 5706 |
Now it works as you would expect (giving the value of the specified |
| 5707 |
module binding). |
| 5708 |
|
| 5709 |
** Have `scm_take_locale_symbol ()' return an interned symbol (bug #25865) |
| 5710 |
|
| 5711 |
|
| 5712 |
Changes in 1.8.6 (since 1.8.5) |
| 5713 |
|
| 5714 |
* New features (see the manual for details) |
| 5715 |
|
| 5716 |
** New convenience function `scm_c_symbol_length ()' |
| 5717 |
|
| 5718 |
** Single stepping through code from Emacs |
| 5719 |
|
| 5720 |
When you use GDS to evaluate Scheme code from Emacs, you can now use |
| 5721 |
`C-u' to indicate that you want to single step through that code. See |
| 5722 |
`Evaluating Scheme Code' in the manual for more details. |
| 5723 |
|
| 5724 |
** New "guile(1)" man page! |
| 5725 |
|
| 5726 |
* Changes to the distribution |
| 5727 |
|
| 5728 |
** Automake's `AM_MAINTAINER_MODE' is no longer used |
| 5729 |
|
| 5730 |
Thus, the `--enable-maintainer-mode' configure option is no longer |
| 5731 |
available: Guile is now always configured in "maintainer mode". |
| 5732 |
|
| 5733 |
** `ChangeLog' files are no longer updated |
| 5734 |
|
| 5735 |
Instead, changes are detailed in the version control system's logs. See |
| 5736 |
the top-level `ChangeLog' files for details. |
| 5737 |
|
| 5738 |
|
| 5739 |
* Bugs fixed |
| 5740 |
|
| 5741 |
** `symbol->string' now returns a read-only string, as per R5RS |
| 5742 |
** Fix incorrect handling of the FLAGS argument of `fold-matches' |
| 5743 |
** `guile-config link' now prints `-L$libdir' before `-lguile' |
| 5744 |
** Fix memory corruption involving GOOPS' `class-redefinition' |
| 5745 |
** Fix possible deadlock in `mutex-lock' |
| 5746 |
** Fix build issue on Tru64 and ia64-hp-hpux11.23 (`SCM_UNPACK' macro) |
| 5747 |
** Fix build issue on mips, mipsel, powerpc and ia64 (stack direction) |
| 5748 |
** Fix build issue on hppa2.0w-hp-hpux11.11 (`dirent64' and `readdir64_r') |
| 5749 |
** Fix build issue on i386-unknown-freebsd7.0 ("break strict-aliasing rules") |
| 5750 |
** Fix misleading output from `(help rationalize)' |
| 5751 |
** Fix build failure on Debian hppa architecture (bad stack growth detection) |
| 5752 |
** Fix `gcd' when called with a single, negative argument. |
| 5753 |
** Fix `Stack overflow' errors seen when building on some platforms |
| 5754 |
** Fix bug when `scm_with_guile ()' was called several times from the |
| 5755 |
same thread |
| 5756 |
** The handler of SRFI-34 `with-exception-handler' is now invoked in the |
| 5757 |
dynamic environment of the call to `raise' |
| 5758 |
** Fix potential deadlock in `make-struct' |
| 5759 |
** Fix compilation problem with libltdl from Libtool 2.2.x |
| 5760 |
** Fix sloppy bound checking in `string-{ref,set!}' with the empty string |
| 5761 |
|
| 5762 |
|
| 5763 |
Changes in 1.8.5 (since 1.8.4) |
| 5764 |
|
| 5765 |
* Infrastructure changes |
| 5766 |
|
| 5767 |
** Guile repository switched from CVS to Git |
| 5768 |
|
| 5769 |
The new repository can be accessed using |
| 5770 |
"git-clone git://git.sv.gnu.org/guile.git", or can be browsed on-line at |
| 5771 |
http://git.sv.gnu.org/gitweb/?p=guile.git . See `README' for details. |
| 5772 |
|
| 5773 |
** Add support for `pkg-config' |
| 5774 |
|
| 5775 |
See "Autoconf Support" in the manual for details. |
| 5776 |
|
| 5777 |
* New modules (see the manual for details) |
| 5778 |
|
| 5779 |
** `(srfi srfi-88)' |
| 5780 |
|
| 5781 |
* New features (see the manual for details) |
| 5782 |
|
| 5783 |
** New `postfix' read option, for SRFI-88 keyword syntax |
| 5784 |
** Some I/O primitives have been inlined, which improves I/O performance |
| 5785 |
** New object-based traps infrastructure |
| 5786 |
|
| 5787 |
This is a GOOPS-based infrastructure that builds on Guile's low-level |
| 5788 |
evaluator trap calls and facilitates the development of debugging |
| 5789 |
features like single-stepping, breakpoints, tracing and profiling. |
| 5790 |
See the `Traps' node of the manual for details. |
| 5791 |
|
| 5792 |
** New support for working on Guile code from within Emacs |
| 5793 |
|
| 5794 |
Guile now incorporates the `GDS' library (previously distributed |
| 5795 |
separately) for working on Guile code from within Emacs. See the |
| 5796 |
`Using Guile In Emacs' node of the manual for details. |
| 5797 |
|
| 5798 |
* Bugs fixed |
| 5799 |
|
| 5800 |
** `scm_add_slot ()' no longer segfaults (fixes bug #22369) |
| 5801 |
** Fixed `(ice-9 match)' for patterns like `((_ ...) ...)' |
| 5802 |
|
| 5803 |
Previously, expressions like `(match '((foo) (bar)) (((_ ...) ...) #t))' |
| 5804 |
would trigger an unbound variable error for `match:andmap'. |
| 5805 |
|
| 5806 |
** `(oop goops describe)' now properly provides the `describe' feature |
| 5807 |
** Fixed `args-fold' from `(srfi srfi-37)' |
| 5808 |
|
| 5809 |
Previously, parsing short option names of argument-less options would |
| 5810 |
lead to a stack overflow. |
| 5811 |
|
| 5812 |
** `(srfi srfi-35)' is now visible through `cond-expand' |
| 5813 |
** Fixed type-checking for the second argument of `eval' |
| 5814 |
** Fixed type-checking for SRFI-1 `partition' |
| 5815 |
** Fixed `struct-ref' and `struct-set!' on "light structs" |
| 5816 |
** Honor struct field access rights in GOOPS |
| 5817 |
** Changed the storage strategy of source properties, which fixes a deadlock |
| 5818 |
** Allow compilation of Guile-using programs in C99 mode with GCC 4.3 and later |
| 5819 |
** Fixed build issue for GNU/Linux on IA64 |
| 5820 |
** Fixed build issues on NetBSD 1.6 |
| 5821 |
** Fixed build issue on Solaris 2.10 x86_64 |
| 5822 |
** Fixed build issue with DEC/Compaq/HP's compiler |
| 5823 |
** Fixed `scm_from_complex_double' build issue on FreeBSD |
| 5824 |
** Fixed `alloca' build issue on FreeBSD 6 |
| 5825 |
** Removed use of non-portable makefile constructs |
| 5826 |
** Fixed shadowing of libc's <random.h> on Tru64, which broke compilation |
| 5827 |
** Make sure all tests honor `$TMPDIR' |
| 5828 |
|
| 5829 |
|
| 5830 |
Changes in 1.8.4 (since 1.8.3) |
| 5831 |
|
| 5832 |
* Bugs fixed |
| 5833 |
|
| 5834 |
** CR (ASCII 0x0d) is (again) recognized as a token delimiter by the reader |
| 5835 |
** Fixed a segmentation fault which occurred when displaying the |
| 5836 |
backtrace of a stack with a promise object (made by `delay') in it. |
| 5837 |
** Make `accept' leave guile mode while blocking |
| 5838 |
** `scm_c_read ()' and `scm_c_write ()' now type-check their port argument |
| 5839 |
** Fixed a build problem on AIX (use of func_data identifier) |
| 5840 |
** Fixed a segmentation fault which occurred when hashx-ref or hashx-set! was |
| 5841 |
called with an associator proc that returns neither a pair nor #f. |
| 5842 |
** Secondary threads now always return a valid module for (current-module). |
| 5843 |
** Avoid MacOS build problems caused by incorrect combination of "64" |
| 5844 |
system and library calls. |
| 5845 |
** `guile-snarf' now honors `$TMPDIR' |
| 5846 |
** `guile-config compile' now reports CPPFLAGS used at compile-time |
| 5847 |
** Fixed build with Sun Studio (Solaris 9) |
| 5848 |
** Fixed wrong-type-arg errors when creating zero length SRFI-4 |
| 5849 |
uniform vectors on AIX. |
| 5850 |
** Fixed a deadlock that occurs upon GC with multiple threads. |
| 5851 |
** Fixed compile problem with GCC on Solaris and AIX (use of _Complex_I) |
| 5852 |
** Fixed autotool-derived build problems on AIX 6.1. |
| 5853 |
** Fixed NetBSD/alpha support |
| 5854 |
** Fixed MacOS build problem caused by use of rl_get_keymap(_name) |
| 5855 |
|
| 5856 |
* New modules (see the manual for details) |
| 5857 |
|
| 5858 |
** `(srfi srfi-69)' |
| 5859 |
|
| 5860 |
* Documentation fixes and improvements |
| 5861 |
|
| 5862 |
** Removed premature breakpoint documentation |
| 5863 |
|
| 5864 |
The features described are not available in the series of 1.8.x |
| 5865 |
releases, so the documentation was misleading and has been removed. |
| 5866 |
|
| 5867 |
** More about Guile's default *random-state* variable |
| 5868 |
|
| 5869 |
** GOOPS: more about how to use `next-method' |
| 5870 |
|
| 5871 |
* Changes to the distribution |
| 5872 |
|
| 5873 |
** Corrected a few files that referred incorrectly to the old GPL + special exception licence |
| 5874 |
|
| 5875 |
In fact Guile since 1.8.0 has been licensed with the GNU Lesser |
| 5876 |
General Public License, and the few incorrect files have now been |
| 5877 |
fixed to agree with the rest of the Guile distribution. |
| 5878 |
|
| 5879 |
** Removed unnecessary extra copies of COPYING* |
| 5880 |
|
| 5881 |
The distribution now contains a single COPYING.LESSER at its top level. |
| 5882 |
|
| 5883 |
|
| 5884 |
Changes in 1.8.3 (since 1.8.2) |
| 5885 |
|
| 5886 |
* New modules (see the manual for details) |
| 5887 |
|
| 5888 |
** `(srfi srfi-35)' |
| 5889 |
** `(srfi srfi-37)' |
| 5890 |
|
| 5891 |
* Bugs fixed |
| 5892 |
|
| 5893 |
** The `(ice-9 slib)' module now works as expected |
| 5894 |
** Expressions like "(set! 'x #t)" no longer yield a crash |
| 5895 |
** Warnings about duplicate bindings now go to stderr |
| 5896 |
** A memory leak in `make-socket-address' was fixed |
| 5897 |
** Alignment issues (e.g., on SPARC) in network routines were fixed |
| 5898 |
** A threading issue that showed up at least on NetBSD was fixed |
| 5899 |
** Build problems on Solaris and IRIX fixed |
| 5900 |
|
| 5901 |
* Implementation improvements |
| 5902 |
|
| 5903 |
** The reader is now faster, which reduces startup time |
| 5904 |
** Procedures returned by `record-accessor' and `record-modifier' are faster |
| 5905 |
|
| 5906 |
|
| 5907 |
Changes in 1.8.2 (since 1.8.1): |
| 5908 |
|
| 5909 |
* New procedures (see the manual for details) |
| 5910 |
|
| 5911 |
** set-program-arguments |
| 5912 |
** make-vtable |
| 5913 |
|
| 5914 |
* Incompatible changes |
| 5915 |
|
| 5916 |
** The body of a top-level `define' no longer sees the binding being created |
| 5917 |
|
| 5918 |
In a top-level `define', the binding being created is no longer visible |
| 5919 |
from the `define' body. This breaks code like |
| 5920 |
"(define foo (begin (set! foo 1) (+ foo 1)))", where `foo' is now |
| 5921 |
unbound in the body. However, such code was not R5RS-compliant anyway, |
| 5922 |
per Section 5.2.1. |
| 5923 |
|
| 5924 |
* Bugs fixed |
| 5925 |
|
| 5926 |
** Fractions were not `equal?' if stored in unreduced form. |
| 5927 |
(A subtle problem, since printing a value reduced it, making it work.) |
| 5928 |
** srfi-60 `copy-bit' failed on 64-bit systems |
| 5929 |
** "guile --use-srfi" option at the REPL can replace core functions |
| 5930 |
(Programs run with that option were ok, but in the interactive REPL |
| 5931 |
the core bindings got priority, preventing SRFI replacements or |
| 5932 |
extensions.) |
| 5933 |
** `regexp-exec' doesn't abort() on #\nul in the input or bad flags arg |
| 5934 |
** `kill' on mingw throws an error for a PID other than oneself |
| 5935 |
** Procedure names are attached to procedure-with-setters |
| 5936 |
** Array read syntax works with negative lower bound |
| 5937 |
** `array-in-bounds?' fix if an array has different lower bounds on each index |
| 5938 |
** `*' returns exact 0 for "(* inexact 0)" |
| 5939 |
This follows what it always did for "(* 0 inexact)". |
| 5940 |
** SRFI-19: Value returned by `(current-time time-process)' was incorrect |
| 5941 |
** SRFI-19: `date->julian-day' did not account for timezone offset |
| 5942 |
** `ttyname' no longer crashes when passed a non-tty argument |
| 5943 |
** `inet-ntop' no longer crashes on SPARC when passed an `AF_INET' address |
| 5944 |
** Small memory leaks have been fixed in `make-fluid' and `add-history' |
| 5945 |
** GOOPS: Fixed a bug in `method-more-specific?' |
| 5946 |
** Build problems on Solaris fixed |
| 5947 |
** Build problems on HP-UX IA64 fixed |
| 5948 |
** Build problems on MinGW fixed |
| 5949 |
|
| 5950 |
|
| 5951 |
Changes in 1.8.1 (since 1.8.0): |
| 5952 |
|
| 5953 |
* LFS functions are now used to access 64-bit files on 32-bit systems. |
| 5954 |
|
| 5955 |
* New procedures (see the manual for details) |
| 5956 |
|
| 5957 |
** primitive-_exit - [Scheme] the-root-module |
| 5958 |
** scm_primitive__exit - [C] |
| 5959 |
** make-completion-function - [Scheme] (ice-9 readline) |
| 5960 |
** scm_c_locale_stringn_to_number - [C] |
| 5961 |
** scm_srfi1_append_reverse [C] |
| 5962 |
** scm_srfi1_append_reverse_x [C] |
| 5963 |
** scm_log - [C] |
| 5964 |
** scm_log10 - [C] |
| 5965 |
** scm_exp - [C] |
| 5966 |
** scm_sqrt - [C] |
| 5967 |
|
| 5968 |
* Bugs fixed |
| 5969 |
|
| 5970 |
** Build problems have been fixed on MacOS, SunOS, and QNX. |
| 5971 |
|
| 5972 |
** `strftime' fix sign of %z timezone offset. |
| 5973 |
|
| 5974 |
** A one-dimensional array can now be 'equal?' to a vector. |
| 5975 |
|
| 5976 |
** Structures, records, and SRFI-9 records can now be compared with `equal?'. |
| 5977 |
|
| 5978 |
** SRFI-14 standard char sets are recomputed upon a successful `setlocale'. |
| 5979 |
|
| 5980 |
** `record-accessor' and `record-modifier' now have strict type checks. |
| 5981 |
|
| 5982 |
Record accessor and modifier procedures now throw an error if the |
| 5983 |
record type of the record they're given is not the type expected. |
| 5984 |
(Previously accessors returned #f and modifiers silently did nothing). |
| 5985 |
|
| 5986 |
** It is now OK to use both autoload and use-modules on a given module. |
| 5987 |
|
| 5988 |
** `apply' checks the number of arguments more carefully on "0 or 1" funcs. |
| 5989 |
|
| 5990 |
Previously there was no checking on primatives like make-vector that |
| 5991 |
accept "one or two" arguments. Now there is. |
| 5992 |
|
| 5993 |
** The srfi-1 assoc function now calls its equality predicate properly. |
| 5994 |
|
| 5995 |
Previously srfi-1 assoc would call the equality predicate with the key |
| 5996 |
last. According to the SRFI, the key should be first. |
| 5997 |
|
| 5998 |
** A bug in n-par-for-each and n-for-each-par-map has been fixed. |
| 5999 |
|
| 6000 |
** The array-set! procedure no longer segfaults when given a bit vector. |
| 6001 |
|
| 6002 |
** Bugs in make-shared-array have been fixed. |
| 6003 |
|
| 6004 |
** string<? and friends now follow char<? etc order on 8-bit chars. |
| 6005 |
|
| 6006 |
** The format procedure now handles inf and nan values for ~f correctly. |
| 6007 |
|
| 6008 |
** exact->inexact should no longer overflow when given certain large fractions. |
| 6009 |
|
| 6010 |
** srfi-9 accessor and modifier procedures now have strict record type checks. |
| 6011 |
|
| 6012 |
This matches the srfi-9 specification. |
| 6013 |
|
| 6014 |
** (ice-9 ftw) procedures won't ignore different files with same inode number. |
| 6015 |
|
| 6016 |
Previously the (ice-9 ftw) procedures would ignore any file that had |
| 6017 |
the same inode number as a file they had already seen, even if that |
| 6018 |
file was on a different device. |
| 6019 |
|
| 6020 |
|
| 6021 |
Changes in 1.8.0 (changes since the 1.6.x series): |
| 6022 |
|
| 6023 |
* Changes to the distribution |
| 6024 |
|
| 6025 |
** Guile is now licensed with the GNU Lesser General Public License. |
| 6026 |
|
| 6027 |
** The manual is now licensed with the GNU Free Documentation License. |
| 6028 |
|
| 6029 |
** Guile now requires GNU MP (http://swox.com/gmp). |
| 6030 |
|
| 6031 |
Guile now uses the GNU MP library for arbitrary precision arithmetic. |
| 6032 |
|
| 6033 |
** Guile now has separate private and public configuration headers. |
| 6034 |
|
| 6035 |
That is, things like HAVE_STRING_H no longer leak from Guile's |
| 6036 |
headers. |
| 6037 |
|
| 6038 |
** Guile now provides and uses an "effective" version number. |
| 6039 |
|
| 6040 |
Guile now provides scm_effective_version and effective-version |
| 6041 |
functions which return the "effective" version number. This is just |
| 6042 |
the normal full version string without the final micro-version number, |
| 6043 |
so the current effective-version is "1.8". The effective version |
| 6044 |
should remain unchanged during a stable series, and should be used for |
| 6045 |
items like the versioned share directory name |
| 6046 |
i.e. /usr/share/guile/1.8. |
| 6047 |
|
| 6048 |
Providing an unchanging version number during a stable release for |
| 6049 |
things like the versioned share directory can be particularly |
| 6050 |
important for Guile "add-on" packages, since it provides a directory |
| 6051 |
that they can install to that won't be changed out from under them |
| 6052 |
with each micro release during a stable series. |
| 6053 |
|
| 6054 |
** Thread implementation has changed. |
| 6055 |
|
| 6056 |
When you configure "--with-threads=null", you will get the usual |
| 6057 |
threading API (call-with-new-thread, make-mutex, etc), but you can't |
| 6058 |
actually create new threads. Also, "--with-threads=no" is now |
| 6059 |
equivalent to "--with-threads=null". This means that the thread API |
| 6060 |
is always present, although you might not be able to create new |
| 6061 |
threads. |
| 6062 |
|
| 6063 |
When you configure "--with-threads=pthreads" or "--with-threads=yes", |
| 6064 |
you will get threads that are implemented with the portable POSIX |
| 6065 |
threads. These threads can run concurrently (unlike the previous |
| 6066 |
"coop" thread implementation), but need to cooperate for things like |
| 6067 |
the GC. |
| 6068 |
|
| 6069 |
The default is "pthreads", unless your platform doesn't have pthreads, |
| 6070 |
in which case "null" threads are used. |
| 6071 |
|
| 6072 |
See the manual for details, nodes "Initialization", "Multi-Threading", |
| 6073 |
"Blocking", and others. |
| 6074 |
|
| 6075 |
** There is the new notion of 'discouraged' features. |
| 6076 |
|
| 6077 |
This is a milder form of deprecation. |
| 6078 |
|
| 6079 |
Things that are discouraged should not be used in new code, but it is |
| 6080 |
OK to leave them in old code for now. When a discouraged feature is |
| 6081 |
used, no warning message is printed like there is for 'deprecated' |
| 6082 |
features. Also, things that are merely discouraged are nevertheless |
| 6083 |
implemented efficiently, while deprecated features can be very slow. |
| 6084 |
|
| 6085 |
You can omit discouraged features from libguile by configuring it with |
| 6086 |
the '--disable-discouraged' option. |
| 6087 |
|
| 6088 |
** Deprecation warnings can be controlled at run-time. |
| 6089 |
|
| 6090 |
(debug-enable 'warn-deprecated) switches them on and (debug-disable |
| 6091 |
'warn-deprecated) switches them off. |
| 6092 |
|
| 6093 |
** Support for SRFI 61, extended cond syntax for multiple values has |
| 6094 |
been added. |
| 6095 |
|
| 6096 |
This SRFI is always available. |
| 6097 |
|
| 6098 |
** Support for require-extension, SRFI-55, has been added. |
| 6099 |
|
| 6100 |
The SRFI-55 special form `require-extension' has been added. It is |
| 6101 |
available at startup, and provides a portable way to load Scheme |
| 6102 |
extensions. SRFI-55 only requires support for one type of extension, |
| 6103 |
"srfi"; so a set of SRFIs may be loaded via (require-extension (srfi 1 |
| 6104 |
13 14)). |
| 6105 |
|
| 6106 |
** New module (srfi srfi-26) provides support for `cut' and `cute'. |
| 6107 |
|
| 6108 |
The (srfi srfi-26) module is an implementation of SRFI-26 which |
| 6109 |
provides the `cut' and `cute' syntax. These may be used to specialize |
| 6110 |
parameters without currying. |
| 6111 |
|
| 6112 |
** New module (srfi srfi-31) |
| 6113 |
|
| 6114 |
This is an implementation of SRFI-31 which provides a special form |
| 6115 |
`rec' for recursive evaluation. |
| 6116 |
|
| 6117 |
** The modules (srfi srfi-13), (srfi srfi-14) and (srfi srfi-4) have |
| 6118 |
been merged with the core, making their functionality always |
| 6119 |
available. |
| 6120 |
|
| 6121 |
The modules are still available, tho, and you could use them together |
| 6122 |
with a renaming import, for example. |
| 6123 |
|
| 6124 |
** Guile no longer includes its own version of libltdl. |
| 6125 |
|
| 6126 |
The official version is good enough now. |
| 6127 |
|
| 6128 |
** The --enable-htmldoc option has been removed from 'configure'. |
| 6129 |
|
| 6130 |
Support for translating the documentation into HTML is now always |
| 6131 |
provided. Use 'make html'. |
| 6132 |
|
| 6133 |
** New module (ice-9 serialize): |
| 6134 |
|
| 6135 |
(serialize FORM1 ...) and (parallelize FORM1 ...) are useful when you |
| 6136 |
don't trust the thread safety of most of your program, but where you |
| 6137 |
have some section(s) of code which you consider can run in parallel to |
| 6138 |
other sections. See ice-9/serialize.scm for more information. |
| 6139 |
|
| 6140 |
** The configure option '--disable-arrays' has been removed. |
| 6141 |
|
| 6142 |
Support for arrays and uniform numeric arrays is now always included |
| 6143 |
in Guile. |
| 6144 |
|
| 6145 |
* Changes to the stand-alone interpreter |
| 6146 |
|
| 6147 |
** New command line option `-L'. |
| 6148 |
|
| 6149 |
This option adds a directory to the front of the load path. |
| 6150 |
|
| 6151 |
** New command line option `--no-debug'. |
| 6152 |
|
| 6153 |
Specifying `--no-debug' on the command line will keep the debugging |
| 6154 |
evaluator turned off, even for interactive sessions. |
| 6155 |
|
| 6156 |
** User-init file ~/.guile is now loaded with the debugging evaluator. |
| 6157 |
|
| 6158 |
Previously, the normal evaluator would have been used. Using the |
| 6159 |
debugging evaluator gives better error messages. |
| 6160 |
|
| 6161 |
** The '-e' option now 'read's its argument. |
| 6162 |
|
| 6163 |
This is to allow the new '(@ MODULE-NAME VARIABLE-NAME)' construct to |
| 6164 |
be used with '-e'. For example, you can now write a script like |
| 6165 |
|
| 6166 |
#! /bin/sh |
| 6167 |
exec guile -e '(@ (demo) main)' -s "$0" "$@" |
| 6168 |
!# |
| 6169 |
|
| 6170 |
(define-module (demo) |
| 6171 |
:export (main)) |
| 6172 |
|
| 6173 |
(define (main args) |
| 6174 |
(format #t "Demo: ~a~%" args)) |
| 6175 |
|
| 6176 |
|
| 6177 |
* Changes to Scheme functions and syntax |
| 6178 |
|
| 6179 |
** Guardians have changed back to their original semantics |
| 6180 |
|
| 6181 |
Guardians now behave like described in the paper by Dybvig et al. In |
| 6182 |
particular, they no longer make guarantees about the order in which |
| 6183 |
they return objects, and they can no longer be greedy. |
| 6184 |
|
| 6185 |
They no longer drop cyclic data structures. |
| 6186 |
|
| 6187 |
The C function scm_make_guardian has been changed incompatibly and no |
| 6188 |
longer takes the 'greedy_p' argument. |
| 6189 |
|
| 6190 |
** New function hashx-remove! |
| 6191 |
|
| 6192 |
This function completes the set of 'hashx' functions. |
| 6193 |
|
| 6194 |
** The concept of dynamic roots has been factored into continuation |
| 6195 |
barriers and dynamic states. |
| 6196 |
|
| 6197 |
Each thread has a current dynamic state that carries the values of the |
| 6198 |
fluids. You can create and copy dynamic states and use them as the |
| 6199 |
second argument for 'eval'. See "Fluids and Dynamic States" in the |
| 6200 |
manual. |
| 6201 |
|
| 6202 |
To restrict the influence that captured continuations can have on the |
| 6203 |
control flow, you can errect continuation barriers. See "Continuation |
| 6204 |
Barriers" in the manual. |
| 6205 |
|
| 6206 |
The function call-with-dynamic-root now essentially temporarily |
| 6207 |
installs a new dynamic state and errects a continuation barrier. |
| 6208 |
|
| 6209 |
** The default load path no longer includes "." at the end. |
| 6210 |
|
| 6211 |
Automatically loading modules from the current directory should not |
| 6212 |
happen by default. If you want to allow it in a more controlled |
| 6213 |
manner, set the environment variable GUILE_LOAD_PATH or the Scheme |
| 6214 |
variable %load-path. |
| 6215 |
|
| 6216 |
** The uniform vector and array support has been overhauled. |
| 6217 |
|
| 6218 |
It now complies with SRFI-4 and the weird prototype based uniform |
| 6219 |
array creation has been deprecated. See the manual for more details. |
| 6220 |
|
| 6221 |
Some non-compatible changes have been made: |
| 6222 |
- characters can no longer be stored into byte arrays. |
| 6223 |
- strings and bit vectors are no longer considered to be uniform numeric |
| 6224 |
vectors. |
| 6225 |
- array-rank throws an error for non-arrays instead of returning zero. |
| 6226 |
- array-ref does no longer accept non-arrays when no indices are given. |
| 6227 |
|
| 6228 |
There is the new notion of 'generalized vectors' and corresponding |
| 6229 |
procedures like 'generalized-vector-ref'. Generalized vectors include |
| 6230 |
strings, bitvectors, ordinary vectors, and uniform numeric vectors. |
| 6231 |
|
| 6232 |
Arrays use generalized vectors as their storage, so that you still |
| 6233 |
have arrays of characters, bits, etc. However, uniform-array-read! |
| 6234 |
and uniform-array-write can no longer read/write strings and |
| 6235 |
bitvectors. |
| 6236 |
|
| 6237 |
** There is now support for copy-on-write substrings, mutation-sharing |
| 6238 |
substrings and read-only strings. |
| 6239 |
|
| 6240 |
Three new procedures are related to this: substring/shared, |
| 6241 |
substring/copy, and substring/read-only. See the manual for more |
| 6242 |
information. |
| 6243 |
|
| 6244 |
** Backtraces will now highlight the value that caused the error. |
| 6245 |
|
| 6246 |
By default, these values are enclosed in "{...}", such as in this |
| 6247 |
example: |
| 6248 |
|
| 6249 |
guile> (car 'a) |
| 6250 |
|
| 6251 |
Backtrace: |
| 6252 |
In current input: |
| 6253 |
1: 0* [car {a}] |
| 6254 |
|
| 6255 |
<unnamed port>:1:1: In procedure car in expression (car (quote a)): |
| 6256 |
<unnamed port>:1:1: Wrong type (expecting pair): a |
| 6257 |
ABORT: (wrong-type-arg) |
| 6258 |
|
| 6259 |
The prefix and suffix used for highlighting can be set via the two new |
| 6260 |
printer options 'highlight-prefix' and 'highlight-suffix'. For |
| 6261 |
example, putting this into ~/.guile will output the bad value in bold |
| 6262 |
on an ANSI terminal: |
| 6263 |
|
| 6264 |
(print-set! highlight-prefix "\x1b[1m") |
| 6265 |
(print-set! highlight-suffix "\x1b[22m") |
| 6266 |
|
| 6267 |
|
| 6268 |
** 'gettext' support for internationalization has been added. |
| 6269 |
|
| 6270 |
See the manual for details. |
| 6271 |
|
| 6272 |
** New syntax '@' and '@@': |
| 6273 |
|
| 6274 |
You can now directly refer to variables exported from a module by |
| 6275 |
writing |
| 6276 |
|
| 6277 |
(@ MODULE-NAME VARIABLE-NAME) |
| 6278 |
|
| 6279 |
For example (@ (ice-9 pretty-print) pretty-print) will directly access |
| 6280 |
the pretty-print variable exported from the (ice-9 pretty-print) |
| 6281 |
module. You don't need to 'use' that module first. You can also use |
| 6282 |
'@' as a target of 'set!', as in (set! (@ mod var) val). |
| 6283 |
|
| 6284 |
The related syntax (@@ MODULE-NAME VARIABLE-NAME) works just like '@', |
| 6285 |
but it can also access variables that have not been exported. It is |
| 6286 |
intended only for kluges and temporary fixes and for debugging, not |
| 6287 |
for ordinary code. |
| 6288 |
|
| 6289 |
** Keyword syntax has been made more disciplined. |
| 6290 |
|
| 6291 |
Previously, the name of a keyword was read as a 'token' but printed as |
| 6292 |
a symbol. Now, it is read as a general Scheme datum which must be a |
| 6293 |
symbol. |
| 6294 |
|
| 6295 |
Previously: |
| 6296 |
|
| 6297 |
guile> #:12 |
| 6298 |
#:#{12}# |
| 6299 |
guile> #:#{12}# |
| 6300 |
#:#{\#{12}\#}# |
| 6301 |
guile> #:(a b c) |
| 6302 |
#:#{}# |
| 6303 |
ERROR: In expression (a b c): |
| 6304 |
Unbound variable: a |
| 6305 |
guile> #: foo |
| 6306 |
#:#{}# |
| 6307 |
ERROR: Unbound variable: foo |
| 6308 |
|
| 6309 |
Now: |
| 6310 |
|
| 6311 |
guile> #:12 |
| 6312 |
ERROR: Wrong type (expecting symbol): 12 |
| 6313 |
guile> #:#{12}# |
| 6314 |
#:#{12}# |
| 6315 |
guile> #:(a b c) |
| 6316 |
ERROR: Wrong type (expecting symbol): (a b c) |
| 6317 |
guile> #: foo |
| 6318 |
#:foo |
| 6319 |
|
| 6320 |
** The printing of symbols that might look like keywords can be |
| 6321 |
controlled. |
| 6322 |
|
| 6323 |
The new printer option 'quote-keywordish-symbols' controls how symbols |
| 6324 |
are printed that have a colon as their first or last character. The |
| 6325 |
default now is to only quote a symbol with #{...}# when the read |
| 6326 |
option 'keywords' is not '#f'. Thus: |
| 6327 |
|
| 6328 |
guile> (define foo (string->symbol ":foo")) |
| 6329 |
guile> (read-set! keywords #f) |
| 6330 |
guile> foo |
| 6331 |
:foo |
| 6332 |
guile> (read-set! keywords 'prefix) |
| 6333 |
guile> foo |
| 6334 |
#{:foo}# |
| 6335 |
guile> (print-set! quote-keywordish-symbols #f) |
| 6336 |
guile> foo |
| 6337 |
:foo |
| 6338 |
|
| 6339 |
** 'while' now provides 'break' and 'continue' |
| 6340 |
|
| 6341 |
break and continue were previously bound in a while loop, but not |
| 6342 |
documented, and continue didn't quite work properly. The undocumented |
| 6343 |
parameter to break which gave a return value for the while has been |
| 6344 |
dropped. |
| 6345 |
|
| 6346 |
** 'call-with-current-continuation' is now also available under the name |
| 6347 |
'call/cc'. |
| 6348 |
|
| 6349 |
** The module system now checks for duplicate bindings. |
| 6350 |
|
| 6351 |
The module system now can check for name conflicts among imported |
| 6352 |
bindings. |
| 6353 |
|
| 6354 |
The behavior can be controlled by specifying one or more 'duplicates' |
| 6355 |
handlers. For example, to make Guile return an error for every name |
| 6356 |
collision, write: |
| 6357 |
|
| 6358 |
(define-module (foo) |
| 6359 |
:use-module (bar) |
| 6360 |
:use-module (baz) |
| 6361 |
:duplicates check) |
| 6362 |
|
| 6363 |
The new default behavior of the module system when a name collision |
| 6364 |
has been detected is to |
| 6365 |
|
| 6366 |
1. Give priority to bindings marked as a replacement. |
| 6367 |
2. Issue a warning (different warning if overriding core binding). |
| 6368 |
3. Give priority to the last encountered binding (this corresponds to |
| 6369 |
the old behavior). |
| 6370 |
|
| 6371 |
If you want the old behavior back without replacements or warnings you |
| 6372 |
can add the line: |
| 6373 |
|
| 6374 |
(default-duplicate-binding-handler 'last) |
| 6375 |
|
| 6376 |
to your .guile init file. |
| 6377 |
|
| 6378 |
** New define-module option: :replace |
| 6379 |
|
| 6380 |
:replace works as :export, but, in addition, marks the binding as a |
| 6381 |
replacement. |
| 6382 |
|
| 6383 |
A typical example is `format' in (ice-9 format) which is a replacement |
| 6384 |
for the core binding `format'. |
| 6385 |
|
| 6386 |
** Adding prefixes to imported bindings in the module system |
| 6387 |
|
| 6388 |
There is now a new :use-module option :prefix. It can be used to add |
| 6389 |
a prefix to all imported bindings. |
| 6390 |
|
| 6391 |
(define-module (foo) |
| 6392 |
:use-module ((bar) :prefix bar:)) |
| 6393 |
|
| 6394 |
will import all bindings exported from bar, but rename them by adding |
| 6395 |
the prefix `bar:'. |
| 6396 |
|
| 6397 |
** Conflicting generic functions can be automatically merged. |
| 6398 |
|
| 6399 |
When two imported bindings conflict and they are both generic |
| 6400 |
functions, the two functions can now be merged automatically. This is |
| 6401 |
activated with the 'duplicates' handler 'merge-generics'. |
| 6402 |
|
| 6403 |
** New function: effective-version |
| 6404 |
|
| 6405 |
Returns the "effective" version number. This is just the normal full |
| 6406 |
version string without the final micro-version number. See "Changes |
| 6407 |
to the distribution" above. |
| 6408 |
|
| 6409 |
** New threading functions: parallel, letpar, par-map, and friends |
| 6410 |
|
| 6411 |
These are convenient ways to run calculations in parallel in new |
| 6412 |
threads. See "Parallel forms" in the manual for details. |
| 6413 |
|
| 6414 |
** New function 'try-mutex'. |
| 6415 |
|
| 6416 |
This function will attempt to lock a mutex but will return immediately |
| 6417 |
instead of blocking and indicate failure. |
| 6418 |
|
| 6419 |
** Waiting on a condition variable can have a timeout. |
| 6420 |
|
| 6421 |
The function 'wait-condition-variable' now takes a third, optional |
| 6422 |
argument that specifies the point in time where the waiting should be |
| 6423 |
aborted. |
| 6424 |
|
| 6425 |
** New function 'broadcast-condition-variable'. |
| 6426 |
|
| 6427 |
** New functions 'all-threads' and 'current-thread'. |
| 6428 |
|
| 6429 |
** Signals and system asyncs work better with threads. |
| 6430 |
|
| 6431 |
The function 'sigaction' now takes a fourth, optional, argument that |
| 6432 |
specifies the thread that the handler should run in. When the |
| 6433 |
argument is omitted, the handler will run in the thread that called |
| 6434 |
'sigaction'. |
| 6435 |
|
| 6436 |
Likewise, 'system-async-mark' takes a second, optional, argument that |
| 6437 |
specifies the thread that the async should run in. When it is |
| 6438 |
omitted, the async will run in the thread that called |
| 6439 |
'system-async-mark'. |
| 6440 |
|
| 6441 |
C code can use the new functions scm_sigaction_for_thread and |
| 6442 |
scm_system_async_mark_for_thread to pass the new thread argument. |
| 6443 |
|
| 6444 |
When a thread blocks on a mutex, a condition variable or is waiting |
| 6445 |
for IO to be possible, it will still execute system asyncs. This can |
| 6446 |
be used to interrupt such a thread by making it execute a 'throw', for |
| 6447 |
example. |
| 6448 |
|
| 6449 |
** The function 'system-async' is deprecated. |
| 6450 |
|
| 6451 |
You can now pass any zero-argument procedure to 'system-async-mark'. |
| 6452 |
The function 'system-async' will just return its argument unchanged |
| 6453 |
now. |
| 6454 |
|
| 6455 |
** New functions 'call-with-blocked-asyncs' and |
| 6456 |
'call-with-unblocked-asyncs' |
| 6457 |
|
| 6458 |
The expression (call-with-blocked-asyncs PROC) will call PROC and will |
| 6459 |
block execution of system asyncs for the current thread by one level |
| 6460 |
while PROC runs. Likewise, call-with-unblocked-asyncs will call a |
| 6461 |
procedure and will unblock the execution of system asyncs by one |
| 6462 |
level for the current thread. |
| 6463 |
|
| 6464 |
Only system asyncs are affected by these functions. |
| 6465 |
|
| 6466 |
** The functions 'mask-signals' and 'unmask-signals' are deprecated. |
| 6467 |
|
| 6468 |
Use 'call-with-blocked-asyncs' or 'call-with-unblocked-asyncs' |
| 6469 |
instead. Those functions are easier to use correctly and can be |
| 6470 |
nested. |
| 6471 |
|
| 6472 |
** New function 'unsetenv'. |
| 6473 |
|
| 6474 |
** New macro 'define-syntax-public'. |
| 6475 |
|
| 6476 |
It works like 'define-syntax' and also exports the defined macro (but |
| 6477 |
only on top-level). |
| 6478 |
|
| 6479 |
** There is support for Infinity and NaNs. |
| 6480 |
|
| 6481 |
Following PLT Scheme, Guile can now work with infinite numbers, and |
| 6482 |
'not-a-numbers'. |
| 6483 |
|
| 6484 |
There is new syntax for numbers: "+inf.0" (infinity), "-inf.0" |
| 6485 |
(negative infinity), "+nan.0" (not-a-number), and "-nan.0" (same as |
| 6486 |
"+nan.0"). These numbers are inexact and have no exact counterpart. |
| 6487 |
|
| 6488 |
Dividing by an inexact zero returns +inf.0 or -inf.0, depending on the |
| 6489 |
sign of the dividend. The infinities are integers, and they answer #t |
| 6490 |
for both 'even?' and 'odd?'. The +nan.0 value is not an integer and is |
| 6491 |
not '=' to itself, but '+nan.0' is 'eqv?' to itself. |
| 6492 |
|
| 6493 |
For example |
| 6494 |
|
| 6495 |
(/ 1 0.0) |
| 6496 |
=> +inf.0 |
| 6497 |
|
| 6498 |
(/ 0 0.0) |
| 6499 |
=> +nan.0 |
| 6500 |
|
| 6501 |
(/ 0) |
| 6502 |
ERROR: Numerical overflow |
| 6503 |
|
| 6504 |
Two new predicates 'inf?' and 'nan?' can be used to test for the |
| 6505 |
special values. |
| 6506 |
|
| 6507 |
** Inexact zero can have a sign. |
| 6508 |
|
| 6509 |
Guile can now distinguish between plus and minus inexact zero, if your |
| 6510 |
platform supports this, too. The two zeros are equal according to |
| 6511 |
'=', but not according to 'eqv?'. For example |
| 6512 |
|
| 6513 |
(- 0.0) |
| 6514 |
=> -0.0 |
| 6515 |
|
| 6516 |
(= 0.0 (- 0.0)) |
| 6517 |
=> #t |
| 6518 |
|
| 6519 |
(eqv? 0.0 (- 0.0)) |
| 6520 |
=> #f |
| 6521 |
|
| 6522 |
** Guile now has exact rationals. |
| 6523 |
|
| 6524 |
Guile can now represent fractions such as 1/3 exactly. Computing with |
| 6525 |
them is also done exactly, of course: |
| 6526 |
|
| 6527 |
(* 1/3 3/2) |
| 6528 |
=> 1/2 |
| 6529 |
|
| 6530 |
** 'floor', 'ceiling', 'round' and 'truncate' now return exact numbers |
| 6531 |
for exact arguments. |
| 6532 |
|
| 6533 |
For example: (floor 2) now returns an exact 2 where in the past it |
| 6534 |
returned an inexact 2.0. Likewise, (floor 5/4) returns an exact 1. |
| 6535 |
|
| 6536 |
** inexact->exact no longer returns only integers. |
| 6537 |
|
| 6538 |
Without exact rationals, the closest exact number was always an |
| 6539 |
integer, but now inexact->exact returns the fraction that is exactly |
| 6540 |
equal to a floating point number. For example: |
| 6541 |
|
| 6542 |
(inexact->exact 1.234) |
| 6543 |
=> 694680242521899/562949953421312 |
| 6544 |
|
| 6545 |
When you want the old behavior, use 'round' explicitly: |
| 6546 |
|
| 6547 |
(inexact->exact (round 1.234)) |
| 6548 |
=> 1 |
| 6549 |
|
| 6550 |
** New function 'rationalize'. |
| 6551 |
|
| 6552 |
This function finds a simple fraction that is close to a given real |
| 6553 |
number. For example (and compare with inexact->exact above): |
| 6554 |
|
| 6555 |
(rationalize (inexact->exact 1.234) 1/2000) |
| 6556 |
=> 58/47 |
| 6557 |
|
| 6558 |
Note that, as required by R5RS, rationalize returns only then an exact |
| 6559 |
result when both its arguments are exact. |
| 6560 |
|
| 6561 |
** 'odd?' and 'even?' work also for inexact integers. |
| 6562 |
|
| 6563 |
Previously, (odd? 1.0) would signal an error since only exact integers |
| 6564 |
were recognized as integers. Now (odd? 1.0) returns #t, (odd? 2.0) |
| 6565 |
returns #f and (odd? 1.5) signals an error. |
| 6566 |
|
| 6567 |
** Guile now has uninterned symbols. |
| 6568 |
|
| 6569 |
The new function 'make-symbol' will return an uninterned symbol. This |
| 6570 |
is a symbol that is unique and is guaranteed to remain unique. |
| 6571 |
However, uninterned symbols can not yet be read back in. |
| 6572 |
|
| 6573 |
Use the new function 'symbol-interned?' to check whether a symbol is |
| 6574 |
interned or not. |
| 6575 |
|
| 6576 |
** pretty-print has more options. |
| 6577 |
|
| 6578 |
The function pretty-print from the (ice-9 pretty-print) module can now |
| 6579 |
also be invoked with keyword arguments that control things like |
| 6580 |
maximum output width. See the manual for details. |
| 6581 |
|
| 6582 |
** Variables have no longer a special behavior for `equal?'. |
| 6583 |
|
| 6584 |
Previously, comparing two variables with `equal?' would recursivly |
| 6585 |
compare their values. This is no longer done. Variables are now only |
| 6586 |
`equal?' if they are `eq?'. |
| 6587 |
|
| 6588 |
** `(begin)' is now valid. |
| 6589 |
|
| 6590 |
You can now use an empty `begin' form. It will yield #<unspecified> |
| 6591 |
when evaluated and simply be ignored in a definition context. |
| 6592 |
|
| 6593 |
** Deprecated: procedure->macro |
| 6594 |
|
| 6595 |
Change your code to use 'define-macro' or r5rs macros. Also, be aware |
| 6596 |
that macro expansion will not be done during evaluation, but prior to |
| 6597 |
evaluation. |
| 6598 |
|
| 6599 |
** Soft ports now allow a `char-ready?' procedure |
| 6600 |
|
| 6601 |
The vector argument to `make-soft-port' can now have a length of |
| 6602 |
either 5 or 6. (Previously the length had to be 5.) The optional 6th |
| 6603 |
element is interpreted as an `input-waiting' thunk -- i.e. a thunk |
| 6604 |
that returns the number of characters that can be read immediately |
| 6605 |
without the soft port blocking. |
| 6606 |
|
| 6607 |
** Deprecated: undefine |
| 6608 |
|
| 6609 |
There is no replacement for undefine. |
| 6610 |
|
| 6611 |
** The functions make-keyword-from-dash-symbol and keyword-dash-symbol |
| 6612 |
have been discouraged. |
| 6613 |
|
| 6614 |
They are relics from a time where a keyword like #:foo was used |
| 6615 |
directly as a Tcl option "-foo" and thus keywords were internally |
| 6616 |
stored as a symbol with a starting dash. We now store a symbol |
| 6617 |
without the dash. |
| 6618 |
|
| 6619 |
Use symbol->keyword and keyword->symbol instead. |
| 6620 |
|
| 6621 |
** The `cheap' debug option is now obsolete |
| 6622 |
|
| 6623 |
Evaluator trap calls are now unconditionally "cheap" - in other words, |
| 6624 |
they pass a debug object to the trap handler rather than a full |
| 6625 |
continuation. The trap handler code can capture a full continuation |
| 6626 |
by using `call-with-current-continuation' in the usual way, if it so |
| 6627 |
desires. |
| 6628 |
|
| 6629 |
The `cheap' option is retained for now so as not to break existing |
| 6630 |
code which gets or sets it, but setting it now has no effect. It will |
| 6631 |
be removed in the next major Guile release. |
| 6632 |
|
| 6633 |
** Evaluator trap calls now support `tweaking' |
| 6634 |
|
| 6635 |
`Tweaking' means that the trap handler code can modify the Scheme |
| 6636 |
expression that is about to be evaluated (in the case of an |
| 6637 |
enter-frame trap) or the value that is being returned (in the case of |
| 6638 |
an exit-frame trap). The trap handler code indicates that it wants to |
| 6639 |
do this by returning a pair whose car is the symbol 'instead and whose |
| 6640 |
cdr is the modified expression or return value. |
| 6641 |
|
| 6642 |
* Changes to the C interface |
| 6643 |
|
| 6644 |
** The functions scm_hash_fn_remove_x and scm_hashx_remove_x no longer |
| 6645 |
take a 'delete' function argument. |
| 6646 |
|
| 6647 |
This argument makes no sense since the delete function is used to |
| 6648 |
remove a pair from an alist, and this must not be configurable. |
| 6649 |
|
| 6650 |
This is an incompatible change. |
| 6651 |
|
| 6652 |
** The GH interface is now subject to the deprecation mechanism |
| 6653 |
|
| 6654 |
The GH interface has been deprecated for quite some time but now it is |
| 6655 |
actually removed from Guile when it is configured with |
| 6656 |
--disable-deprecated. |
| 6657 |
|
| 6658 |
See the manual "Transitioning away from GH" for more information. |
| 6659 |
|
| 6660 |
** A new family of functions for converting between C values and |
| 6661 |
Scheme values has been added. |
| 6662 |
|
| 6663 |
These functions follow a common naming scheme and are designed to be |
| 6664 |
easier to use, thread-safe and more future-proof than the older |
| 6665 |
alternatives. |
| 6666 |
|
| 6667 |
- int scm_is_* (...) |
| 6668 |
|
| 6669 |
These are predicates that return a C boolean: 1 or 0. Instead of |
| 6670 |
SCM_NFALSEP, you can now use scm_is_true, for example. |
| 6671 |
|
| 6672 |
- <type> scm_to_<type> (SCM val, ...) |
| 6673 |
|
| 6674 |
These are functions that convert a Scheme value into an appropriate |
| 6675 |
C value. For example, you can use scm_to_int to safely convert from |
| 6676 |
a SCM to an int. |
| 6677 |
|
| 6678 |
- SCM scm_from_<type> (<type> val, ...) |
| 6679 |
|
| 6680 |
These functions convert from a C type to a SCM value; for example, |
| 6681 |
scm_from_int for ints. |
| 6682 |
|
| 6683 |
There is a huge number of these functions, for numbers, strings, |
| 6684 |
symbols, vectors, etc. They are documented in the reference manual in |
| 6685 |
the API section together with the types that they apply to. |
| 6686 |
|
| 6687 |
** New functions for dealing with complex numbers in C have been added. |
| 6688 |
|
| 6689 |
The new functions are scm_c_make_rectangular, scm_c_make_polar, |
| 6690 |
scm_c_real_part, scm_c_imag_part, scm_c_magnitude and scm_c_angle. |
| 6691 |
They work like scm_make_rectangular etc but take or return doubles |
| 6692 |
directly. |
| 6693 |
|
| 6694 |
** The function scm_make_complex has been discouraged. |
| 6695 |
|
| 6696 |
Use scm_c_make_rectangular instead. |
| 6697 |
|
| 6698 |
** The INUM macros have been deprecated. |
| 6699 |
|
| 6700 |
A lot of code uses these macros to do general integer conversions, |
| 6701 |
although the macros only work correctly with fixnums. Use the |
| 6702 |
following alternatives. |
| 6703 |
|
| 6704 |
SCM_INUMP -> scm_is_integer or similar |
| 6705 |
SCM_NINUMP -> !scm_is_integer or similar |
| 6706 |
SCM_MAKINUM -> scm_from_int or similar |
| 6707 |
SCM_INUM -> scm_to_int or similar |
| 6708 |
|
| 6709 |
SCM_VALIDATE_INUM_* -> Do not use these; scm_to_int, etc. will |
| 6710 |
do the validating for you. |
| 6711 |
|
| 6712 |
** The scm_num2<type> and scm_<type>2num functions and scm_make_real |
| 6713 |
have been discouraged. |
| 6714 |
|
| 6715 |
Use the newer scm_to_<type> and scm_from_<type> functions instead for |
| 6716 |
new code. The functions have been discouraged since they don't fit |
| 6717 |
the naming scheme. |
| 6718 |
|
| 6719 |
** The 'boolean' macros SCM_FALSEP etc have been discouraged. |
| 6720 |
|
| 6721 |
They have strange names, especially SCM_NFALSEP, and SCM_BOOLP |
| 6722 |
evaluates its argument twice. Use scm_is_true, etc. instead for new |
| 6723 |
code. |
| 6724 |
|
| 6725 |
** The macro SCM_EQ_P has been discouraged. |
| 6726 |
|
| 6727 |
Use scm_is_eq for new code, which fits better into the naming |
| 6728 |
conventions. |
| 6729 |
|
| 6730 |
** The macros SCM_CONSP, SCM_NCONSP, SCM_NULLP, and SCM_NNULLP have |
| 6731 |
been discouraged. |
| 6732 |
|
| 6733 |
Use the function scm_is_pair or scm_is_null instead. |
| 6734 |
|
| 6735 |
** The functions scm_round and scm_truncate have been deprecated and |
| 6736 |
are now available as scm_c_round and scm_c_truncate, respectively. |
| 6737 |
|
| 6738 |
These functions occupy the names that scm_round_number and |
| 6739 |
scm_truncate_number should have. |
| 6740 |
|
| 6741 |
** The functions scm_c_string2str, scm_c_substring2str, and |
| 6742 |
scm_c_symbol2str have been deprecated. |
| 6743 |
|
| 6744 |
Use scm_to_locale_stringbuf or similar instead, maybe together with |
| 6745 |
scm_substring. |
| 6746 |
|
| 6747 |
** New functions scm_c_make_string, scm_c_string_length, |
| 6748 |
scm_c_string_ref, scm_c_string_set_x, scm_c_substring, |
| 6749 |
scm_c_substring_shared, scm_c_substring_copy. |
| 6750 |
|
| 6751 |
These are like scm_make_string, scm_length, etc. but are slightly |
| 6752 |
easier to use from C. |
| 6753 |
|
| 6754 |
** The macros SCM_STRINGP, SCM_STRING_CHARS, SCM_STRING_LENGTH, |
| 6755 |
SCM_SYMBOL_CHARS, and SCM_SYMBOL_LENGTH have been deprecated. |
| 6756 |
|
| 6757 |
They export too many assumptions about the implementation of strings |
| 6758 |
and symbols that are no longer true in the presence of |
| 6759 |
mutation-sharing substrings and when Guile switches to some form of |
| 6760 |
Unicode. |
| 6761 |
|
| 6762 |
When working with strings, it is often best to use the normal string |
| 6763 |
functions provided by Guile, such as scm_c_string_ref, |
| 6764 |
scm_c_string_set_x, scm_string_append, etc. Be sure to look in the |
| 6765 |
manual since many more such functions are now provided than |
| 6766 |
previously. |
| 6767 |
|
| 6768 |
When you want to convert a SCM string to a C string, use the |
| 6769 |
scm_to_locale_string function or similar instead. For symbols, use |
| 6770 |
scm_symbol_to_string and then work with that string. Because of the |
| 6771 |
new string representation, scm_symbol_to_string does not need to copy |
| 6772 |
and is thus quite efficient. |
| 6773 |
|
| 6774 |
** Some string, symbol and keyword functions have been discouraged. |
| 6775 |
|
| 6776 |
They don't fit into the uniform naming scheme and are not explicit |
| 6777 |
about the character encoding. |
| 6778 |
|
| 6779 |
Replace according to the following table: |
| 6780 |
|
| 6781 |
scm_allocate_string -> scm_c_make_string |
| 6782 |
scm_take_str -> scm_take_locale_stringn |
| 6783 |
scm_take0str -> scm_take_locale_string |
| 6784 |
scm_mem2string -> scm_from_locale_stringn |
| 6785 |
scm_str2string -> scm_from_locale_string |
| 6786 |
scm_makfrom0str -> scm_from_locale_string |
| 6787 |
scm_mem2symbol -> scm_from_locale_symboln |
| 6788 |
scm_mem2uninterned_symbol -> scm_from_locale_stringn + scm_make_symbol |
| 6789 |
scm_str2symbol -> scm_from_locale_symbol |
| 6790 |
|
| 6791 |
SCM_SYMBOL_HASH -> scm_hashq |
| 6792 |
SCM_SYMBOL_INTERNED_P -> scm_symbol_interned_p |
| 6793 |
|
| 6794 |
scm_c_make_keyword -> scm_from_locale_keyword |
| 6795 |
|
| 6796 |
** The functions scm_keyword_to_symbol and sym_symbol_to_keyword are |
| 6797 |
now also available to C code. |
| 6798 |
|
| 6799 |
** SCM_KEYWORDP and SCM_KEYWORDSYM have been deprecated. |
| 6800 |
|
| 6801 |
Use scm_is_keyword and scm_keyword_to_symbol instead, but note that |
| 6802 |
the latter returns the true name of the keyword, not the 'dash name', |
| 6803 |
as SCM_KEYWORDSYM used to do. |
| 6804 |
|
| 6805 |
** A new way to access arrays in a thread-safe and efficient way has |
| 6806 |
been added. |
| 6807 |
|
| 6808 |
See the manual, node "Accessing Arrays From C". |
| 6809 |
|
| 6810 |
** The old uniform vector and bitvector implementations have been |
| 6811 |
unceremoniously removed. |
| 6812 |
|
| 6813 |
This implementation exposed the details of the tagging system of |
| 6814 |
Guile. Use the new C API explained in the manual in node "Uniform |
| 6815 |
Numeric Vectors" and "Bit Vectors", respectively. |
| 6816 |
|
| 6817 |
The following macros are gone: SCM_UVECTOR_BASE, SCM_SET_UVECTOR_BASE, |
| 6818 |
SCM_UVECTOR_MAXLENGTH, SCM_UVECTOR_LENGTH, SCM_MAKE_UVECTOR_TAG, |
| 6819 |
SCM_SET_UVECTOR_LENGTH, SCM_BITVECTOR_P, SCM_BITVECTOR_BASE, |
| 6820 |
SCM_SET_BITVECTOR_BASE, SCM_BITVECTOR_MAX_LENGTH, |
| 6821 |
SCM_BITVECTOR_LENGTH, SCM_MAKE_BITVECTOR_TAG, |
| 6822 |
SCM_SET_BITVECTOR_LENGTH, SCM_BITVEC_REF, SCM_BITVEC_SET, |
| 6823 |
SCM_BITVEC_CLR. |
| 6824 |
|
| 6825 |
** The macros dealing with vectors have been deprecated. |
| 6826 |
|
| 6827 |
Use the new functions scm_is_vector, scm_vector_elements, |
| 6828 |
scm_vector_writable_elements, etc, or scm_is_simple_vector, |
| 6829 |
SCM_SIMPLE_VECTOR_REF, SCM_SIMPLE_VECTOR_SET, etc instead. See the |
| 6830 |
manual for more details. |
| 6831 |
|
| 6832 |
Deprecated are SCM_VECTORP, SCM_VELTS, SCM_VECTOR_MAX_LENGTH, |
| 6833 |
SCM_VECTOR_LENGTH, SCM_VECTOR_REF, SCM_VECTOR_SET, SCM_WRITABLE_VELTS. |
| 6834 |
|
| 6835 |
The following macros have been removed: SCM_VECTOR_BASE, |
| 6836 |
SCM_SET_VECTOR_BASE, SCM_MAKE_VECTOR_TAG, SCM_SET_VECTOR_LENGTH, |
| 6837 |
SCM_VELTS_AS_STACKITEMS, SCM_SETVELTS, SCM_GC_WRITABLE_VELTS. |
| 6838 |
|
| 6839 |
** Some C functions and macros related to arrays have been deprecated. |
| 6840 |
|
| 6841 |
Migrate according to the following table: |
| 6842 |
|
| 6843 |
scm_make_uve -> scm_make_typed_array, scm_make_u8vector etc. |
| 6844 |
scm_make_ra -> scm_make_array |
| 6845 |
scm_shap2ra -> scm_make_array |
| 6846 |
scm_cvref -> scm_c_generalized_vector_ref |
| 6847 |
scm_ra_set_contp -> do not use |
| 6848 |
scm_aind -> scm_array_handle_pos |
| 6849 |
scm_raprin1 -> scm_display or scm_write |
| 6850 |
|
| 6851 |
SCM_ARRAYP -> scm_is_array |
| 6852 |
SCM_ARRAY_NDIM -> scm_c_array_rank |
| 6853 |
SCM_ARRAY_DIMS -> scm_array_handle_dims |
| 6854 |
SCM_ARRAY_CONTP -> do not use |
| 6855 |
SCM_ARRAY_MEM -> do not use |
| 6856 |
SCM_ARRAY_V -> scm_array_handle_elements or similar |
| 6857 |
SCM_ARRAY_BASE -> do not use |
| 6858 |
|
| 6859 |
** SCM_CELL_WORD_LOC has been deprecated. |
| 6860 |
|
| 6861 |
Use the new macro SCM_CELL_OBJECT_LOC instead, which returns a pointer |
| 6862 |
to a SCM, as opposed to a pointer to a scm_t_bits. |
| 6863 |
|
| 6864 |
This was done to allow the correct use of pointers into the Scheme |
| 6865 |
heap. Previously, the heap words were of type scm_t_bits and local |
| 6866 |
variables and function arguments were of type SCM, making it |
| 6867 |
non-standards-conformant to have a pointer that can point to both. |
| 6868 |
|
| 6869 |
** New macros SCM_SMOB_DATA_2, SCM_SMOB_DATA_3, etc. |
| 6870 |
|
| 6871 |
These macros should be used instead of SCM_CELL_WORD_2/3 to access the |
| 6872 |
second and third words of double smobs. Likewise for |
| 6873 |
SCM_SET_SMOB_DATA_2 and SCM_SET_SMOB_DATA_3. |
| 6874 |
|
| 6875 |
Also, there is SCM_SMOB_FLAGS and SCM_SET_SMOB_FLAGS that should be |
| 6876 |
used to get and set the 16 exra bits in the zeroth word of a smob. |
| 6877 |
|
| 6878 |
And finally, there is SCM_SMOB_OBJECT and SCM_SMOB_SET_OBJECT for |
| 6879 |
accesing the first immediate word of a smob as a SCM value, and there |
| 6880 |
is SCM_SMOB_OBJECT_LOC for getting a pointer to the first immediate |
| 6881 |
smob word. Like wise for SCM_SMOB_OBJECT_2, etc. |
| 6882 |
|
| 6883 |
** New way to deal with non-local exits and re-entries. |
| 6884 |
|
| 6885 |
There is a new set of functions that essentially do what |
| 6886 |
scm_internal_dynamic_wind does, but in a way that is more convenient |
| 6887 |
for C code in some situations. Here is a quick example of how to |
| 6888 |
prevent a potential memory leak: |
| 6889 |
|
| 6890 |
void |
| 6891 |
foo () |
| 6892 |
{ |
| 6893 |
char *mem; |
| 6894 |
|
| 6895 |
scm_dynwind_begin (0); |
| 6896 |
|
| 6897 |
mem = scm_malloc (100); |
| 6898 |
scm_dynwind_unwind_handler (free, mem, SCM_F_WIND_EXPLICITLY); |
| 6899 |
|
| 6900 |
/* MEM would leak if BAR throws an error. |
| 6901 |
SCM_DYNWIND_UNWIND_HANDLER frees it nevertheless. |
| 6902 |
*/ |
| 6903 |
|
| 6904 |
bar (); |
| 6905 |
|
| 6906 |
scm_dynwind_end (); |
| 6907 |
|
| 6908 |
/* Because of SCM_F_WIND_EXPLICITLY, MEM will be freed by |
| 6909 |
SCM_DYNWIND_END as well. |
| 6910 |
*/ |
| 6911 |
} |
| 6912 |
|
| 6913 |
For full documentation, see the node "Dynamic Wind" in the manual. |
| 6914 |
|
| 6915 |
** New function scm_dynwind_free |
| 6916 |
|
| 6917 |
This function calls 'free' on a given pointer when a dynwind context |
| 6918 |
is left. Thus the call to scm_dynwind_unwind_handler above could be |
| 6919 |
replaced with simply scm_dynwind_free (mem). |
| 6920 |
|
| 6921 |
** New functions scm_c_call_with_blocked_asyncs and |
| 6922 |
scm_c_call_with_unblocked_asyncs |
| 6923 |
|
| 6924 |
Like scm_call_with_blocked_asyncs etc. but for C functions. |
| 6925 |
|
| 6926 |
** New functions scm_dynwind_block_asyncs and scm_dynwind_unblock_asyncs |
| 6927 |
|
| 6928 |
In addition to scm_c_call_with_blocked_asyncs you can now also use |
| 6929 |
scm_dynwind_block_asyncs in a 'dynwind context' (see above). Likewise for |
| 6930 |
scm_c_call_with_unblocked_asyncs and scm_dynwind_unblock_asyncs. |
| 6931 |
|
| 6932 |
** The macros SCM_DEFER_INTS, SCM_ALLOW_INTS, SCM_REDEFER_INTS, |
| 6933 |
SCM_REALLOW_INTS have been deprecated. |
| 6934 |
|
| 6935 |
They do no longer fulfill their original role of blocking signal |
| 6936 |
delivery. Depending on what you want to achieve, replace a pair of |
| 6937 |
SCM_DEFER_INTS and SCM_ALLOW_INTS with a dynwind context that locks a |
| 6938 |
mutex, blocks asyncs, or both. See node "Critical Sections" in the |
| 6939 |
manual. |
| 6940 |
|
| 6941 |
** The value 'scm_mask_ints' is no longer writable. |
| 6942 |
|
| 6943 |
Previously, you could set scm_mask_ints directly. This is no longer |
| 6944 |
possible. Use scm_c_call_with_blocked_asyncs and |
| 6945 |
scm_c_call_with_unblocked_asyncs instead. |
| 6946 |
|
| 6947 |
** New way to temporarily set the current input, output or error ports |
| 6948 |
|
| 6949 |
C code can now use scm_dynwind_current_<foo>_port in a 'dynwind |
| 6950 |
context' (see above). <foo> is one of "input", "output" or "error". |
| 6951 |
|
| 6952 |
** New way to temporarily set fluids |
| 6953 |
|
| 6954 |
C code can now use scm_dynwind_fluid in a 'dynwind context' (see |
| 6955 |
above) to temporarily set the value of a fluid. |
| 6956 |
|
| 6957 |
** New types scm_t_intmax and scm_t_uintmax. |
| 6958 |
|
| 6959 |
On platforms that have them, these types are identical to intmax_t and |
| 6960 |
uintmax_t, respectively. On other platforms, they are identical to |
| 6961 |
the largest integer types that Guile knows about. |
| 6962 |
|
| 6963 |
** The functions scm_unmemocopy and scm_unmemoize have been removed. |
| 6964 |
|
| 6965 |
You should not have used them. |
| 6966 |
|
| 6967 |
** Many public #defines with generic names have been made private. |
| 6968 |
|
| 6969 |
#defines with generic names like HAVE_FOO or SIZEOF_FOO have been made |
| 6970 |
private or renamed with a more suitable public name. |
| 6971 |
|
| 6972 |
** The macro SCM_TYP16S has been deprecated. |
| 6973 |
|
| 6974 |
This macro is not intended for public use. |
| 6975 |
|
| 6976 |
** The macro SCM_SLOPPY_INEXACTP has been deprecated. |
| 6977 |
|
| 6978 |
Use scm_is_true (scm_inexact_p (...)) instead. |
| 6979 |
|
| 6980 |
** The macro SCM_SLOPPY_REALP has been deprecated. |
| 6981 |
|
| 6982 |
Use scm_is_real instead. |
| 6983 |
|
| 6984 |
** The macro SCM_SLOPPY_COMPLEXP has been deprecated. |
| 6985 |
|
| 6986 |
Use scm_is_complex instead. |
| 6987 |
|
| 6988 |
** Some preprocessor defines have been deprecated. |
| 6989 |
|
| 6990 |
These defines indicated whether a certain feature was present in Guile |
| 6991 |
or not. Going forward, assume that the features are always present. |
| 6992 |
|
| 6993 |
The macros are: USE_THREADS, GUILE_ISELECT, READER_EXTENSIONS, |
| 6994 |
DEBUG_EXTENSIONS, DYNAMIC_LINKING. |
| 6995 |
|
| 6996 |
The following macros have been removed completely: MEMOIZE_LOCALS, |
| 6997 |
SCM_RECKLESS, SCM_CAUTIOUS. |
| 6998 |
|
| 6999 |
** The preprocessor define STACK_DIRECTION has been deprecated. |
| 7000 |
|
| 7001 |
There should be no need to know about the stack direction for ordinary |
| 7002 |
programs. |
| 7003 |
|
| 7004 |
** New function: scm_effective_version |
| 7005 |
|
| 7006 |
Returns the "effective" version number. This is just the normal full |
| 7007 |
version string without the final micro-version number. See "Changes |
| 7008 |
to the distribution" above. |
| 7009 |
|
| 7010 |
** The function scm_call_with_new_thread has a new prototype. |
| 7011 |
|
| 7012 |
Instead of taking a list with the thunk and handler, these two |
| 7013 |
arguments are now passed directly: |
| 7014 |
|
| 7015 |
SCM scm_call_with_new_thread (SCM thunk, SCM handler); |
| 7016 |
|
| 7017 |
This is an incompatible change. |
| 7018 |
|
| 7019 |
** New snarfer macro SCM_DEFINE_PUBLIC. |
| 7020 |
|
| 7021 |
This is like SCM_DEFINE, but also calls scm_c_export for the defined |
| 7022 |
function in the init section. |
| 7023 |
|
| 7024 |
** The snarfer macro SCM_SNARF_INIT is now officially supported. |
| 7025 |
|
| 7026 |
** Garbage collector rewrite. |
| 7027 |
|
| 7028 |
The garbage collector is cleaned up a lot, and now uses lazy |
| 7029 |
sweeping. This is reflected in the output of (gc-stats); since cells |
| 7030 |
are being freed when they are allocated, the cells-allocated field |
| 7031 |
stays roughly constant. |
| 7032 |
|
| 7033 |
For malloc related triggers, the behavior is changed. It uses the same |
| 7034 |
heuristic as the cell-triggered collections. It may be tuned with the |
| 7035 |
environment variables GUILE_MIN_YIELD_MALLOC. This is the percentage |
| 7036 |
for minimum yield of malloc related triggers. The default is 40. |
| 7037 |
GUILE_INIT_MALLOC_LIMIT sets the initial trigger for doing a GC. The |
| 7038 |
default is 200 kb. |
| 7039 |
|
| 7040 |
Debugging operations for the freelist have been deprecated, along with |
| 7041 |
the C variables that control garbage collection. The environment |
| 7042 |
variables GUILE_MAX_SEGMENT_SIZE, GUILE_INIT_SEGMENT_SIZE_2, |
| 7043 |
GUILE_INIT_SEGMENT_SIZE_1, and GUILE_MIN_YIELD_2 should be used. |
| 7044 |
|
| 7045 |
For understanding the memory usage of a GUILE program, the routine |
| 7046 |
gc-live-object-stats returns an alist containing the number of live |
| 7047 |
objects for every type. |
| 7048 |
|
| 7049 |
|
| 7050 |
** The function scm_definedp has been renamed to scm_defined_p |
| 7051 |
|
| 7052 |
The name scm_definedp is deprecated. |
| 7053 |
|
| 7054 |
** The struct scm_cell type has been renamed to scm_t_cell |
| 7055 |
|
| 7056 |
This is in accordance to Guile's naming scheme for types. Note that |
| 7057 |
the name scm_cell is now used for a function that allocates and |
| 7058 |
initializes a new cell (see below). |
| 7059 |
|
| 7060 |
** New functions for memory management |
| 7061 |
|
| 7062 |
A new set of functions for memory management has been added since the |
| 7063 |
old way (scm_must_malloc, scm_must_free, etc) was error prone and |
| 7064 |
indeed, Guile itself contained some long standing bugs that could |
| 7065 |
cause aborts in long running programs. |
| 7066 |
|
| 7067 |
The new functions are more symmetrical and do not need cooperation |
| 7068 |
from smob free routines, among other improvements. |
| 7069 |
|
| 7070 |
The new functions are scm_malloc, scm_realloc, scm_calloc, scm_strdup, |
| 7071 |
scm_strndup, scm_gc_malloc, scm_gc_calloc, scm_gc_realloc, |
| 7072 |
scm_gc_free, scm_gc_register_collectable_memory, and |
| 7073 |
scm_gc_unregister_collectable_memory. Refer to the manual for more |
| 7074 |
details and for upgrading instructions. |
| 7075 |
|
| 7076 |
The old functions for memory management have been deprecated. They |
| 7077 |
are: scm_must_malloc, scm_must_realloc, scm_must_free, |
| 7078 |
scm_must_strdup, scm_must_strndup, scm_done_malloc, scm_done_free. |
| 7079 |
|
| 7080 |
** Declarations of exported features are marked with SCM_API. |
| 7081 |
|
| 7082 |
Every declaration of a feature that belongs to the exported Guile API |
| 7083 |
has been marked by adding the macro "SCM_API" to the start of the |
| 7084 |
declaration. This macro can expand into different things, the most |
| 7085 |
common of which is just "extern" for Unix platforms. On Win32, it can |
| 7086 |
be used to control which symbols are exported from a DLL. |
| 7087 |
|
| 7088 |
If you `#define SCM_IMPORT' before including <libguile.h>, SCM_API |
| 7089 |
will expand into "__declspec (dllimport) extern", which is needed for |
| 7090 |
linking to the Guile DLL in Windows. |
| 7091 |
|
| 7092 |
There are also SCM_RL_IMPORT, SCM_SRFI1314_IMPORT, and |
| 7093 |
SCM_SRFI4_IMPORT, for the corresponding libraries. |
| 7094 |
|
| 7095 |
** SCM_NEWCELL and SCM_NEWCELL2 have been deprecated. |
| 7096 |
|
| 7097 |
Use the new functions scm_cell and scm_double_cell instead. The old |
| 7098 |
macros had problems because with them allocation and initialization |
| 7099 |
was separated and the GC could sometimes observe half initialized |
| 7100 |
cells. Only careful coding by the user of SCM_NEWCELL and |
| 7101 |
SCM_NEWCELL2 could make this safe and efficient. |
| 7102 |
|
| 7103 |
** CHECK_ENTRY, CHECK_APPLY and CHECK_EXIT have been deprecated. |
| 7104 |
|
| 7105 |
Use the variables scm_check_entry_p, scm_check_apply_p and scm_check_exit_p |
| 7106 |
instead. |
| 7107 |
|
| 7108 |
** SRCBRKP has been deprecated. |
| 7109 |
|
| 7110 |
Use scm_c_source_property_breakpoint_p instead. |
| 7111 |
|
| 7112 |
** Deprecated: scm_makmacro |
| 7113 |
|
| 7114 |
Change your code to use either scm_makmmacro or to define macros in |
| 7115 |
Scheme, using 'define-macro'. |
| 7116 |
|
| 7117 |
** New function scm_c_port_for_each. |
| 7118 |
|
| 7119 |
This function is like scm_port_for_each but takes a pointer to a C |
| 7120 |
function as the callback instead of a SCM value. |
| 7121 |
|
| 7122 |
** The names scm_internal_select, scm_thread_sleep, and |
| 7123 |
scm_thread_usleep have been discouraged. |
| 7124 |
|
| 7125 |
Use scm_std_select, scm_std_sleep, scm_std_usleep instead. |
| 7126 |
|
| 7127 |
** The GC can no longer be blocked. |
| 7128 |
|
| 7129 |
The global flags scm_gc_heap_lock and scm_block_gc have been removed. |
| 7130 |
The GC can now run (partially) concurrently with other code and thus |
| 7131 |
blocking it is not well defined. |
| 7132 |
|
| 7133 |
** Many definitions have been removed that were previously deprecated. |
| 7134 |
|
| 7135 |
scm_lisp_nil, scm_lisp_t, s_nil_ify, scm_m_nil_ify, s_t_ify, |
| 7136 |
scm_m_t_ify, s_0_cond, scm_m_0_cond, s_0_ify, scm_m_0_ify, s_1_ify, |
| 7137 |
scm_m_1_ify, scm_debug_newcell, scm_debug_newcell2, |
| 7138 |
scm_tc16_allocated, SCM_SET_SYMBOL_HASH, SCM_IM_NIL_IFY, SCM_IM_T_IFY, |
| 7139 |
SCM_IM_0_COND, SCM_IM_0_IFY, SCM_IM_1_IFY, SCM_GC_SET_ALLOCATED, |
| 7140 |
scm_debug_newcell, scm_debug_newcell2, SCM_HUP_SIGNAL, SCM_INT_SIGNAL, |
| 7141 |
SCM_FPE_SIGNAL, SCM_BUS_SIGNAL, SCM_SEGV_SIGNAL, SCM_ALRM_SIGNAL, |
| 7142 |
SCM_GC_SIGNAL, SCM_TICK_SIGNAL, SCM_SIG_ORD, SCM_ORD_SIG, |
| 7143 |
SCM_NUM_SIGS, scm_top_level_lookup_closure_var, |
| 7144 |
*top-level-lookup-closure*, scm_system_transformer, scm_eval_3, |
| 7145 |
scm_eval2, root_module_lookup_closure, SCM_SLOPPY_STRINGP, |
| 7146 |
SCM_RWSTRINGP, scm_read_only_string_p, scm_make_shared_substring, |
| 7147 |
scm_tc7_substring, sym_huh, SCM_VARVCELL, SCM_UDVARIABLEP, |
| 7148 |
SCM_DEFVARIABLEP, scm_mkbig, scm_big2inum, scm_adjbig, scm_normbig, |
| 7149 |
scm_copybig, scm_2ulong2big, scm_dbl2big, scm_big2dbl, SCM_FIXNUM_BIT, |
| 7150 |
SCM_SETCHARS, SCM_SLOPPY_SUBSTRP, SCM_SUBSTR_STR, SCM_SUBSTR_OFFSET, |
| 7151 |
SCM_LENGTH_MAX, SCM_SETLENGTH, SCM_ROSTRINGP, SCM_ROLENGTH, |
| 7152 |
SCM_ROCHARS, SCM_ROUCHARS, SCM_SUBSTRP, SCM_COERCE_SUBSTR, |
| 7153 |
scm_sym2vcell, scm_intern, scm_intern0, scm_sysintern, scm_sysintern0, |
| 7154 |
scm_sysintern0_no_module_lookup, scm_init_symbols_deprecated, |
| 7155 |
scm_vector_set_length_x, scm_contregs, scm_debug_info, |
| 7156 |
scm_debug_frame, SCM_DSIDEVAL, SCM_CONST_LONG, SCM_VCELL, |
| 7157 |
SCM_GLOBAL_VCELL, SCM_VCELL_INIT, SCM_GLOBAL_VCELL_INIT, |
| 7158 |
SCM_HUGE_LENGTH, SCM_VALIDATE_STRINGORSUBSTR, SCM_VALIDATE_ROSTRING, |
| 7159 |
SCM_VALIDATE_ROSTRING_COPY, SCM_VALIDATE_NULLORROSTRING_COPY, |
| 7160 |
SCM_VALIDATE_RWSTRING, DIGITS, scm_small_istr2int, scm_istr2int, |
| 7161 |
scm_istr2flo, scm_istring2number, scm_istr2int, scm_istr2flo, |
| 7162 |
scm_istring2number, scm_vtable_index_vcell, scm_si_vcell, SCM_ECONSP, |
| 7163 |
SCM_NECONSP, SCM_GLOC_VAR, SCM_GLOC_VAL, SCM_GLOC_SET_VAL, |
| 7164 |
SCM_GLOC_VAL_LOC, scm_make_gloc, scm_gloc_p, scm_tc16_variable, |
| 7165 |
SCM_CHARS, SCM_LENGTH, SCM_SET_STRING_CHARS, SCM_SET_STRING_LENGTH. |
| 7166 |
|
| 7167 |
* Changes to bundled modules |
| 7168 |
|
| 7169 |
** (ice-9 debug) |
| 7170 |
|
| 7171 |
Using the (ice-9 debug) module no longer automatically switches Guile |
| 7172 |
to use the debugging evaluator. If you want to switch to the |
| 7173 |
debugging evaluator (which is needed for backtrace information if you |
| 7174 |
hit an error), please add an explicit "(debug-enable 'debug)" to your |
| 7175 |
code just after the code to use (ice-9 debug). |
| 7176 |
|
| 7177 |
|
| 7178 |
Changes since Guile 1.4: |
| 7179 |
|
| 7180 |
* Changes to the distribution |
| 7181 |
|
| 7182 |
** A top-level TODO file is included. |
| 7183 |
|
| 7184 |
** Guile now uses a versioning scheme similar to that of the Linux kernel. |
| 7185 |
|
| 7186 |
Guile now always uses three numbers to represent the version, |
| 7187 |
i.e. "1.6.5". The first number, 1, is the major version number, the |
| 7188 |
second number, 6, is the minor version number, and the third number, |
| 7189 |
5, is the micro version number. Changes in major version number |
| 7190 |
indicate major changes in Guile. |
| 7191 |
|
| 7192 |
Minor version numbers that are even denote stable releases, and odd |
| 7193 |
minor version numbers denote development versions (which may be |
| 7194 |
unstable). The micro version number indicates a minor sub-revision of |
| 7195 |
a given MAJOR.MINOR release. |
| 7196 |
|
| 7197 |
In keeping with the new scheme, (minor-version) and scm_minor_version |
| 7198 |
no longer return everything but the major version number. They now |
| 7199 |
just return the minor version number. Two new functions |
| 7200 |
(micro-version) and scm_micro_version have been added to report the |
| 7201 |
micro version number. |
| 7202 |
|
| 7203 |
In addition, ./GUILE-VERSION now defines GUILE_MICRO_VERSION. |
| 7204 |
|
| 7205 |
** New preprocessor definitions are available for checking versions. |
| 7206 |
|
| 7207 |
version.h now #defines SCM_MAJOR_VERSION, SCM_MINOR_VERSION, and |
| 7208 |
SCM_MICRO_VERSION to the appropriate integer values. |
| 7209 |
|
| 7210 |
** Guile now actively warns about deprecated features. |
| 7211 |
|
| 7212 |
The new configure option `--enable-deprecated=LEVEL' and the |
| 7213 |
environment variable GUILE_WARN_DEPRECATED control this mechanism. |
| 7214 |
See INSTALL and README for more information. |
| 7215 |
|
| 7216 |
** Guile is much more likely to work on 64-bit architectures. |
| 7217 |
|
| 7218 |
Guile now compiles and passes "make check" with only two UNRESOLVED GC |
| 7219 |
cases on Alpha and ia64 based machines now. Thanks to John Goerzen |
| 7220 |
for the use of a test machine, and thanks to Stefan Jahn for ia64 |
| 7221 |
patches. |
| 7222 |
|
| 7223 |
** New functions: setitimer and getitimer. |
| 7224 |
|
| 7225 |
These implement a fairly direct interface to the libc functions of the |
| 7226 |
same name. |
| 7227 |
|
| 7228 |
** The #. reader extension is now disabled by default. |
| 7229 |
|
| 7230 |
For safety reasons, #. evaluation is disabled by default. To |
| 7231 |
re-enable it, set the fluid read-eval? to #t. For example: |
| 7232 |
|
| 7233 |
(fluid-set! read-eval? #t) |
| 7234 |
|
| 7235 |
but make sure you realize the potential security risks involved. With |
| 7236 |
read-eval? enabled, reading a data file from an untrusted source can |
| 7237 |
be dangerous. |
| 7238 |
|
| 7239 |
** New SRFI modules have been added: |
| 7240 |
|
| 7241 |
SRFI-0 `cond-expand' is now supported in Guile, without requiring |
| 7242 |
using a module. |
| 7243 |
|
| 7244 |
(srfi srfi-1) is a library containing many useful pair- and list-processing |
| 7245 |
procedures. |
| 7246 |
|
| 7247 |
(srfi srfi-2) exports and-let*. |
| 7248 |
|
| 7249 |
(srfi srfi-4) implements homogeneous numeric vector datatypes. |
| 7250 |
|
| 7251 |
(srfi srfi-6) is a dummy module for now, since guile already provides |
| 7252 |
all of the srfi-6 procedures by default: open-input-string, |
| 7253 |
open-output-string, get-output-string. |
| 7254 |
|
| 7255 |
(srfi srfi-8) exports receive. |
| 7256 |
|
| 7257 |
(srfi srfi-9) exports define-record-type. |
| 7258 |
|
| 7259 |
(srfi srfi-10) exports define-reader-ctor and implements the reader |
| 7260 |
extension #,(). |
| 7261 |
|
| 7262 |
(srfi srfi-11) exports let-values and let*-values. |
| 7263 |
|
| 7264 |
(srfi srfi-13) implements the SRFI String Library. |
| 7265 |
|
| 7266 |
(srfi srfi-14) implements the SRFI Character-Set Library. |
| 7267 |
|
| 7268 |
(srfi srfi-17) implements setter and getter-with-setter and redefines |
| 7269 |
some accessor procedures as procedures with getters. (such as car, |
| 7270 |
cdr, vector-ref etc.) |
| 7271 |
|
| 7272 |
(srfi srfi-19) implements the SRFI Time/Date Library. |
| 7273 |
|
| 7274 |
** New scripts / "executable modules" |
| 7275 |
|
| 7276 |
Subdirectory "scripts" contains Scheme modules that are packaged to |
| 7277 |
also be executable as scripts. At this time, these scripts are available: |
| 7278 |
|
| 7279 |
display-commentary |
| 7280 |
doc-snarf |
| 7281 |
generate-autoload |
| 7282 |
punify |
| 7283 |
read-scheme-source |
| 7284 |
use2dot |
| 7285 |
|
| 7286 |
See README there for more info. |
| 7287 |
|
| 7288 |
These scripts can be invoked from the shell with the new program |
| 7289 |
"guile-tools", which keeps track of installation directory for you. |
| 7290 |
For example: |
| 7291 |
|
| 7292 |
$ guile-tools display-commentary srfi/*.scm |
| 7293 |
|
| 7294 |
guile-tools is copied to the standard $bindir on "make install". |
| 7295 |
|
| 7296 |
** New module (ice-9 stack-catch): |
| 7297 |
|
| 7298 |
stack-catch is like catch, but saves the current state of the stack in |
| 7299 |
the fluid the-last-stack. This fluid can be useful when using the |
| 7300 |
debugger and when re-throwing an error. |
| 7301 |
|
| 7302 |
** The module (ice-9 and-let*) has been renamed to (ice-9 and-let-star) |
| 7303 |
|
| 7304 |
This has been done to prevent problems on lesser operating systems |
| 7305 |
that can't tolerate `*'s in file names. The exported macro continues |
| 7306 |
to be named `and-let*', of course. |
| 7307 |
|
| 7308 |
On systems that support it, there is also a compatibility module named |
| 7309 |
(ice-9 and-let*). It will go away in the next release. |
| 7310 |
|
| 7311 |
** New modules (oop goops) etc.: |
| 7312 |
|
| 7313 |
(oop goops) |
| 7314 |
(oop goops describe) |
| 7315 |
(oop goops save) |
| 7316 |
(oop goops active-slot) |
| 7317 |
(oop goops composite-slot) |
| 7318 |
|
| 7319 |
The Guile Object Oriented Programming System (GOOPS) has been |
| 7320 |
integrated into Guile. For further information, consult the GOOPS |
| 7321 |
manual and tutorial in the `doc' directory. |
| 7322 |
|
| 7323 |
** New module (ice-9 rdelim). |
| 7324 |
|
| 7325 |
This exports the following procedures which were previously defined |
| 7326 |
in the default environment: |
| 7327 |
|
| 7328 |
read-line read-line! read-delimited read-delimited! %read-delimited! |
| 7329 |
%read-line write-line |
| 7330 |
|
| 7331 |
For backwards compatibility the definitions are still imported into the |
| 7332 |
default environment in this version of Guile. However you should add: |
| 7333 |
|
| 7334 |
(use-modules (ice-9 rdelim)) |
| 7335 |
|
| 7336 |
to any program which uses the definitions, since this may change in |
| 7337 |
future. |
| 7338 |
|
| 7339 |
Alternatively, if guile-scsh is installed, the (scsh rdelim) module |
| 7340 |
can be used for similar functionality. |
| 7341 |
|
| 7342 |
** New module (ice-9 rw) |
| 7343 |
|
| 7344 |
This is a subset of the (scsh rw) module from guile-scsh. Currently |
| 7345 |
it defines two procedures: |
| 7346 |
|
| 7347 |
*** New function: read-string!/partial str [port_or_fdes [start [end]]] |
| 7348 |
|
| 7349 |
Read characters from a port or file descriptor into a string STR. |
| 7350 |
A port must have an underlying file descriptor -- a so-called |
| 7351 |
fport. This procedure is scsh-compatible and can efficiently read |
| 7352 |
large strings. |
| 7353 |
|
| 7354 |
*** New function: write-string/partial str [port_or_fdes [start [end]]] |
| 7355 |
|
| 7356 |
Write characters from a string STR to a port or file descriptor. |
| 7357 |
A port must have an underlying file descriptor -- a so-called |
| 7358 |
fport. This procedure is mostly compatible and can efficiently |
| 7359 |
write large strings. |
| 7360 |
|
| 7361 |
** New module (ice-9 match) |
| 7362 |
|
| 7363 |
This module includes Andrew K. Wright's pattern matcher. See |
| 7364 |
ice-9/match.scm for brief description or |
| 7365 |
|
| 7366 |
http://www.star-lab.com/wright/code.html |
| 7367 |
|
| 7368 |
for complete documentation. |
| 7369 |
|
| 7370 |
** New module (ice-9 buffered-input) |
| 7371 |
|
| 7372 |
This module provides procedures to construct an input port from an |
| 7373 |
underlying source of input that reads and returns its input in chunks. |
| 7374 |
The underlying input source is a Scheme procedure, specified by the |
| 7375 |
caller, which the port invokes whenever it needs more input. |
| 7376 |
|
| 7377 |
This is useful when building an input port whose back end is Readline |
| 7378 |
or a UI element such as the GtkEntry widget. |
| 7379 |
|
| 7380 |
** Documentation |
| 7381 |
|
| 7382 |
The reference and tutorial documentation that was previously |
| 7383 |
distributed separately, as `guile-doc', is now included in the core |
| 7384 |
Guile distribution. The documentation consists of the following |
| 7385 |
manuals. |
| 7386 |
|
| 7387 |
- The Guile Tutorial (guile-tut.texi) contains a tutorial introduction |
| 7388 |
to using Guile. |
| 7389 |
|
| 7390 |
- The Guile Reference Manual (guile.texi) contains (or is intended to |
| 7391 |
contain) reference documentation on all aspects of Guile. |
| 7392 |
|
| 7393 |
- The GOOPS Manual (goops.texi) contains both tutorial-style and |
| 7394 |
reference documentation for using GOOPS, Guile's Object Oriented |
| 7395 |
Programming System. |
| 7396 |
|
| 7397 |
- The Revised^5 Report on the Algorithmic Language Scheme |
| 7398 |
(r5rs.texi). |
| 7399 |
|
| 7400 |
See the README file in the `doc' directory for more details. |
| 7401 |
|
| 7402 |
** There are a couple of examples in the examples/ directory now. |
| 7403 |
|
| 7404 |
* Changes to the stand-alone interpreter |
| 7405 |
|
| 7406 |
** New command line option `--use-srfi' |
| 7407 |
|
| 7408 |
Using this option, SRFI modules can be loaded on startup and be |
| 7409 |
available right from the beginning. This makes programming portable |
| 7410 |
Scheme programs easier. |
| 7411 |
|
| 7412 |
The option `--use-srfi' expects a comma-separated list of numbers, |
| 7413 |
each representing a SRFI number to be loaded into the interpreter |
| 7414 |
before starting evaluating a script file or the REPL. Additionally, |
| 7415 |
the feature identifier for the loaded SRFIs is recognized by |
| 7416 |
`cond-expand' when using this option. |
| 7417 |
|
| 7418 |
Example: |
| 7419 |
$ guile --use-srfi=8,13 |
| 7420 |
guile> (receive (x z) (values 1 2) (+ 1 2)) |
| 7421 |
3 |
| 7422 |
guile> (string-pad "bla" 20) |
| 7423 |
" bla" |
| 7424 |
|
| 7425 |
** Guile now always starts up in the `(guile-user)' module. |
| 7426 |
|
| 7427 |
Previously, scripts executed via the `-s' option would run in the |
| 7428 |
`(guile)' module and the repl would run in the `(guile-user)' module. |
| 7429 |
Now every user action takes place in the `(guile-user)' module by |
| 7430 |
default. |
| 7431 |
|
| 7432 |
* Changes to Scheme functions and syntax |
| 7433 |
|
| 7434 |
** Character classifiers work for non-ASCII characters. |
| 7435 |
|
| 7436 |
The predicates `char-alphabetic?', `char-numeric?', |
| 7437 |
`char-whitespace?', `char-lower?', `char-upper?' and `char-is-both?' |
| 7438 |
no longer check whether their arguments are ASCII characters. |
| 7439 |
Previously, a character would only be considered alphabetic when it |
| 7440 |
was also ASCII, for example. |
| 7441 |
|
| 7442 |
** Previously deprecated Scheme functions have been removed: |
| 7443 |
|
| 7444 |
tag - no replacement. |
| 7445 |
fseek - replaced by seek. |
| 7446 |
list* - replaced by cons*. |
| 7447 |
|
| 7448 |
** It's now possible to create modules with controlled environments |
| 7449 |
|
| 7450 |
Example: |
| 7451 |
|
| 7452 |
(use-modules (ice-9 safe)) |
| 7453 |
(define m (make-safe-module)) |
| 7454 |
;;; m will now be a module containing only a safe subset of R5RS |
| 7455 |
(eval '(+ 1 2) m) --> 3 |
| 7456 |
(eval 'load m) --> ERROR: Unbound variable: load |
| 7457 |
|
| 7458 |
** Evaluation of "()", the empty list, is now an error. |
| 7459 |
|
| 7460 |
Previously, the expression "()" evaluated to the empty list. This has |
| 7461 |
been changed to signal a "missing expression" error. The correct way |
| 7462 |
to write the empty list as a literal constant is to use quote: "'()". |
| 7463 |
|
| 7464 |
** New concept of `Guile Extensions'. |
| 7465 |
|
| 7466 |
A Guile Extension is just a ordinary shared library that can be linked |
| 7467 |
at run-time. We found it advantageous to give this simple concept a |
| 7468 |
dedicated name to distinguish the issues related to shared libraries |
| 7469 |
from the issues related to the module system. |
| 7470 |
|
| 7471 |
*** New function: load-extension |
| 7472 |
|
| 7473 |
Executing (load-extension lib init) is mostly equivalent to |
| 7474 |
|
| 7475 |
(dynamic-call init (dynamic-link lib)) |
| 7476 |
|
| 7477 |
except when scm_register_extension has been called previously. |
| 7478 |
Whenever appropriate, you should use `load-extension' instead of |
| 7479 |
dynamic-link and dynamic-call. |
| 7480 |
|
| 7481 |
*** New C function: scm_c_register_extension |
| 7482 |
|
| 7483 |
This function registers a initialization function for use by |
| 7484 |
`load-extension'. Use it when you don't want specific extensions to |
| 7485 |
be loaded as shared libraries (for example on platforms that don't |
| 7486 |
support dynamic linking). |
| 7487 |
|
| 7488 |
** Auto-loading of compiled-code modules is deprecated. |
| 7489 |
|
| 7490 |
Guile used to be able to automatically find and link a shared |
| 7491 |
library to satisfy requests for a module. For example, the module |
| 7492 |
`(foo bar)' could be implemented by placing a shared library named |
| 7493 |
"foo/libbar.so" (or with a different extension) in a directory on the |
| 7494 |
load path of Guile. |
| 7495 |
|
| 7496 |
This has been found to be too tricky, and is no longer supported. The |
| 7497 |
shared libraries are now called "extensions". You should now write a |
| 7498 |
small Scheme file that calls `load-extension' to load the shared |
| 7499 |
library and initialize it explicitly. |
| 7500 |
|
| 7501 |
The shared libraries themselves should be installed in the usual |
| 7502 |
places for shared libraries, with names like "libguile-foo-bar". |
| 7503 |
|
| 7504 |
For example, place this into a file "foo/bar.scm" |
| 7505 |
|
| 7506 |
(define-module (foo bar)) |
| 7507 |
|
| 7508 |
(load-extension "libguile-foo-bar" "foobar_init") |
| 7509 |
|
| 7510 |
** Backward incompatible change: eval EXP ENVIRONMENT-SPECIFIER |
| 7511 |
|
| 7512 |
`eval' is now R5RS, that is it takes two arguments. |
| 7513 |
The second argument is an environment specifier, i.e. either |
| 7514 |
|
| 7515 |
(scheme-report-environment 5) |
| 7516 |
(null-environment 5) |
| 7517 |
(interaction-environment) |
| 7518 |
|
| 7519 |
or |
| 7520 |
|
| 7521 |
any module. |
| 7522 |
|
| 7523 |
** The module system has been made more disciplined. |
| 7524 |
|
| 7525 |
The function `eval' will save and restore the current module around |
| 7526 |
the evaluation of the specified expression. While this expression is |
| 7527 |
evaluated, `(current-module)' will now return the right module, which |
| 7528 |
is the module specified as the second argument to `eval'. |
| 7529 |
|
| 7530 |
A consequence of this change is that `eval' is not particularly |
| 7531 |
useful when you want allow the evaluated code to change what module is |
| 7532 |
designated as the current module and have this change persist from one |
| 7533 |
call to `eval' to the next. The read-eval-print-loop is an example |
| 7534 |
where `eval' is now inadequate. To compensate, there is a new |
| 7535 |
function `primitive-eval' that does not take a module specifier and |
| 7536 |
that does not save/restore the current module. You should use this |
| 7537 |
function together with `set-current-module', `current-module', etc |
| 7538 |
when you want to have more control over the state that is carried from |
| 7539 |
one eval to the next. |
| 7540 |
|
| 7541 |
Additionally, it has been made sure that forms that are evaluated at |
| 7542 |
the top level are always evaluated with respect to the current module. |
| 7543 |
Previously, subforms of top-level forms such as `begin', `case', |
| 7544 |
etc. did not respect changes to the current module although these |
| 7545 |
subforms are at the top-level as well. |
| 7546 |
|
| 7547 |
To prevent strange behavior, the forms `define-module', |
| 7548 |
`use-modules', `use-syntax', and `export' have been restricted to only |
| 7549 |
work on the top level. The forms `define-public' and |
| 7550 |
`defmacro-public' only export the new binding on the top level. They |
| 7551 |
behave just like `define' and `defmacro', respectively, when they are |
| 7552 |
used in a lexical environment. |
| 7553 |
|
| 7554 |
Also, `export' will no longer silently re-export bindings imported |
| 7555 |
from a used module. It will emit a `deprecation' warning and will |
| 7556 |
cease to perform any re-export in the next version. If you actually |
| 7557 |
want to re-export bindings, use the new `re-export' in place of |
| 7558 |
`export'. The new `re-export' will not make copies of variables when |
| 7559 |
rexporting them, as `export' did wrongly. |
| 7560 |
|
| 7561 |
** Module system now allows selection and renaming of imported bindings |
| 7562 |
|
| 7563 |
Previously, when using `use-modules' or the `#:use-module' clause in |
| 7564 |
the `define-module' form, all the bindings (association of symbols to |
| 7565 |
values) for imported modules were added to the "current module" on an |
| 7566 |
as-is basis. This has been changed to allow finer control through two |
| 7567 |
new facilities: selection and renaming. |
| 7568 |
|
| 7569 |
You can now select which of the imported module's bindings are to be |
| 7570 |
visible in the current module by using the `:select' clause. This |
| 7571 |
clause also can be used to rename individual bindings. For example: |
| 7572 |
|
| 7573 |
;; import all bindings no questions asked |
| 7574 |
(use-modules (ice-9 common-list)) |
| 7575 |
|
| 7576 |
;; import four bindings, renaming two of them; |
| 7577 |
;; the current module sees: every some zonk-y zonk-n |
| 7578 |
(use-modules ((ice-9 common-list) |
| 7579 |
:select (every some |
| 7580 |
(remove-if . zonk-y) |
| 7581 |
(remove-if-not . zonk-n)))) |
| 7582 |
|
| 7583 |
You can also programmatically rename all selected bindings using the |
| 7584 |
`:renamer' clause, which specifies a proc that takes a symbol and |
| 7585 |
returns another symbol. Because it is common practice to use a prefix, |
| 7586 |
we now provide the convenience procedure `symbol-prefix-proc'. For |
| 7587 |
example: |
| 7588 |
|
| 7589 |
;; import four bindings, renaming two of them specifically, |
| 7590 |
;; and all four w/ prefix "CL:"; |
| 7591 |
;; the current module sees: CL:every CL:some CL:zonk-y CL:zonk-n |
| 7592 |
(use-modules ((ice-9 common-list) |
| 7593 |
:select (every some |
| 7594 |
(remove-if . zonk-y) |
| 7595 |
(remove-if-not . zonk-n)) |
| 7596 |
:renamer (symbol-prefix-proc 'CL:))) |
| 7597 |
|
| 7598 |
;; import four bindings, renaming two of them specifically, |
| 7599 |
;; and all four by upcasing. |
| 7600 |
;; the current module sees: EVERY SOME ZONK-Y ZONK-N |
| 7601 |
(define (upcase-symbol sym) |
| 7602 |
(string->symbol (string-upcase (symbol->string sym)))) |
| 7603 |
|
| 7604 |
(use-modules ((ice-9 common-list) |
| 7605 |
:select (every some |
| 7606 |
(remove-if . zonk-y) |
| 7607 |
(remove-if-not . zonk-n)) |
| 7608 |
:renamer upcase-symbol)) |
| 7609 |
|
| 7610 |
Note that programmatic renaming is done *after* individual renaming. |
| 7611 |
Also, the above examples show `use-modules', but the same facilities are |
| 7612 |
available for the `#:use-module' clause of `define-module'. |
| 7613 |
|
| 7614 |
See manual for more info. |
| 7615 |
|
| 7616 |
** The semantics of guardians have changed. |
| 7617 |
|
| 7618 |
The changes are for the most part compatible. An important criterion |
| 7619 |
was to keep the typical usage of guardians as simple as before, but to |
| 7620 |
make the semantics safer and (as a result) more useful. |
| 7621 |
|
| 7622 |
*** All objects returned from guardians are now properly alive. |
| 7623 |
|
| 7624 |
It is now guaranteed that any object referenced by an object returned |
| 7625 |
from a guardian is alive. It's now impossible for a guardian to |
| 7626 |
return a "contained" object before its "containing" object. |
| 7627 |
|
| 7628 |
One incompatible (but probably not very important) change resulting |
| 7629 |
from this is that it is no longer possible to guard objects that |
| 7630 |
indirectly reference themselves (i.e. are parts of cycles). If you do |
| 7631 |
so accidentally, you'll get a warning. |
| 7632 |
|
| 7633 |
*** There are now two types of guardians: greedy and sharing. |
| 7634 |
|
| 7635 |
If you call (make-guardian #t) or just (make-guardian), you'll get a |
| 7636 |
greedy guardian, and for (make-guardian #f) a sharing guardian. |
| 7637 |
|
| 7638 |
Greedy guardians are the default because they are more "defensive". |
| 7639 |
You can only greedily guard an object once. If you guard an object |
| 7640 |
more than once, once in a greedy guardian and the rest of times in |
| 7641 |
sharing guardians, then it is guaranteed that the object won't be |
| 7642 |
returned from sharing guardians as long as it is greedily guarded |
| 7643 |
and/or alive. |
| 7644 |
|
| 7645 |
Guardians returned by calls to `make-guardian' can now take one more |
| 7646 |
optional parameter, which says whether to throw an error in case an |
| 7647 |
attempt is made to greedily guard an object that is already greedily |
| 7648 |
guarded. The default is true, i.e. throw an error. If the parameter |
| 7649 |
is false, the guardian invocation returns #t if guarding was |
| 7650 |
successful and #f if it wasn't. |
| 7651 |
|
| 7652 |
Also, since greedy guarding is, in effect, a side-effecting operation |
| 7653 |
on objects, a new function is introduced: `destroy-guardian!'. |
| 7654 |
Invoking this function on a guardian renders it unoperative and, if |
| 7655 |
the guardian is greedy, clears the "greedily guarded" property of the |
| 7656 |
objects that were guarded by it, thus undoing the side effect. |
| 7657 |
|
| 7658 |
Note that all this hair is hardly very important, since guardian |
| 7659 |
objects are usually permanent. |
| 7660 |
|
| 7661 |
** Continuations created by call-with-current-continuation now accept |
| 7662 |
any number of arguments, as required by R5RS. |
| 7663 |
|
| 7664 |
** New function `issue-deprecation-warning' |
| 7665 |
|
| 7666 |
This function is used to display the deprecation messages that are |
| 7667 |
controlled by GUILE_WARN_DEPRECATION as explained in the README. |
| 7668 |
|
| 7669 |
(define (id x) |
| 7670 |
(issue-deprecation-warning "`id' is deprecated. Use `identity' instead.") |
| 7671 |
(identity x)) |
| 7672 |
|
| 7673 |
guile> (id 1) |
| 7674 |
;; `id' is deprecated. Use `identity' instead. |
| 7675 |
1 |
| 7676 |
guile> (id 1) |
| 7677 |
1 |
| 7678 |
|
| 7679 |
** New syntax `begin-deprecated' |
| 7680 |
|
| 7681 |
When deprecated features are included (as determined by the configure |
| 7682 |
option --enable-deprecated), `begin-deprecated' is identical to |
| 7683 |
`begin'. When deprecated features are excluded, it always evaluates |
| 7684 |
to `#f', ignoring the body forms. |
| 7685 |
|
| 7686 |
** New function `make-object-property' |
| 7687 |
|
| 7688 |
This function returns a new `procedure with setter' P that can be used |
| 7689 |
to attach a property to objects. When calling P as |
| 7690 |
|
| 7691 |
(set! (P obj) val) |
| 7692 |
|
| 7693 |
where `obj' is any kind of object, it attaches `val' to `obj' in such |
| 7694 |
a way that it can be retrieved by calling P as |
| 7695 |
|
| 7696 |
(P obj) |
| 7697 |
|
| 7698 |
This function will replace procedure properties, symbol properties and |
| 7699 |
source properties eventually. |
| 7700 |
|
| 7701 |
** Module (ice-9 optargs) now uses keywords instead of `#&'. |
| 7702 |
|
| 7703 |
Instead of #&optional, #&key, etc you should now use #:optional, |
| 7704 |
#:key, etc. Since #:optional is a keyword, you can write it as just |
| 7705 |
:optional when (read-set! keywords 'prefix) is active. |
| 7706 |
|
| 7707 |
The old reader syntax `#&' is still supported, but deprecated. It |
| 7708 |
will be removed in the next release. |
| 7709 |
|
| 7710 |
** New define-module option: pure |
| 7711 |
|
| 7712 |
Tells the module system not to include any bindings from the root |
| 7713 |
module. |
| 7714 |
|
| 7715 |
Example: |
| 7716 |
|
| 7717 |
(define-module (totally-empty-module) |
| 7718 |
:pure) |
| 7719 |
|
| 7720 |
** New define-module option: export NAME1 ... |
| 7721 |
|
| 7722 |
Export names NAME1 ... |
| 7723 |
|
| 7724 |
This option is required if you want to be able to export bindings from |
| 7725 |
a module which doesn't import one of `define-public' or `export'. |
| 7726 |
|
| 7727 |
Example: |
| 7728 |
|
| 7729 |
(define-module (foo) |
| 7730 |
:pure |
| 7731 |
:use-module (ice-9 r5rs) |
| 7732 |
:export (bar)) |
| 7733 |
|
| 7734 |
;;; Note that we're pure R5RS below this point! |
| 7735 |
|
| 7736 |
(define (bar) |
| 7737 |
...) |
| 7738 |
|
| 7739 |
** New function: object->string OBJ |
| 7740 |
|
| 7741 |
Return a Scheme string obtained by printing a given object. |
| 7742 |
|
| 7743 |
** New function: port? X |
| 7744 |
|
| 7745 |
Returns a boolean indicating whether X is a port. Equivalent to |
| 7746 |
`(or (input-port? X) (output-port? X))'. |
| 7747 |
|
| 7748 |
** New function: file-port? |
| 7749 |
|
| 7750 |
Determines whether a given object is a port that is related to a file. |
| 7751 |
|
| 7752 |
** New function: port-for-each proc |
| 7753 |
|
| 7754 |
Apply PROC to each port in the Guile port table in turn. The return |
| 7755 |
value is unspecified. More specifically, PROC is applied exactly once |
| 7756 |
to every port that exists in the system at the time PORT-FOR-EACH is |
| 7757 |
invoked. Changes to the port table while PORT-FOR-EACH is running |
| 7758 |
have no effect as far as PORT-FOR-EACH is concerned. |
| 7759 |
|
| 7760 |
** New function: dup2 oldfd newfd |
| 7761 |
|
| 7762 |
A simple wrapper for the `dup2' system call. Copies the file |
| 7763 |
descriptor OLDFD to descriptor number NEWFD, replacing the |
| 7764 |
previous meaning of NEWFD. Both OLDFD and NEWFD must be integers. |
| 7765 |
Unlike for dup->fdes or primitive-move->fdes, no attempt is made |
| 7766 |
to move away ports which are using NEWFD. The return value is |
| 7767 |
unspecified. |
| 7768 |
|
| 7769 |
** New function: close-fdes fd |
| 7770 |
|
| 7771 |
A simple wrapper for the `close' system call. Close file |
| 7772 |
descriptor FD, which must be an integer. Unlike close (*note |
| 7773 |
close: Ports and File Descriptors.), the file descriptor will be |
| 7774 |
closed even if a port is using it. The return value is |
| 7775 |
unspecified. |
| 7776 |
|
| 7777 |
** New function: crypt password salt |
| 7778 |
|
| 7779 |
Encrypts `password' using the standard unix password encryption |
| 7780 |
algorithm. |
| 7781 |
|
| 7782 |
** New function: chroot path |
| 7783 |
|
| 7784 |
Change the root directory of the running process to `path'. |
| 7785 |
|
| 7786 |
** New functions: getlogin, cuserid |
| 7787 |
|
| 7788 |
Return the login name or the user name of the current effective user |
| 7789 |
id, respectively. |
| 7790 |
|
| 7791 |
** New functions: getpriority which who, setpriority which who prio |
| 7792 |
|
| 7793 |
Get or set the priority of the running process. |
| 7794 |
|
| 7795 |
** New function: getpass prompt |
| 7796 |
|
| 7797 |
Read a password from the terminal, first displaying `prompt' and |
| 7798 |
disabling echoing. |
| 7799 |
|
| 7800 |
** New function: flock file operation |
| 7801 |
|
| 7802 |
Set/remove an advisory shared or exclusive lock on `file'. |
| 7803 |
|
| 7804 |
** New functions: sethostname name, gethostname |
| 7805 |
|
| 7806 |
Set or get the hostname of the machine the current process is running |
| 7807 |
on. |
| 7808 |
|
| 7809 |
** New function: mkstemp! tmpl |
| 7810 |
|
| 7811 |
mkstemp creates a new unique file in the file system and returns a |
| 7812 |
new buffered port open for reading and writing to the file. TMPL |
| 7813 |
is a string specifying where the file should be created: it must |
| 7814 |
end with `XXXXXX' and will be changed in place to return the name |
| 7815 |
of the temporary file. |
| 7816 |
|
| 7817 |
** New function: open-input-string string |
| 7818 |
|
| 7819 |
Return an input string port which delivers the characters from |
| 7820 |
`string'. This procedure, together with `open-output-string' and |
| 7821 |
`get-output-string' implements SRFI-6. |
| 7822 |
|
| 7823 |
** New function: open-output-string |
| 7824 |
|
| 7825 |
Return an output string port which collects all data written to it. |
| 7826 |
The data can then be retrieved by `get-output-string'. |
| 7827 |
|
| 7828 |
** New function: get-output-string |
| 7829 |
|
| 7830 |
Return the contents of an output string port. |
| 7831 |
|
| 7832 |
** New function: identity |
| 7833 |
|
| 7834 |
Return the argument. |
| 7835 |
|
| 7836 |
** socket, connect, accept etc., now have support for IPv6. IPv6 addresses |
| 7837 |
are represented in Scheme as integers with normal host byte ordering. |
| 7838 |
|
| 7839 |
** New function: inet-pton family address |
| 7840 |
|
| 7841 |
Convert a printable string network address into an integer. Note that |
| 7842 |
unlike the C version of this function, the result is an integer with |
| 7843 |
normal host byte ordering. FAMILY can be `AF_INET' or `AF_INET6'. |
| 7844 |
e.g., |
| 7845 |
|
| 7846 |
(inet-pton AF_INET "127.0.0.1") => 2130706433 |
| 7847 |
(inet-pton AF_INET6 "::1") => 1 |
| 7848 |
|
| 7849 |
** New function: inet-ntop family address |
| 7850 |
|
| 7851 |
Convert an integer network address into a printable string. Note that |
| 7852 |
unlike the C version of this function, the input is an integer with |
| 7853 |
normal host byte ordering. FAMILY can be `AF_INET' or `AF_INET6'. |
| 7854 |
e.g., |
| 7855 |
|
| 7856 |
(inet-ntop AF_INET 2130706433) => "127.0.0.1" |
| 7857 |
(inet-ntop AF_INET6 (- (expt 2 128) 1)) => |
| 7858 |
ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff |
| 7859 |
|
| 7860 |
** Deprecated: id |
| 7861 |
|
| 7862 |
Use `identity' instead. |
| 7863 |
|
| 7864 |
** Deprecated: -1+ |
| 7865 |
|
| 7866 |
Use `1-' instead. |
| 7867 |
|
| 7868 |
** Deprecated: return-it |
| 7869 |
|
| 7870 |
Do without it. |
| 7871 |
|
| 7872 |
** Deprecated: string-character-length |
| 7873 |
|
| 7874 |
Use `string-length' instead. |
| 7875 |
|
| 7876 |
** Deprecated: flags |
| 7877 |
|
| 7878 |
Use `logior' instead. |
| 7879 |
|
| 7880 |
** Deprecated: close-all-ports-except. |
| 7881 |
|
| 7882 |
This was intended for closing ports in a child process after a fork, |
| 7883 |
but it has the undesirable side effect of flushing buffers. |
| 7884 |
port-for-each is more flexible. |
| 7885 |
|
| 7886 |
** The (ice-9 popen) module now attempts to set up file descriptors in |
| 7887 |
the child process from the current Scheme ports, instead of using the |
| 7888 |
current values of file descriptors 0, 1, and 2 in the parent process. |
| 7889 |
|
| 7890 |
** Removed function: builtin-weak-bindings |
| 7891 |
|
| 7892 |
There is no such concept as a weak binding any more. |
| 7893 |
|
| 7894 |
** Removed constants: bignum-radix, scm-line-incrementors |
| 7895 |
|
| 7896 |
** define-method: New syntax mandatory. |
| 7897 |
|
| 7898 |
The new method syntax is now mandatory: |
| 7899 |
|
| 7900 |
(define-method (NAME ARG-SPEC ...) BODY ...) |
| 7901 |
(define-method (NAME ARG-SPEC ... . REST-ARG) BODY ...) |
| 7902 |
|
| 7903 |
ARG-SPEC ::= ARG-NAME | (ARG-NAME TYPE) |
| 7904 |
REST-ARG ::= ARG-NAME |
| 7905 |
|
| 7906 |
If you have old code using the old syntax, import |
| 7907 |
(oop goops old-define-method) before (oop goops) as in: |
| 7908 |
|
| 7909 |
(use-modules (oop goops old-define-method) (oop goops)) |
| 7910 |
|
| 7911 |
** Deprecated function: builtin-variable |
| 7912 |
Removed function: builtin-bindings |
| 7913 |
|
| 7914 |
There is no longer a distinction between builtin or other variables. |
| 7915 |
Use module system operations for all variables. |
| 7916 |
|
| 7917 |
** Lazy-catch handlers are no longer allowed to return. |
| 7918 |
|
| 7919 |
That is, a call to `throw', `error', etc is now guaranteed to not |
| 7920 |
return. |
| 7921 |
|
| 7922 |
** Bugfixes for (ice-9 getopt-long) |
| 7923 |
|
| 7924 |
This module is now tested using test-suite/tests/getopt-long.test. |
| 7925 |
The following bugs have been fixed: |
| 7926 |
|
| 7927 |
*** Parsing for options that are specified to have `optional' args now checks |
| 7928 |
if the next element is an option instead of unconditionally taking it as the |
| 7929 |
option arg. |
| 7930 |
|
| 7931 |
*** An error is now thrown for `--opt=val' when the option description |
| 7932 |
does not specify `(value #t)' or `(value optional)'. This condition used to |
| 7933 |
be accepted w/o error, contrary to the documentation. |
| 7934 |
|
| 7935 |
*** The error message for unrecognized options is now more informative. |
| 7936 |
It used to be "not a record", an artifact of the implementation. |
| 7937 |
|
| 7938 |
*** The error message for `--opt' terminating the arg list (no value), when |
| 7939 |
`(value #t)' is specified, is now more informative. It used to be "not enough |
| 7940 |
args". |
| 7941 |
|
| 7942 |
*** "Clumped" single-char args now preserve trailing string, use it as arg. |
| 7943 |
The expansion used to be like so: |
| 7944 |
|
| 7945 |
("-abc5d" "--xyz") => ("-a" "-b" "-c" "--xyz") |
| 7946 |
|
| 7947 |
Note that the "5d" is dropped. Now it is like so: |
| 7948 |
|
| 7949 |
("-abc5d" "--xyz") => ("-a" "-b" "-c" "5d" "--xyz") |
| 7950 |
|
| 7951 |
This enables single-char options to have adjoining arguments as long as their |
| 7952 |
constituent characters are not potential single-char options. |
| 7953 |
|
| 7954 |
** (ice-9 session) procedure `arity' now works with (ice-9 optargs) `lambda*' |
| 7955 |
|
| 7956 |
The `lambda*' and derivative forms in (ice-9 optargs) now set a procedure |
| 7957 |
property `arglist', which can be retrieved by `arity'. The result is that |
| 7958 |
`arity' can give more detailed information than before: |
| 7959 |
|
| 7960 |
Before: |
| 7961 |
|
| 7962 |
guile> (use-modules (ice-9 optargs)) |
| 7963 |
guile> (define* (foo #:optional a b c) a) |
| 7964 |
guile> (arity foo) |
| 7965 |
0 or more arguments in `lambda*:G0'. |
| 7966 |
|
| 7967 |
After: |
| 7968 |
|
| 7969 |
guile> (arity foo) |
| 7970 |
3 optional arguments: `a', `b' and `c'. |
| 7971 |
guile> (define* (bar a b #:key c d #:allow-other-keys) a) |
| 7972 |
guile> (arity bar) |
| 7973 |
2 required arguments: `a' and `b', 2 keyword arguments: `c' |
| 7974 |
and `d', other keywords allowed. |
| 7975 |
guile> (define* (baz a b #:optional c #:rest r) a) |
| 7976 |
guile> (arity baz) |
| 7977 |
2 required arguments: `a' and `b', 1 optional argument: `c', |
| 7978 |
the rest in `r'. |
| 7979 |
|
| 7980 |
* Changes to the C interface |
| 7981 |
|
| 7982 |
** Types have been renamed from scm_*_t to scm_t_*. |
| 7983 |
|
| 7984 |
This has been done for POSIX sake. It reserves identifiers ending |
| 7985 |
with "_t". What a concept. |
| 7986 |
|
| 7987 |
The old names are still available with status `deprecated'. |
| 7988 |
|
| 7989 |
** scm_t_bits (former scm_bits_t) is now a unsigned type. |
| 7990 |
|
| 7991 |
** Deprecated features have been removed. |
| 7992 |
|
| 7993 |
*** Macros removed |
| 7994 |
|
| 7995 |
SCM_INPORTP, SCM_OUTPORTP SCM_ICHRP, SCM_ICHR, SCM_MAKICHR |
| 7996 |
SCM_SETJMPBUF SCM_NSTRINGP SCM_NRWSTRINGP SCM_NVECTORP SCM_DOUBLE_CELLP |
| 7997 |
|
| 7998 |
*** C Functions removed |
| 7999 |
|
| 8000 |
scm_sysmissing scm_tag scm_tc16_flo scm_tc_flo |
| 8001 |
scm_fseek - replaced by scm_seek. |
| 8002 |
gc-thunk - replaced by after-gc-hook. |
| 8003 |
gh_int2scmb - replaced by gh_bool2scm. |
| 8004 |
scm_tc_dblr - replaced by scm_tc16_real. |
| 8005 |
scm_tc_dblc - replaced by scm_tc16_complex. |
| 8006 |
scm_list_star - replaced by scm_cons_star. |
| 8007 |
|
| 8008 |
** Deprecated: scm_makfromstr |
| 8009 |
|
| 8010 |
Use scm_mem2string instead. |
| 8011 |
|
| 8012 |
** Deprecated: scm_make_shared_substring |
| 8013 |
|
| 8014 |
Explicit shared substrings will disappear from Guile. |
| 8015 |
|
| 8016 |
Instead, "normal" strings will be implemented using sharing |
| 8017 |
internally, combined with a copy-on-write strategy. |
| 8018 |
|
| 8019 |
** Deprecated: scm_read_only_string_p |
| 8020 |
|
| 8021 |
The concept of read-only strings will disappear in next release of |
| 8022 |
Guile. |
| 8023 |
|
| 8024 |
** Deprecated: scm_sloppy_memq, scm_sloppy_memv, scm_sloppy_member |
| 8025 |
|
| 8026 |
Instead, use scm_c_memq or scm_memq, scm_memv, scm_member. |
| 8027 |
|
| 8028 |
** New functions: scm_call_0, scm_call_1, scm_call_2, scm_call_3 |
| 8029 |
|
| 8030 |
Call a procedure with the indicated number of arguments. See "Fly |
| 8031 |
Evaluation" in the manual. |
| 8032 |
|
| 8033 |
** New functions: scm_apply_0, scm_apply_1, scm_apply_2, scm_apply_3 |
| 8034 |
|
| 8035 |
Call a procedure with the indicated number of arguments and a list of |
| 8036 |
further arguments. See "Fly Evaluation" in the manual. |
| 8037 |
|
| 8038 |
** New functions: scm_list_1, scm_list_2, scm_list_3, scm_list_4, scm_list_5 |
| 8039 |
|
| 8040 |
Create a list of the given number of elements. See "List |
| 8041 |
Constructors" in the manual. |
| 8042 |
|
| 8043 |
** Renamed function: scm_listify has been replaced by scm_list_n. |
| 8044 |
|
| 8045 |
** Deprecated macros: SCM_LIST0, SCM_LIST1, SCM_LIST2, SCM_LIST3, SCM_LIST4, |
| 8046 |
SCM_LIST5, SCM_LIST6, SCM_LIST7, SCM_LIST8, SCM_LIST9. |
| 8047 |
|
| 8048 |
Use functions scm_list_N instead. |
| 8049 |
|
| 8050 |
** New function: scm_c_read (SCM port, void *buffer, scm_sizet size) |
| 8051 |
|
| 8052 |
Used by an application to read arbitrary number of bytes from a port. |
| 8053 |
Same semantics as libc read, except that scm_c_read only returns less |
| 8054 |
than SIZE bytes if at end-of-file. |
| 8055 |
|
| 8056 |
Warning: Doesn't update port line and column counts! |
| 8057 |
|
| 8058 |
** New function: scm_c_write (SCM port, const void *ptr, scm_sizet size) |
| 8059 |
|
| 8060 |
Used by an application to write arbitrary number of bytes to an SCM |
| 8061 |
port. Similar semantics as libc write. However, unlike libc |
| 8062 |
write, scm_c_write writes the requested number of bytes and has no |
| 8063 |
return value. |
| 8064 |
|
| 8065 |
Warning: Doesn't update port line and column counts! |
| 8066 |
|
| 8067 |
** New function: scm_init_guile () |
| 8068 |
|
| 8069 |
In contrast to scm_boot_guile, scm_init_guile will return normally |
| 8070 |
after initializing Guile. It is not available on all systems, tho. |
| 8071 |
|
| 8072 |
** New functions: scm_str2symbol, scm_mem2symbol |
| 8073 |
|
| 8074 |
The function scm_str2symbol takes a const char* pointing to a zero-terminated |
| 8075 |
field of characters and creates a scheme symbol object from that C string. |
| 8076 |
The function scm_mem2symbol takes a const char* and a number of characters and |
| 8077 |
creates a symbol from the characters in that memory area. |
| 8078 |
|
| 8079 |
** New functions: scm_primitive_make_property |
| 8080 |
scm_primitive_property_ref |
| 8081 |
scm_primitive_property_set_x |
| 8082 |
scm_primitive_property_del_x |
| 8083 |
|
| 8084 |
These functions implement a new way to deal with object properties. |
| 8085 |
See libguile/properties.c for their documentation. |
| 8086 |
|
| 8087 |
** New function: scm_done_free (long size) |
| 8088 |
|
| 8089 |
This function is the inverse of scm_done_malloc. Use it to report the |
| 8090 |
amount of smob memory you free. The previous method, which involved |
| 8091 |
calling scm_done_malloc with negative argument, was somewhat |
| 8092 |
unintuitive (and is still available, of course). |
| 8093 |
|
| 8094 |
** New function: scm_c_memq (SCM obj, SCM list) |
| 8095 |
|
| 8096 |
This function provides a fast C level alternative for scm_memq for the case |
| 8097 |
that the list parameter is known to be a proper list. The function is a |
| 8098 |
replacement for scm_sloppy_memq, but is stricter in its requirements on its |
| 8099 |
list input parameter, since for anything else but a proper list the function's |
| 8100 |
behaviour is undefined - it may even crash or loop endlessly. Further, for |
| 8101 |
the case that the object is not found in the list, scm_c_memq returns #f which |
| 8102 |
is similar to scm_memq, but different from scm_sloppy_memq's behaviour. |
| 8103 |
|
| 8104 |
** New functions: scm_remember_upto_here_1, scm_remember_upto_here_2, |
| 8105 |
scm_remember_upto_here |
| 8106 |
|
| 8107 |
These functions replace the function scm_remember. |
| 8108 |
|
| 8109 |
** Deprecated function: scm_remember |
| 8110 |
|
| 8111 |
Use one of the new functions scm_remember_upto_here_1, |
| 8112 |
scm_remember_upto_here_2 or scm_remember_upto_here instead. |
| 8113 |
|
| 8114 |
** New function: scm_allocate_string |
| 8115 |
|
| 8116 |
This function replaces the function scm_makstr. |
| 8117 |
|
| 8118 |
** Deprecated function: scm_makstr |
| 8119 |
|
| 8120 |
Use the new function scm_allocate_string instead. |
| 8121 |
|
| 8122 |
** New global variable scm_gc_running_p introduced. |
| 8123 |
|
| 8124 |
Use this variable to find out if garbage collection is being executed. Up to |
| 8125 |
now applications have used scm_gc_heap_lock to test if garbage collection was |
| 8126 |
running, which also works because of the fact that up to know only the garbage |
| 8127 |
collector has set this variable. But, this is an implementation detail that |
| 8128 |
may change. Further, scm_gc_heap_lock is not set throughout gc, thus the use |
| 8129 |
of this variable is (and has been) not fully safe anyway. |
| 8130 |
|
| 8131 |
** New macros: SCM_BITVECTOR_MAX_LENGTH, SCM_UVECTOR_MAX_LENGTH |
| 8132 |
|
| 8133 |
Use these instead of SCM_LENGTH_MAX. |
| 8134 |
|
| 8135 |
** New macros: SCM_CONTINUATION_LENGTH, SCM_CCLO_LENGTH, SCM_STACK_LENGTH, |
| 8136 |
SCM_STRING_LENGTH, SCM_SYMBOL_LENGTH, SCM_UVECTOR_LENGTH, |
| 8137 |
SCM_BITVECTOR_LENGTH, SCM_VECTOR_LENGTH. |
| 8138 |
|
| 8139 |
Use these instead of SCM_LENGTH. |
| 8140 |
|
| 8141 |
** New macros: SCM_SET_CONTINUATION_LENGTH, SCM_SET_STRING_LENGTH, |
| 8142 |
SCM_SET_SYMBOL_LENGTH, SCM_SET_VECTOR_LENGTH, SCM_SET_UVECTOR_LENGTH, |
| 8143 |
SCM_SET_BITVECTOR_LENGTH |
| 8144 |
|
| 8145 |
Use these instead of SCM_SETLENGTH |
| 8146 |
|
| 8147 |
** New macros: SCM_STRING_CHARS, SCM_SYMBOL_CHARS, SCM_CCLO_BASE, |
| 8148 |
SCM_VECTOR_BASE, SCM_UVECTOR_BASE, SCM_BITVECTOR_BASE, SCM_COMPLEX_MEM, |
| 8149 |
SCM_ARRAY_MEM |
| 8150 |
|
| 8151 |
Use these instead of SCM_CHARS, SCM_UCHARS, SCM_ROCHARS, SCM_ROUCHARS or |
| 8152 |
SCM_VELTS. |
| 8153 |
|
| 8154 |
** New macros: SCM_SET_BIGNUM_BASE, SCM_SET_STRING_CHARS, |
| 8155 |
SCM_SET_SYMBOL_CHARS, SCM_SET_UVECTOR_BASE, SCM_SET_BITVECTOR_BASE, |
| 8156 |
SCM_SET_VECTOR_BASE |
| 8157 |
|
| 8158 |
Use these instead of SCM_SETCHARS. |
| 8159 |
|
| 8160 |
** New macro: SCM_BITVECTOR_P |
| 8161 |
|
| 8162 |
** New macro: SCM_STRING_COERCE_0TERMINATION_X |
| 8163 |
|
| 8164 |
Use instead of SCM_COERCE_SUBSTR. |
| 8165 |
|
| 8166 |
** New macros: SCM_DIR_OPEN_P, SCM_DIR_FLAG_OPEN |
| 8167 |
|
| 8168 |
For directory objects, use these instead of SCM_OPDIRP and SCM_OPN. |
| 8169 |
|
| 8170 |
** Deprecated macros: SCM_OUTOFRANGE, SCM_NALLOC, SCM_HUP_SIGNAL, |
| 8171 |
SCM_INT_SIGNAL, SCM_FPE_SIGNAL, SCM_BUS_SIGNAL, SCM_SEGV_SIGNAL, |
| 8172 |
SCM_ALRM_SIGNAL, SCM_GC_SIGNAL, SCM_TICK_SIGNAL, SCM_SIG_ORD, |
| 8173 |
SCM_ORD_SIG, SCM_NUM_SIGS, SCM_SYMBOL_SLOTS, SCM_SLOTS, SCM_SLOPPY_STRINGP, |
| 8174 |
SCM_VALIDATE_STRINGORSUBSTR, SCM_FREEP, SCM_NFREEP, SCM_CHARS, SCM_UCHARS, |
| 8175 |
SCM_VALIDATE_ROSTRING, SCM_VALIDATE_ROSTRING_COPY, |
| 8176 |
SCM_VALIDATE_NULLORROSTRING_COPY, SCM_ROLENGTH, SCM_LENGTH, SCM_HUGE_LENGTH, |
| 8177 |
SCM_SUBSTRP, SCM_SUBSTR_STR, SCM_SUBSTR_OFFSET, SCM_COERCE_SUBSTR, |
| 8178 |
SCM_ROSTRINGP, SCM_RWSTRINGP, SCM_VALIDATE_RWSTRING, SCM_ROCHARS, |
| 8179 |
SCM_ROUCHARS, SCM_SETLENGTH, SCM_SETCHARS, SCM_LENGTH_MAX, SCM_GC8MARKP, |
| 8180 |
SCM_SETGC8MARK, SCM_CLRGC8MARK, SCM_GCTYP16, SCM_GCCDR, SCM_SUBR_DOC, |
| 8181 |
SCM_OPDIRP, SCM_VALIDATE_OPDIR, SCM_WTA, RETURN_SCM_WTA, SCM_CONST_LONG, |
| 8182 |
SCM_WNA, SCM_FUNC_NAME, SCM_VALIDATE_NUMBER_COPY, |
| 8183 |
SCM_VALIDATE_NUMBER_DEF_COPY, SCM_SLOPPY_CONSP, SCM_SLOPPY_NCONSP, |
| 8184 |
SCM_SETAND_CDR, SCM_SETOR_CDR, SCM_SETAND_CAR, SCM_SETOR_CAR |
| 8185 |
|
| 8186 |
Use SCM_ASSERT_RANGE or SCM_VALIDATE_XXX_RANGE instead of SCM_OUTOFRANGE. |
| 8187 |
Use scm_memory_error instead of SCM_NALLOC. |
| 8188 |
Use SCM_STRINGP instead of SCM_SLOPPY_STRINGP. |
| 8189 |
Use SCM_VALIDATE_STRING instead of SCM_VALIDATE_STRINGORSUBSTR. |
| 8190 |
Use SCM_FREE_CELL_P instead of SCM_FREEP/SCM_NFREEP |
| 8191 |
Use a type specific accessor macro instead of SCM_CHARS/SCM_UCHARS. |
| 8192 |
Use a type specific accessor instead of SCM(_|_RO|_HUGE_)LENGTH. |
| 8193 |
Use SCM_VALIDATE_(SYMBOL|STRING) instead of SCM_VALIDATE_ROSTRING. |
| 8194 |
Use SCM_STRING_COERCE_0TERMINATION_X instead of SCM_COERCE_SUBSTR. |
| 8195 |
Use SCM_STRINGP or SCM_SYMBOLP instead of SCM_ROSTRINGP. |
| 8196 |
Use SCM_STRINGP instead of SCM_RWSTRINGP. |
| 8197 |
Use SCM_VALIDATE_STRING instead of SCM_VALIDATE_RWSTRING. |
| 8198 |
Use SCM_STRING_CHARS instead of SCM_ROCHARS. |
| 8199 |
Use SCM_STRING_UCHARS instead of SCM_ROUCHARS. |
| 8200 |
Use a type specific setter macro instead of SCM_SETLENGTH. |
| 8201 |
Use a type specific setter macro instead of SCM_SETCHARS. |
| 8202 |
Use a type specific length macro instead of SCM_LENGTH_MAX. |
| 8203 |
Use SCM_GCMARKP instead of SCM_GC8MARKP. |
| 8204 |
Use SCM_SETGCMARK instead of SCM_SETGC8MARK. |
| 8205 |
Use SCM_CLRGCMARK instead of SCM_CLRGC8MARK. |
| 8206 |
Use SCM_TYP16 instead of SCM_GCTYP16. |
| 8207 |
Use SCM_CDR instead of SCM_GCCDR. |
| 8208 |
Use SCM_DIR_OPEN_P instead of SCM_OPDIRP. |
| 8209 |
Use SCM_MISC_ERROR or SCM_WRONG_TYPE_ARG instead of SCM_WTA. |
| 8210 |
Use SCM_MISC_ERROR or SCM_WRONG_TYPE_ARG instead of RETURN_SCM_WTA. |
| 8211 |
Use SCM_VCELL_INIT instead of SCM_CONST_LONG. |
| 8212 |
Use SCM_WRONG_NUM_ARGS instead of SCM_WNA. |
| 8213 |
Use SCM_CONSP instead of SCM_SLOPPY_CONSP. |
| 8214 |
Use !SCM_CONSP instead of SCM_SLOPPY_NCONSP. |
| 8215 |
|
| 8216 |
** Removed function: scm_struct_init |
| 8217 |
|
| 8218 |
** Removed variable: scm_symhash_dim |
| 8219 |
|
| 8220 |
** Renamed function: scm_make_cont has been replaced by |
| 8221 |
scm_make_continuation, which has a different interface. |
| 8222 |
|
| 8223 |
** Deprecated function: scm_call_catching_errors |
| 8224 |
|
| 8225 |
Use scm_catch or scm_lazy_catch from throw.[ch] instead. |
| 8226 |
|
| 8227 |
** Deprecated function: scm_strhash |
| 8228 |
|
| 8229 |
Use scm_string_hash instead. |
| 8230 |
|
| 8231 |
** Deprecated function: scm_vector_set_length_x |
| 8232 |
|
| 8233 |
Instead, create a fresh vector of the desired size and copy the contents. |
| 8234 |
|
| 8235 |
** scm_gensym has changed prototype |
| 8236 |
|
| 8237 |
scm_gensym now only takes one argument. |
| 8238 |
|
| 8239 |
** Deprecated type tags: scm_tc7_ssymbol, scm_tc7_msymbol, scm_tcs_symbols, |
| 8240 |
scm_tc7_lvector |
| 8241 |
|
| 8242 |
There is now only a single symbol type scm_tc7_symbol. |
| 8243 |
The tag scm_tc7_lvector was not used anyway. |
| 8244 |
|
| 8245 |
** Deprecated function: scm_make_smob_type_mfpe, scm_set_smob_mfpe. |
| 8246 |
|
| 8247 |
Use scm_make_smob_type and scm_set_smob_XXX instead. |
| 8248 |
|
| 8249 |
** New function scm_set_smob_apply. |
| 8250 |
|
| 8251 |
This can be used to set an apply function to a smob type. |
| 8252 |
|
| 8253 |
** Deprecated function: scm_strprint_obj |
| 8254 |
|
| 8255 |
Use scm_object_to_string instead. |
| 8256 |
|
| 8257 |
** Deprecated function: scm_wta |
| 8258 |
|
| 8259 |
Use scm_wrong_type_arg, or another appropriate error signalling function |
| 8260 |
instead. |
| 8261 |
|
| 8262 |
** Explicit support for obarrays has been deprecated. |
| 8263 |
|
| 8264 |
Use `scm_str2symbol' and the generic hashtable functions instead. |
| 8265 |
|
| 8266 |
** The concept of `vcells' has been deprecated. |
| 8267 |
|
| 8268 |
The data type `variable' is now used exclusively. `Vcells' have been |
| 8269 |
a low-level concept so you are likely not affected by this change. |
| 8270 |
|
| 8271 |
*** Deprecated functions: scm_sym2vcell, scm_sysintern, |
| 8272 |
scm_sysintern0, scm_symbol_value0, scm_intern, scm_intern0. |
| 8273 |
|
| 8274 |
Use scm_c_define or scm_c_lookup instead, as appropriate. |
| 8275 |
|
| 8276 |
*** New functions: scm_c_module_lookup, scm_c_lookup, |
| 8277 |
scm_c_module_define, scm_c_define, scm_module_lookup, scm_lookup, |
| 8278 |
scm_module_define, scm_define. |
| 8279 |
|
| 8280 |
These functions work with variables instead of with vcells. |
| 8281 |
|
| 8282 |
** New functions for creating and defining `subr's and `gsubr's. |
| 8283 |
|
| 8284 |
The new functions more clearly distinguish between creating a subr (or |
| 8285 |
gsubr) object and adding it to the current module. |
| 8286 |
|
| 8287 |
These new functions are available: scm_c_make_subr, scm_c_define_subr, |
| 8288 |
scm_c_make_subr_with_generic, scm_c_define_subr_with_generic, |
| 8289 |
scm_c_make_gsubr, scm_c_define_gsubr, scm_c_make_gsubr_with_generic, |
| 8290 |
scm_c_define_gsubr_with_generic. |
| 8291 |
|
| 8292 |
** Deprecated functions: scm_make_subr, scm_make_subr_opt, |
| 8293 |
scm_make_subr_with_generic, scm_make_gsubr, |
| 8294 |
scm_make_gsubr_with_generic. |
| 8295 |
|
| 8296 |
Use the new ones from above instead. |
| 8297 |
|
| 8298 |
** C interface to the module system has changed. |
| 8299 |
|
| 8300 |
While we suggest that you avoid as many explicit module system |
| 8301 |
operations from C as possible for the time being, the C interface has |
| 8302 |
been made more similar to the high-level Scheme module system. |
| 8303 |
|
| 8304 |
*** New functions: scm_c_define_module, scm_c_use_module, |
| 8305 |
scm_c_export, scm_c_resolve_module. |
| 8306 |
|
| 8307 |
They mostly work like their Scheme namesakes. scm_c_define_module |
| 8308 |
takes a function that is called a context where the new module is |
| 8309 |
current. |
| 8310 |
|
| 8311 |
*** Deprecated functions: scm_the_root_module, scm_make_module, |
| 8312 |
scm_ensure_user_module, scm_load_scheme_module. |
| 8313 |
|
| 8314 |
Use the new functions instead. |
| 8315 |
|
| 8316 |
** Renamed function: scm_internal_with_fluids becomes |
| 8317 |
scm_c_with_fluids. |
| 8318 |
|
| 8319 |
scm_internal_with_fluids is available as a deprecated function. |
| 8320 |
|
| 8321 |
** New function: scm_c_with_fluid. |
| 8322 |
|
| 8323 |
Just like scm_c_with_fluids, but takes one fluid and one value instead |
| 8324 |
of lists of same. |
| 8325 |
|
| 8326 |
** Deprecated typedefs: long_long, ulong_long. |
| 8327 |
|
| 8328 |
They are of questionable utility and they pollute the global |
| 8329 |
namespace. |
| 8330 |
|
| 8331 |
** Deprecated typedef: scm_sizet |
| 8332 |
|
| 8333 |
It is of questionable utility now that Guile requires ANSI C, and is |
| 8334 |
oddly named. |
| 8335 |
|
| 8336 |
** Deprecated typedefs: scm_port_rw_active, scm_port, |
| 8337 |
scm_ptob_descriptor, scm_debug_info, scm_debug_frame, scm_fport, |
| 8338 |
scm_option, scm_rstate, scm_rng, scm_array, scm_array_dim. |
| 8339 |
|
| 8340 |
Made more compliant with the naming policy by adding a _t at the end. |
| 8341 |
|
| 8342 |
** Deprecated functions: scm_mkbig, scm_big2num, scm_adjbig, |
| 8343 |
scm_normbig, scm_copybig, scm_2ulong2big, scm_dbl2big, scm_big2dbl |
| 8344 |
|
| 8345 |
With the exception of the mysterious scm_2ulong2big, they are still |
| 8346 |
available under new names (scm_i_mkbig etc). These functions are not |
| 8347 |
intended to be used in user code. You should avoid dealing with |
| 8348 |
bignums directly, and should deal with numbers in general (which can |
| 8349 |
be bignums). |
| 8350 |
|
| 8351 |
** Change in behavior: scm_num2long, scm_num2ulong |
| 8352 |
|
| 8353 |
The scm_num2[u]long functions don't any longer accept an inexact |
| 8354 |
argument. This change in behavior is motivated by concordance with |
| 8355 |
R5RS: It is more common that a primitive doesn't want to accept an |
| 8356 |
inexact for an exact. |
| 8357 |
|
| 8358 |
** New functions: scm_short2num, scm_ushort2num, scm_int2num, |
| 8359 |
scm_uint2num, scm_size2num, scm_ptrdiff2num, scm_num2short, |
| 8360 |
scm_num2ushort, scm_num2int, scm_num2uint, scm_num2ptrdiff, |
| 8361 |
scm_num2size. |
| 8362 |
|
| 8363 |
These are conversion functions between the various ANSI C integral |
| 8364 |
types and Scheme numbers. NOTE: The scm_num2xxx functions don't |
| 8365 |
accept an inexact argument. |
| 8366 |
|
| 8367 |
** New functions: scm_float2num, scm_double2num, |
| 8368 |
scm_num2float, scm_num2double. |
| 8369 |
|
| 8370 |
These are conversion functions between the two ANSI C float types and |
| 8371 |
Scheme numbers. |
| 8372 |
|
| 8373 |
** New number validation macros: |
| 8374 |
SCM_NUM2{SIZE,PTRDIFF,SHORT,USHORT,INT,UINT}[_DEF] |
| 8375 |
|
| 8376 |
See above. |
| 8377 |
|
| 8378 |
** New functions: scm_gc_protect_object, scm_gc_unprotect_object |
| 8379 |
|
| 8380 |
These are just nicer-named old scm_protect_object and |
| 8381 |
scm_unprotect_object. |
| 8382 |
|
| 8383 |
** Deprecated functions: scm_protect_object, scm_unprotect_object |
| 8384 |
|
| 8385 |
** New functions: scm_gc_[un]register_root, scm_gc_[un]register_roots |
| 8386 |
|
| 8387 |
These functions can be used to register pointers to locations that |
| 8388 |
hold SCM values. |
| 8389 |
|
| 8390 |
** Deprecated function: scm_create_hook. |
| 8391 |
|
| 8392 |
Its sins are: misleading name, non-modularity and lack of general |
| 8393 |
usefulness. |
| 8394 |
|
| 8395 |
|
| 8396 |
Changes since Guile 1.3.4: |
| 8397 |
|
| 8398 |
* Changes to the distribution |
| 8399 |
|
| 8400 |
** Trees from nightly snapshots and CVS now require you to run autogen.sh. |
| 8401 |
|
| 8402 |
We've changed the way we handle generated files in the Guile source |
| 8403 |
repository. As a result, the procedure for building trees obtained |
| 8404 |
from the nightly FTP snapshots or via CVS has changed: |
| 8405 |
- You must have appropriate versions of autoconf, automake, and |
| 8406 |
libtool installed on your system. See README for info on how to |
| 8407 |
obtain these programs. |
| 8408 |
- Before configuring the tree, you must first run the script |
| 8409 |
`autogen.sh' at the top of the source tree. |
| 8410 |
|
| 8411 |
The Guile repository used to contain not only source files, written by |
| 8412 |
humans, but also some generated files, like configure scripts and |
| 8413 |
Makefile.in files. Even though the contents of these files could be |
| 8414 |
derived mechanically from other files present, we thought it would |
| 8415 |
make the tree easier to build if we checked them into CVS. |
| 8416 |
|
| 8417 |
However, this approach means that minor differences between |
| 8418 |
developer's installed tools and habits affected the whole team. |
| 8419 |
So we have removed the generated files from the repository, and |
| 8420 |
added the autogen.sh script, which will reconstruct them |
| 8421 |
appropriately. |
| 8422 |
|
| 8423 |
|
| 8424 |
** configure now has experimental options to remove support for certain |
| 8425 |
features: |
| 8426 |
|
| 8427 |
--disable-arrays omit array and uniform array support |
| 8428 |
--disable-posix omit posix interfaces |
| 8429 |
--disable-networking omit networking interfaces |
| 8430 |
--disable-regex omit regular expression interfaces |
| 8431 |
|
| 8432 |
These are likely to become separate modules some day. |
| 8433 |
|
| 8434 |
** New configure option --enable-debug-freelist |
| 8435 |
|
| 8436 |
This enables a debugging version of SCM_NEWCELL(), and also registers |
| 8437 |
an extra primitive, the setter `gc-set-debug-check-freelist!'. |
| 8438 |
|
| 8439 |
Configure with the --enable-debug-freelist option to enable |
| 8440 |
the gc-set-debug-check-freelist! primitive, and then use: |
| 8441 |
|
| 8442 |
(gc-set-debug-check-freelist! #t) # turn on checking of the freelist |
| 8443 |
(gc-set-debug-check-freelist! #f) # turn off checking |
| 8444 |
|
| 8445 |
Checking of the freelist forces a traversal of the freelist and |
| 8446 |
a garbage collection before each allocation of a cell. This can |
| 8447 |
slow down the interpreter dramatically, so the setter should be used to |
| 8448 |
turn on this extra processing only when necessary. |
| 8449 |
|
| 8450 |
** New configure option --enable-debug-malloc |
| 8451 |
|
| 8452 |
Include code for debugging of calls to scm_must_malloc/realloc/free. |
| 8453 |
|
| 8454 |
Checks that |
| 8455 |
|
| 8456 |
1. objects freed by scm_must_free has been mallocated by scm_must_malloc |
| 8457 |
2. objects reallocated by scm_must_realloc has been allocated by |
| 8458 |
scm_must_malloc |
| 8459 |
3. reallocated objects are reallocated with the same what string |
| 8460 |
|
| 8461 |
But, most importantly, it records the number of allocated objects of |
| 8462 |
each kind. This is useful when searching for memory leaks. |
| 8463 |
|
| 8464 |
A Guile compiled with this option provides the primitive |
| 8465 |
`malloc-stats' which returns an alist with pairs of kind and the |
| 8466 |
number of objects of that kind. |
| 8467 |
|
| 8468 |
** All includes are now referenced relative to the root directory |
| 8469 |
|
| 8470 |
Since some users have had problems with mixups between Guile and |
| 8471 |
system headers, we have decided to always refer to Guile headers via |
| 8472 |
their parent directories. This essentially creates a "private name |
| 8473 |
space" for Guile headers. This means that the compiler only is given |
| 8474 |
-I options for the root build and root source directory. |
| 8475 |
|
| 8476 |
** Header files kw.h and genio.h have been removed. |
| 8477 |
|
| 8478 |
** The module (ice-9 getopt-gnu-style) has been removed. |
| 8479 |
|
| 8480 |
** New module (ice-9 documentation) |
| 8481 |
|
| 8482 |
Implements the interface to documentation strings associated with |
| 8483 |
objects. |
| 8484 |
|
| 8485 |
** New module (ice-9 time) |
| 8486 |
|
| 8487 |
Provides a macro `time', which displays execution time of a given form. |
| 8488 |
|
| 8489 |
** New module (ice-9 history) |
| 8490 |
|
| 8491 |
Loading this module enables value history in the repl. |
| 8492 |
|
| 8493 |
* Changes to the stand-alone interpreter |
| 8494 |
|
| 8495 |
** New command line option --debug |
| 8496 |
|
| 8497 |
Start Guile with debugging evaluator and backtraces enabled. |
| 8498 |
|
| 8499 |
This is useful when debugging your .guile init file or scripts. |
| 8500 |
|
| 8501 |
** New help facility |
| 8502 |
|
| 8503 |
Usage: (help NAME) gives documentation about objects named NAME (a symbol) |
| 8504 |
(help REGEXP) ditto for objects with names matching REGEXP (a string) |
| 8505 |
(help 'NAME) gives documentation for NAME, even if it is not an object |
| 8506 |
(help ,EXPR) gives documentation for object returned by EXPR |
| 8507 |
(help (my module)) gives module commentary for `(my module)' |
| 8508 |
(help) gives this text |
| 8509 |
|
| 8510 |
`help' searches among bindings exported from loaded modules, while |
| 8511 |
`apropos' searches among bindings visible from the "current" module. |
| 8512 |
|
| 8513 |
Examples: (help help) |
| 8514 |
(help cons) |
| 8515 |
(help "output-string") |
| 8516 |
|
| 8517 |
** `help' and `apropos' now prints full module names |
| 8518 |
|
| 8519 |
** Dynamic linking now uses libltdl from the libtool package. |
| 8520 |
|
| 8521 |
The old system dependent code for doing dynamic linking has been |
| 8522 |
replaced with calls to the libltdl functions which do all the hairy |
| 8523 |
details for us. |
| 8524 |
|
| 8525 |
The major improvement is that you can now directly pass libtool |
| 8526 |
library names like "libfoo.la" to `dynamic-link' and `dynamic-link' |
| 8527 |
will be able to do the best shared library job you can get, via |
| 8528 |
libltdl. |
| 8529 |
|
| 8530 |
The way dynamic libraries are found has changed and is not really |
| 8531 |
portable across platforms, probably. It is therefore recommended to |
| 8532 |
use absolute filenames when possible. |
| 8533 |
|
| 8534 |
If you pass a filename without an extension to `dynamic-link', it will |
| 8535 |
try a few appropriate ones. Thus, the most platform ignorant way is |
| 8536 |
to specify a name like "libfoo", without any directories and |
| 8537 |
extensions. |
| 8538 |
|
| 8539 |
** Guile COOP threads are now compatible with LinuxThreads |
| 8540 |
|
| 8541 |
Previously, COOP threading wasn't possible in applications linked with |
| 8542 |
Linux POSIX threads due to their use of the stack pointer to find the |
| 8543 |
thread context. This has now been fixed with a workaround which uses |
| 8544 |
the pthreads to allocate the stack. |
| 8545 |
|
| 8546 |
** New primitives: `pkgdata-dir', `site-dir', `library-dir' |
| 8547 |
|
| 8548 |
** Positions of erring expression in scripts |
| 8549 |
|
| 8550 |
With version 1.3.4, the location of the erring expression in Guile |
| 8551 |
scipts is no longer automatically reported. (This should have been |
| 8552 |
documented before the 1.3.4 release.) |
| 8553 |
|
| 8554 |
You can get this information by enabling recording of positions of |
| 8555 |
source expressions and running the debugging evaluator. Put this at |
| 8556 |
the top of your script (or in your "site" file): |
| 8557 |
|
| 8558 |
(read-enable 'positions) |
| 8559 |
(debug-enable 'debug) |
| 8560 |
|
| 8561 |
** Backtraces in scripts |
| 8562 |
|
| 8563 |
It is now possible to get backtraces in scripts. |
| 8564 |
|
| 8565 |
Put |
| 8566 |
|
| 8567 |
(debug-enable 'debug 'backtrace) |
| 8568 |
|
| 8569 |
at the top of the script. |
| 8570 |
|
| 8571 |
(The first options enables the debugging evaluator. |
| 8572 |
The second enables backtraces.) |
| 8573 |
|
| 8574 |
** Part of module system symbol lookup now implemented in C |
| 8575 |
|
| 8576 |
The eval closure of most modules is now implemented in C. Since this |
| 8577 |
was one of the bottlenecks for loading speed, Guile now loads code |
| 8578 |
substantially faster than before. |
| 8579 |
|
| 8580 |
** Attempting to get the value of an unbound variable now produces |
| 8581 |
an exception with a key of 'unbound-variable instead of 'misc-error. |
| 8582 |
|
| 8583 |
** The initial default output port is now unbuffered if it's using a |
| 8584 |
tty device. Previously in this situation it was line-buffered. |
| 8585 |
|
| 8586 |
** New hook: after-gc-hook |
| 8587 |
|
| 8588 |
after-gc-hook takes over the role of gc-thunk. This hook is run at |
| 8589 |
the first SCM_TICK after a GC. (Thus, the code is run at the same |
| 8590 |
point during evaluation as signal handlers.) |
| 8591 |
|
| 8592 |
Note that this hook should be used only for diagnostic and debugging |
| 8593 |
purposes. It is not certain that it will continue to be well-defined |
| 8594 |
when this hook is run in the future. |
| 8595 |
|
| 8596 |
C programmers: Note the new C level hooks scm_before_gc_c_hook, |
| 8597 |
scm_before_sweep_c_hook, scm_after_gc_c_hook. |
| 8598 |
|
| 8599 |
** Improvements to garbage collector |
| 8600 |
|
| 8601 |
Guile 1.4 has a new policy for triggering heap allocation and |
| 8602 |
determining the sizes of heap segments. It fixes a number of problems |
| 8603 |
in the old GC. |
| 8604 |
|
| 8605 |
1. The new policy can handle two separate pools of cells |
| 8606 |
(2-word/4-word) better. (The old policy would run wild, allocating |
| 8607 |
more and more memory for certain programs.) |
| 8608 |
|
| 8609 |
2. The old code would sometimes allocate far too much heap so that the |
| 8610 |
Guile process became gigantic. The new code avoids this. |
| 8611 |
|
| 8612 |
3. The old code would sometimes allocate too little so that few cells |
| 8613 |
were freed at GC so that, in turn, too much time was spent in GC. |
| 8614 |
|
| 8615 |
4. The old code would often trigger heap allocation several times in a |
| 8616 |
row. (The new scheme predicts how large the segments needs to be |
| 8617 |
in order not to need further allocation.) |
| 8618 |
|
| 8619 |
All in all, the new GC policy will make larger applications more |
| 8620 |
efficient. |
| 8621 |
|
| 8622 |
The new GC scheme also is prepared for POSIX threading. Threads can |
| 8623 |
allocate private pools of cells ("clusters") with just a single |
| 8624 |
function call. Allocation of single cells from such a cluster can |
| 8625 |
then proceed without any need of inter-thread synchronization. |
| 8626 |
|
| 8627 |
** New environment variables controlling GC parameters |
| 8628 |
|
| 8629 |
GUILE_MAX_SEGMENT_SIZE Maximal segment size |
| 8630 |
(default = 2097000) |
| 8631 |
|
| 8632 |
Allocation of 2-word cell heaps: |
| 8633 |
|
| 8634 |
GUILE_INIT_SEGMENT_SIZE_1 Size of initial heap segment in bytes |
| 8635 |
(default = 360000) |
| 8636 |
|
| 8637 |
GUILE_MIN_YIELD_1 Minimum number of freed cells at each |
| 8638 |
GC in percent of total heap size |
| 8639 |
(default = 40) |
| 8640 |
|
| 8641 |
Allocation of 4-word cell heaps |
| 8642 |
(used for real numbers and misc other objects): |
| 8643 |
|
| 8644 |
GUILE_INIT_SEGMENT_SIZE_2, GUILE_MIN_YIELD_2 |
| 8645 |
|
| 8646 |
(See entry "Way for application to customize GC parameters" under |
| 8647 |
section "Changes to the scm_ interface" below.) |
| 8648 |
|
| 8649 |
** Guile now implements reals using 4-word cells |
| 8650 |
|
| 8651 |
This speeds up computation with reals. (They were earlier allocated |
| 8652 |
with `malloc'.) There is still some room for optimizations, however. |
| 8653 |
|
| 8654 |
** Some further steps toward POSIX thread support have been taken |
| 8655 |
|
| 8656 |
*** Guile's critical sections (SCM_DEFER/ALLOW_INTS) |
| 8657 |
don't have much effect any longer, and many of them will be removed in |
| 8658 |
next release. |
| 8659 |
|
| 8660 |
*** Signals |
| 8661 |
are only handled at the top of the evaluator loop, immediately after |
| 8662 |
I/O, and in scm_equalp. |
| 8663 |
|
| 8664 |
*** The GC can allocate thread private pools of pairs. |
| 8665 |
|
| 8666 |
* Changes to Scheme functions and syntax |
| 8667 |
|
| 8668 |
** close-input-port and close-output-port are now R5RS |
| 8669 |
|
| 8670 |
These procedures have been turned into primitives and have R5RS behaviour. |
| 8671 |
|
| 8672 |
** New procedure: simple-format PORT MESSAGE ARG1 ... |
| 8673 |
|
| 8674 |
(ice-9 boot) makes `format' an alias for `simple-format' until possibly |
| 8675 |
extended by the more sophisticated version in (ice-9 format) |
| 8676 |
|
| 8677 |
(simple-format port message . args) |
| 8678 |
Write MESSAGE to DESTINATION, defaulting to `current-output-port'. |
| 8679 |
MESSAGE can contain ~A (was %s) and ~S (was %S) escapes. When printed, |
| 8680 |
the escapes are replaced with corresponding members of ARGS: |
| 8681 |
~A formats using `display' and ~S formats using `write'. |
| 8682 |
If DESTINATION is #t, then use the `current-output-port', |
| 8683 |
if DESTINATION is #f, then return a string containing the formatted text. |
| 8684 |
Does not add a trailing newline." |
| 8685 |
|
| 8686 |
** string-ref: the second argument is no longer optional. |
| 8687 |
|
| 8688 |
** string, list->string: no longer accept strings in their arguments, |
| 8689 |
only characters, for compatibility with R5RS. |
| 8690 |
|
| 8691 |
** New procedure: port-closed? PORT |
| 8692 |
Returns #t if PORT is closed or #f if it is open. |
| 8693 |
|
| 8694 |
** Deprecated: list* |
| 8695 |
|
| 8696 |
The list* functionality is now provided by cons* (SRFI-1 compliant) |
| 8697 |
|
| 8698 |
** New procedure: cons* ARG1 ARG2 ... ARGn |
| 8699 |
|
| 8700 |
Like `list', but the last arg provides the tail of the constructed list, |
| 8701 |
returning (cons ARG1 (cons ARG2 (cons ... ARGn))). |
| 8702 |
|
| 8703 |
Requires at least one argument. If given one argument, that argument |
| 8704 |
is returned as result. |
| 8705 |
|
| 8706 |
This function is called `list*' in some other Schemes and in Common LISP. |
| 8707 |
|
| 8708 |
** Removed deprecated: serial-map, serial-array-copy!, serial-array-map! |
| 8709 |
|
| 8710 |
** New procedure: object-documentation OBJECT |
| 8711 |
|
| 8712 |
Returns the documentation string associated with OBJECT. The |
| 8713 |
procedure uses a caching mechanism so that subsequent lookups are |
| 8714 |
faster. |
| 8715 |
|
| 8716 |
Exported by (ice-9 documentation). |
| 8717 |
|
| 8718 |
** module-name now returns full names of modules |
| 8719 |
|
| 8720 |
Previously, only the last part of the name was returned (`session' for |
| 8721 |
`(ice-9 session)'). Ex: `(ice-9 session)'. |
| 8722 |
|
| 8723 |
* Changes to the gh_ interface |
| 8724 |
|
| 8725 |
** Deprecated: gh_int2scmb |
| 8726 |
|
| 8727 |
Use gh_bool2scm instead. |
| 8728 |
|
| 8729 |
* Changes to the scm_ interface |
| 8730 |
|
| 8731 |
** Guile primitives now carry docstrings! |
| 8732 |
|
| 8733 |
Thanks to Greg Badros! |
| 8734 |
|
| 8735 |
** Guile primitives are defined in a new way: SCM_DEFINE/SCM_DEFINE1/SCM_PROC |
| 8736 |
|
| 8737 |
Now Guile primitives are defined using the SCM_DEFINE/SCM_DEFINE1/SCM_PROC |
| 8738 |
macros and must contain a docstring that is extracted into foo.doc using a new |
| 8739 |
guile-doc-snarf script (that uses guile-doc-snarf.awk). |
| 8740 |
|
| 8741 |
However, a major overhaul of these macros is scheduled for the next release of |
| 8742 |
guile. |
| 8743 |
|
| 8744 |
** Guile primitives use a new technique for validation of arguments |
| 8745 |
|
| 8746 |
SCM_VALIDATE_* macros are defined to ease the redundancy and improve |
| 8747 |
the readability of argument checking. |
| 8748 |
|
| 8749 |
** All (nearly?) K&R prototypes for functions replaced with ANSI C equivalents. |
| 8750 |
|
| 8751 |
** New macros: SCM_PACK, SCM_UNPACK |
| 8752 |
|
| 8753 |
Compose/decompose an SCM value. |
| 8754 |
|
| 8755 |
The SCM type is now treated as an abstract data type and may be defined as a |
| 8756 |
long, a void* or as a struct, depending on the architecture and compile time |
| 8757 |
options. This makes it easier to find several types of bugs, for example when |
| 8758 |
SCM values are treated as integers without conversion. Values of the SCM type |
| 8759 |
should be treated as "atomic" values. These macros are used when |
| 8760 |
composing/decomposing an SCM value, either because you want to access |
| 8761 |
individual bits, or because you want to treat it as an integer value. |
| 8762 |
|
| 8763 |
E.g., in order to set bit 7 in an SCM value x, use the expression |
| 8764 |
|
| 8765 |
SCM_PACK (SCM_UNPACK (x) | 0x80) |
| 8766 |
|
| 8767 |
** The name property of hooks is deprecated. |
| 8768 |
Thus, the use of SCM_HOOK_NAME and scm_make_hook_with_name is deprecated. |
| 8769 |
|
| 8770 |
You can emulate this feature by using object properties. |
| 8771 |
|
| 8772 |
** Deprecated macros: SCM_INPORTP, SCM_OUTPORTP, SCM_CRDY, SCM_ICHRP, |
| 8773 |
SCM_ICHR, SCM_MAKICHR, SCM_SETJMPBUF, SCM_NSTRINGP, SCM_NRWSTRINGP, |
| 8774 |
SCM_NVECTORP |
| 8775 |
|
| 8776 |
These macros will be removed in a future release of Guile. |
| 8777 |
|
| 8778 |
** The following types, functions and macros from numbers.h are deprecated: |
| 8779 |
scm_dblproc, SCM_UNEGFIXABLE, SCM_FLOBUFLEN, SCM_INEXP, SCM_CPLXP, SCM_REAL, |
| 8780 |
SCM_IMAG, SCM_REALPART, scm_makdbl, SCM_SINGP, SCM_NUM2DBL, SCM_NO_BIGDIG |
| 8781 |
|
| 8782 |
** Port internals: the rw_random variable in the scm_port structure |
| 8783 |
must be set to non-zero in any random access port. In recent Guile |
| 8784 |
releases it was only set for bidirectional random-access ports. |
| 8785 |
|
| 8786 |
** Port internals: the seek ptob procedure is now responsible for |
| 8787 |
resetting the buffers if required. The change was made so that in the |
| 8788 |
special case of reading the current position (i.e., seek p 0 SEEK_CUR) |
| 8789 |
the fport and strport ptobs can avoid resetting the buffers, |
| 8790 |
in particular to avoid discarding unread chars. An existing port |
| 8791 |
type can be fixed by adding something like the following to the |
| 8792 |
beginning of the ptob seek procedure: |
| 8793 |
|
| 8794 |
if (pt->rw_active == SCM_PORT_READ) |
| 8795 |
scm_end_input (object); |
| 8796 |
else if (pt->rw_active == SCM_PORT_WRITE) |
| 8797 |
ptob->flush (object); |
| 8798 |
|
| 8799 |
although to actually avoid resetting the buffers and discard unread |
| 8800 |
chars requires further hacking that depends on the characteristics |
| 8801 |
of the ptob. |
| 8802 |
|
| 8803 |
** Deprecated functions: scm_fseek, scm_tag |
| 8804 |
|
| 8805 |
These functions are no longer used and will be removed in a future version. |
| 8806 |
|
| 8807 |
** The scm_sysmissing procedure is no longer used in libguile. |
| 8808 |
Unless it turns out to be unexpectedly useful to somebody, it will be |
| 8809 |
removed in a future version. |
| 8810 |
|
| 8811 |
** The format of error message strings has changed |
| 8812 |
|
| 8813 |
The two C procedures: scm_display_error and scm_error, as well as the |
| 8814 |
primitive `scm-error', now use scm_simple_format to do their work. |
| 8815 |
This means that the message strings of all code must be updated to use |
| 8816 |
~A where %s was used before, and ~S where %S was used before. |
| 8817 |
|
| 8818 |
During the period when there still are a lot of old Guiles out there, |
| 8819 |
you might want to support both old and new versions of Guile. |
| 8820 |
|
| 8821 |
There are basically two methods to achieve this. Both methods use |
| 8822 |
autoconf. Put |
| 8823 |
|
| 8824 |
AC_CHECK_FUNCS(scm_simple_format) |
| 8825 |
|
| 8826 |
in your configure.in. |
| 8827 |
|
| 8828 |
Method 1: Use the string concatenation features of ANSI C's |
| 8829 |
preprocessor. |
| 8830 |
|
| 8831 |
In C: |
| 8832 |
|
| 8833 |
#ifdef HAVE_SCM_SIMPLE_FORMAT |
| 8834 |
#define FMT_S "~S" |
| 8835 |
#else |
| 8836 |
#define FMT_S "%S" |
| 8837 |
#endif |
| 8838 |
|
| 8839 |
Then represent each of your error messages using a preprocessor macro: |
| 8840 |
|
| 8841 |
#define E_SPIDER_ERROR "There's a spider in your " ## FMT_S ## "!!!" |
| 8842 |
|
| 8843 |
In Scheme: |
| 8844 |
|
| 8845 |
(define fmt-s (if (defined? 'simple-format) "~S" "%S")) |
| 8846 |
(define make-message string-append) |
| 8847 |
|
| 8848 |
(define e-spider-error (make-message "There's a spider in your " fmt-s "!!!")) |
| 8849 |
|
| 8850 |
Method 2: Use the oldfmt function found in doc/oldfmt.c. |
| 8851 |
|
| 8852 |
In C: |
| 8853 |
|
| 8854 |
scm_misc_error ("picnic", scm_c_oldfmt0 ("There's a spider in your ~S!!!"), |
| 8855 |
...); |
| 8856 |
|
| 8857 |
In Scheme: |
| 8858 |
|
| 8859 |
(scm-error 'misc-error "picnic" (oldfmt "There's a spider in your ~S!!!") |
| 8860 |
...) |
| 8861 |
|
| 8862 |
|
| 8863 |
** Deprecated: coop_mutex_init, coop_condition_variable_init |
| 8864 |
|
| 8865 |
Don't use the functions coop_mutex_init and |
| 8866 |
coop_condition_variable_init. They will change. |
| 8867 |
|
| 8868 |
Use scm_mutex_init and scm_cond_init instead. |
| 8869 |
|
| 8870 |
** New function: int scm_cond_timedwait (scm_cond_t *COND, scm_mutex_t *MUTEX, const struct timespec *ABSTIME) |
| 8871 |
`scm_cond_timedwait' atomically unlocks MUTEX and waits on |
| 8872 |
COND, as `scm_cond_wait' does, but it also bounds the duration |
| 8873 |
of the wait. If COND has not been signaled before time ABSTIME, |
| 8874 |
the mutex MUTEX is re-acquired and `scm_cond_timedwait' |
| 8875 |
returns the error code `ETIMEDOUT'. |
| 8876 |
|
| 8877 |
The ABSTIME parameter specifies an absolute time, with the same |
| 8878 |
origin as `time' and `gettimeofday': an ABSTIME of 0 corresponds |
| 8879 |
to 00:00:00 GMT, January 1, 1970. |
| 8880 |
|
| 8881 |
** New function: scm_cond_broadcast (scm_cond_t *COND) |
| 8882 |
`scm_cond_broadcast' restarts all the threads that are waiting |
| 8883 |
on the condition variable COND. Nothing happens if no threads are |
| 8884 |
waiting on COND. |
| 8885 |
|
| 8886 |
** New function: scm_key_create (scm_key_t *KEY, void (*destr_function) (void *)) |
| 8887 |
`scm_key_create' allocates a new TSD key. The key is stored in |
| 8888 |
the location pointed to by KEY. There is no limit on the number |
| 8889 |
of keys allocated at a given time. The value initially associated |
| 8890 |
with the returned key is `NULL' in all currently executing threads. |
| 8891 |
|
| 8892 |
The DESTR_FUNCTION argument, if not `NULL', specifies a destructor |
| 8893 |
function associated with the key. When a thread terminates, |
| 8894 |
DESTR_FUNCTION is called on the value associated with the key in |
| 8895 |
that thread. The DESTR_FUNCTION is not called if a key is deleted |
| 8896 |
with `scm_key_delete' or a value is changed with |
| 8897 |
`scm_setspecific'. The order in which destructor functions are |
| 8898 |
called at thread termination time is unspecified. |
| 8899 |
|
| 8900 |
Destructors are not yet implemented. |
| 8901 |
|
| 8902 |
** New function: scm_setspecific (scm_key_t KEY, const void *POINTER) |
| 8903 |
`scm_setspecific' changes the value associated with KEY in the |
| 8904 |
calling thread, storing the given POINTER instead. |
| 8905 |
|
| 8906 |
** New function: scm_getspecific (scm_key_t KEY) |
| 8907 |
`scm_getspecific' returns the value currently associated with |
| 8908 |
KEY in the calling thread. |
| 8909 |
|
| 8910 |
** New function: scm_key_delete (scm_key_t KEY) |
| 8911 |
`scm_key_delete' deallocates a TSD key. It does not check |
| 8912 |
whether non-`NULL' values are associated with that key in the |
| 8913 |
currently executing threads, nor call the destructor function |
| 8914 |
associated with the key. |
| 8915 |
|
| 8916 |
** New function: scm_c_hook_init (scm_c_hook_t *HOOK, void *HOOK_DATA, scm_c_hook_type_t TYPE) |
| 8917 |
|
| 8918 |
Initialize a C level hook HOOK with associated HOOK_DATA and type |
| 8919 |
TYPE. (See scm_c_hook_run ().) |
| 8920 |
|
| 8921 |
** New function: scm_c_hook_add (scm_c_hook_t *HOOK, scm_c_hook_function_t FUNC, void *FUNC_DATA, int APPENDP) |
| 8922 |
|
| 8923 |
Add hook function FUNC with associated FUNC_DATA to HOOK. If APPENDP |
| 8924 |
is true, add it last, otherwise first. The same FUNC can be added |
| 8925 |
multiple times if FUNC_DATA differ and vice versa. |
| 8926 |
|
| 8927 |
** New function: scm_c_hook_remove (scm_c_hook_t *HOOK, scm_c_hook_function_t FUNC, void *FUNC_DATA) |
| 8928 |
|
| 8929 |
Remove hook function FUNC with associated FUNC_DATA from HOOK. A |
| 8930 |
function is only removed if both FUNC and FUNC_DATA matches. |
| 8931 |
|
| 8932 |
** New function: void *scm_c_hook_run (scm_c_hook_t *HOOK, void *DATA) |
| 8933 |
|
| 8934 |
Run hook HOOK passing DATA to the hook functions. |
| 8935 |
|
| 8936 |
If TYPE is SCM_C_HOOK_NORMAL, all hook functions are run. The value |
| 8937 |
returned is undefined. |
| 8938 |
|
| 8939 |
If TYPE is SCM_C_HOOK_OR, hook functions are run until a function |
| 8940 |
returns a non-NULL value. This value is returned as the result of |
| 8941 |
scm_c_hook_run. If all functions return NULL, NULL is returned. |
| 8942 |
|
| 8943 |
If TYPE is SCM_C_HOOK_AND, hook functions are run until a function |
| 8944 |
returns a NULL value, and NULL is returned. If all functions returns |
| 8945 |
a non-NULL value, the last value is returned. |
| 8946 |
|
| 8947 |
** New C level GC hooks |
| 8948 |
|
| 8949 |
Five new C level hooks has been added to the garbage collector. |
| 8950 |
|
| 8951 |
scm_before_gc_c_hook |
| 8952 |
scm_after_gc_c_hook |
| 8953 |
|
| 8954 |
are run before locking and after unlocking the heap. The system is |
| 8955 |
thus in a mode where evaluation can take place. (Except that |
| 8956 |
scm_before_gc_c_hook must not allocate new cells.) |
| 8957 |
|
| 8958 |
scm_before_mark_c_hook |
| 8959 |
scm_before_sweep_c_hook |
| 8960 |
scm_after_sweep_c_hook |
| 8961 |
|
| 8962 |
are run when the heap is locked. These are intended for extension of |
| 8963 |
the GC in a modular fashion. Examples are the weaks and guardians |
| 8964 |
modules. |
| 8965 |
|
| 8966 |
** Way for application to customize GC parameters |
| 8967 |
|
| 8968 |
The application can set up other default values for the GC heap |
| 8969 |
allocation parameters |
| 8970 |
|
| 8971 |
GUILE_INIT_HEAP_SIZE_1, GUILE_MIN_YIELD_1, |
| 8972 |
GUILE_INIT_HEAP_SIZE_2, GUILE_MIN_YIELD_2, |
| 8973 |
GUILE_MAX_SEGMENT_SIZE, |
| 8974 |
|
| 8975 |
by setting |
| 8976 |
|
| 8977 |
scm_default_init_heap_size_1, scm_default_min_yield_1, |
| 8978 |
scm_default_init_heap_size_2, scm_default_min_yield_2, |
| 8979 |
scm_default_max_segment_size |
| 8980 |
|
| 8981 |
respectively before callong scm_boot_guile. |
| 8982 |
|
| 8983 |
(See entry "New environment variables ..." in section |
| 8984 |
"Changes to the stand-alone interpreter" above.) |
| 8985 |
|
| 8986 |
** scm_protect_object/scm_unprotect_object now nest |
| 8987 |
|
| 8988 |
This means that you can call scm_protect_object multiple times on an |
| 8989 |
object and count on the object being protected until |
| 8990 |
scm_unprotect_object has been call the same number of times. |
| 8991 |
|
| 8992 |
The functions also have better time complexity. |
| 8993 |
|
| 8994 |
Still, it is usually possible to structure the application in a way |
| 8995 |
that you don't need to use these functions. For example, if you use a |
| 8996 |
protected standard Guile list to keep track of live objects rather |
| 8997 |
than some custom data type, objects will die a natural death when they |
| 8998 |
are no longer needed. |
| 8999 |
|
| 9000 |
** Deprecated type tags: scm_tc16_flo, scm_tc_flo, scm_tc_dblr, scm_tc_dblc |
| 9001 |
|
| 9002 |
Guile does not provide the float representation for inexact real numbers any |
| 9003 |
more. Now, only doubles are used to represent inexact real numbers. Further, |
| 9004 |
the tag names scm_tc_dblr and scm_tc_dblc have been changed to scm_tc16_real |
| 9005 |
and scm_tc16_complex, respectively. |
| 9006 |
|
| 9007 |
** Removed deprecated type scm_smobfuns |
| 9008 |
|
| 9009 |
** Removed deprecated function scm_newsmob |
| 9010 |
|
| 9011 |
** Warning: scm_make_smob_type_mfpe might become deprecated in a future release |
| 9012 |
|
| 9013 |
There is an ongoing discussion among the developers whether to |
| 9014 |
deprecate `scm_make_smob_type_mfpe' or not. Please use the current |
| 9015 |
standard interface (scm_make_smob_type, scm_set_smob_XXX) in new code |
| 9016 |
until this issue has been settled. |
| 9017 |
|
| 9018 |
** Removed deprecated type tag scm_tc16_kw |
| 9019 |
|
| 9020 |
** Added type tag scm_tc16_keyword |
| 9021 |
|
| 9022 |
(This was introduced already in release 1.3.4 but was not documented |
| 9023 |
until now.) |
| 9024 |
|
| 9025 |
** gdb_print now prints "*** Guile not initialized ***" until Guile initialized |
| 9026 |
|
| 9027 |
* Changes to system call interfaces: |
| 9028 |
|
| 9029 |
** The "select" procedure now tests port buffers for the ability to |
| 9030 |
provide input or accept output. Previously only the underlying file |
| 9031 |
descriptors were checked. |
| 9032 |
|
| 9033 |
** New variable PIPE_BUF: the maximum number of bytes that can be |
| 9034 |
atomically written to a pipe. |
| 9035 |
|
| 9036 |
** If a facility is not available on the system when Guile is |
| 9037 |
compiled, the corresponding primitive procedure will not be defined. |
| 9038 |
Previously it would have been defined but would throw a system-error |
| 9039 |
exception if called. Exception handlers which catch this case may |
| 9040 |
need minor modification: an error will be thrown with key |
| 9041 |
'unbound-variable instead of 'system-error. Alternatively it's |
| 9042 |
now possible to use `defined?' to check whether the facility is |
| 9043 |
available. |
| 9044 |
|
| 9045 |
** Procedures which depend on the timezone should now give the correct |
| 9046 |
result on systems which cache the TZ environment variable, even if TZ |
| 9047 |
is changed without calling tzset. |
| 9048 |
|
| 9049 |
* Changes to the networking interfaces: |
| 9050 |
|
| 9051 |
** New functions: htons, ntohs, htonl, ntohl: for converting short and |
| 9052 |
long integers between network and host format. For now, it's not |
| 9053 |
particularly convenient to do this kind of thing, but consider: |
| 9054 |
|
| 9055 |
(define write-network-long |
| 9056 |
(lambda (value port) |
| 9057 |
(let ((v (make-uniform-vector 1 1 0))) |
| 9058 |
(uniform-vector-set! v 0 (htonl value)) |
| 9059 |
(uniform-vector-write v port)))) |
| 9060 |
|
| 9061 |
(define read-network-long |
| 9062 |
(lambda (port) |
| 9063 |
(let ((v (make-uniform-vector 1 1 0))) |
| 9064 |
(uniform-vector-read! v port) |
| 9065 |
(ntohl (uniform-vector-ref v 0))))) |
| 9066 |
|
| 9067 |
** If inet-aton fails, it now throws an error with key 'misc-error |
| 9068 |
instead of 'system-error, since errno is not relevant. |
| 9069 |
|
| 9070 |
** Certain gethostbyname/gethostbyaddr failures now throw errors with |
| 9071 |
specific keys instead of 'system-error. The latter is inappropriate |
| 9072 |
since errno will not have been set. The keys are: |
| 9073 |
'host-not-found, 'try-again, 'no-recovery and 'no-data. |
| 9074 |
|
| 9075 |
** sethostent, setnetent, setprotoent, setservent: now take an |
| 9076 |
optional argument STAYOPEN, which specifies whether the database |
| 9077 |
remains open after a database entry is accessed randomly (e.g., using |
| 9078 |
gethostbyname for the hosts database.) The default is #f. Previously |
| 9079 |
#t was always used. |
| 9080 |
|
| 9081 |
|
| 9082 |
Changes since Guile 1.3.2: |
| 9083 |
|
| 9084 |
* Changes to the stand-alone interpreter |
| 9085 |
|
| 9086 |
** Debugger |
| 9087 |
|
| 9088 |
An initial version of the Guile debugger written by Chris Hanson has |
| 9089 |
been added. The debugger is still under development but is included |
| 9090 |
in the distribution anyway since it is already quite useful. |
| 9091 |
|
| 9092 |
Type |
| 9093 |
|
| 9094 |
(debug) |
| 9095 |
|
| 9096 |
after an error to enter the debugger. Type `help' inside the debugger |
| 9097 |
for a description of available commands. |
| 9098 |
|
| 9099 |
If you prefer to have stack frames numbered and printed in |
| 9100 |
anti-chronological order and prefer up in the stack to be down on the |
| 9101 |
screen as is the case in gdb, you can put |
| 9102 |
|
| 9103 |
(debug-enable 'backwards) |
| 9104 |
|
| 9105 |
in your .guile startup file. (However, this means that Guile can't |
| 9106 |
use indentation to indicate stack level.) |
| 9107 |
|
| 9108 |
The debugger is autoloaded into Guile at the first use. |
| 9109 |
|
| 9110 |
** Further enhancements to backtraces |
| 9111 |
|
| 9112 |
There is a new debug option `width' which controls the maximum width |
| 9113 |
on the screen of printed stack frames. Fancy printing parameters |
| 9114 |
("level" and "length" as in Common LISP) are adaptively adjusted for |
| 9115 |
each stack frame to give maximum information while still fitting |
| 9116 |
within the bounds. If the stack frame can't be made to fit by |
| 9117 |
adjusting parameters, it is simply cut off at the end. This is marked |
| 9118 |
with a `$'. |
| 9119 |
|
| 9120 |
** Some modules are now only loaded when the repl is started |
| 9121 |
|
| 9122 |
The modules (ice-9 debug), (ice-9 session), (ice-9 threads) and (ice-9 |
| 9123 |
regex) are now loaded into (guile-user) only if the repl has been |
| 9124 |
started. The effect is that the startup time for scripts has been |
| 9125 |
reduced to 30% of what it was previously. |
| 9126 |
|
| 9127 |
Correctly written scripts load the modules they require at the top of |
| 9128 |
the file and should not be affected by this change. |
| 9129 |
|
| 9130 |
** Hooks are now represented as smobs |
| 9131 |
|
| 9132 |
* Changes to Scheme functions and syntax |
| 9133 |
|
| 9134 |
** Readline support has changed again. |
| 9135 |
|
| 9136 |
The old (readline-activator) module is gone. Use (ice-9 readline) |
| 9137 |
instead, which now contains all readline functionality. So the code |
| 9138 |
to activate readline is now |
| 9139 |
|
| 9140 |
(use-modules (ice-9 readline)) |
| 9141 |
(activate-readline) |
| 9142 |
|
| 9143 |
This should work at any time, including from the guile prompt. |
| 9144 |
|
| 9145 |
To avoid confusion about the terms of Guile's license, please only |
| 9146 |
enable readline for your personal use; please don't make it the |
| 9147 |
default for others. Here is why we make this rather odd-sounding |
| 9148 |
request: |
| 9149 |
|
| 9150 |
Guile is normally licensed under a weakened form of the GNU General |
| 9151 |
Public License, which allows you to link code with Guile without |
| 9152 |
placing that code under the GPL. This exception is important to some |
| 9153 |
people. |
| 9154 |
|
| 9155 |
However, since readline is distributed under the GNU General Public |
| 9156 |
License, when you link Guile with readline, either statically or |
| 9157 |
dynamically, you effectively change Guile's license to the strict GPL. |
| 9158 |
Whenever you link any strictly GPL'd code into Guile, uses of Guile |
| 9159 |
which are normally permitted become forbidden. This is a rather |
| 9160 |
non-obvious consequence of the licensing terms. |
| 9161 |
|
| 9162 |
So, to make sure things remain clear, please let people choose for |
| 9163 |
themselves whether to link GPL'd libraries like readline with Guile. |
| 9164 |
|
| 9165 |
** regexp-substitute/global has changed slightly, but incompatibly. |
| 9166 |
|
| 9167 |
If you include a function in the item list, the string of the match |
| 9168 |
object it receives is the same string passed to |
| 9169 |
regexp-substitute/global, not some suffix of that string. |
| 9170 |
Correspondingly, the match's positions are relative to the entire |
| 9171 |
string, not the suffix. |
| 9172 |
|
| 9173 |
If the regexp can match the empty string, the way matches are chosen |
| 9174 |
from the string has changed. regexp-substitute/global recognizes the |
| 9175 |
same set of matches that list-matches does; see below. |
| 9176 |
|
| 9177 |
** New function: list-matches REGEXP STRING [FLAGS] |
| 9178 |
|
| 9179 |
Return a list of match objects, one for every non-overlapping, maximal |
| 9180 |
match of REGEXP in STRING. The matches appear in left-to-right order. |
| 9181 |
list-matches only reports matches of the empty string if there are no |
| 9182 |
other matches which begin on, end at, or include the empty match's |
| 9183 |
position. |
| 9184 |
|
| 9185 |
If present, FLAGS is passed as the FLAGS argument to regexp-exec. |
| 9186 |
|
| 9187 |
** New function: fold-matches REGEXP STRING INIT PROC [FLAGS] |
| 9188 |
|
| 9189 |
For each match of REGEXP in STRING, apply PROC to the match object, |
| 9190 |
and the last value PROC returned, or INIT for the first call. Return |
| 9191 |
the last value returned by PROC. We apply PROC to the matches as they |
| 9192 |
appear from left to right. |
| 9193 |
|
| 9194 |
This function recognizes matches according to the same criteria as |
| 9195 |
list-matches. |
| 9196 |
|
| 9197 |
Thus, you could define list-matches like this: |
| 9198 |
|
| 9199 |
(define (list-matches regexp string . flags) |
| 9200 |
(reverse! (apply fold-matches regexp string '() cons flags))) |
| 9201 |
|
| 9202 |
If present, FLAGS is passed as the FLAGS argument to regexp-exec. |
| 9203 |
|
| 9204 |
** Hooks |
| 9205 |
|
| 9206 |
*** New function: hook? OBJ |
| 9207 |
|
| 9208 |
Return #t if OBJ is a hook, otherwise #f. |
| 9209 |
|
| 9210 |
*** New function: make-hook-with-name NAME [ARITY] |
| 9211 |
|
| 9212 |
Return a hook with name NAME and arity ARITY. The default value for |
| 9213 |
ARITY is 0. The only effect of NAME is that it will appear when the |
| 9214 |
hook object is printed to ease debugging. |
| 9215 |
|
| 9216 |
*** New function: hook-empty? HOOK |
| 9217 |
|
| 9218 |
Return #t if HOOK doesn't contain any procedures, otherwise #f. |
| 9219 |
|
| 9220 |
*** New function: hook->list HOOK |
| 9221 |
|
| 9222 |
Return a list of the procedures that are called when run-hook is |
| 9223 |
applied to HOOK. |
| 9224 |
|
| 9225 |
** `map' signals an error if its argument lists are not all the same length. |
| 9226 |
|
| 9227 |
This is the behavior required by R5RS, so this change is really a bug |
| 9228 |
fix. But it seems to affect a lot of people's code, so we're |
| 9229 |
mentioning it here anyway. |
| 9230 |
|
| 9231 |
** Print-state handling has been made more transparent |
| 9232 |
|
| 9233 |
Under certain circumstances, ports are represented as a port with an |
| 9234 |
associated print state. Earlier, this pair was represented as a pair |
| 9235 |
(see "Some magic has been added to the printer" below). It is now |
| 9236 |
indistinguishable (almost; see `get-print-state') from a port on the |
| 9237 |
user level. |
| 9238 |
|
| 9239 |
*** New function: port-with-print-state OUTPUT-PORT PRINT-STATE |
| 9240 |
|
| 9241 |
Return a new port with the associated print state PRINT-STATE. |
| 9242 |
|
| 9243 |
*** New function: get-print-state OUTPUT-PORT |
| 9244 |
|
| 9245 |
Return the print state associated with this port if it exists, |
| 9246 |
otherwise return #f. |
| 9247 |
|
| 9248 |
*** New function: directory-stream? OBJECT |
| 9249 |
|
| 9250 |
Returns true iff OBJECT is a directory stream --- the sort of object |
| 9251 |
returned by `opendir'. |
| 9252 |
|
| 9253 |
** New function: using-readline? |
| 9254 |
|
| 9255 |
Return #t if readline is in use in the current repl. |
| 9256 |
|
| 9257 |
** structs will be removed in 1.4 |
| 9258 |
|
| 9259 |
Structs will be replaced in Guile 1.4. We will merge GOOPS into Guile |
| 9260 |
and use GOOPS objects as the fundamental record type. |
| 9261 |
|
| 9262 |
* Changes to the scm_ interface |
| 9263 |
|
| 9264 |
** structs will be removed in 1.4 |
| 9265 |
|
| 9266 |
The entire current struct interface (struct.c, struct.h) will be |
| 9267 |
replaced in Guile 1.4. We will merge GOOPS into libguile and use |
| 9268 |
GOOPS objects as the fundamental record type. |
| 9269 |
|
| 9270 |
** The internal representation of subr's has changed |
| 9271 |
|
| 9272 |
Instead of giving a hint to the subr name, the CAR field of the subr |
| 9273 |
now contains an index to a subr entry in scm_subr_table. |
| 9274 |
|
| 9275 |
*** New variable: scm_subr_table |
| 9276 |
|
| 9277 |
An array of subr entries. A subr entry contains the name, properties |
| 9278 |
and documentation associated with the subr. The properties and |
| 9279 |
documentation slots are not yet used. |
| 9280 |
|
| 9281 |
** A new scheme for "forwarding" calls to a builtin to a generic function |
| 9282 |
|
| 9283 |
It is now possible to extend the functionality of some Guile |
| 9284 |
primitives by letting them defer a call to a GOOPS generic function on |
| 9285 |
argument mismatch. This means that there is no loss of efficiency in |
| 9286 |
normal evaluation. |
| 9287 |
|
| 9288 |
Example: |
| 9289 |
|
| 9290 |
(use-modules (oop goops)) ; Must be GOOPS version 0.2. |
| 9291 |
(define-method + ((x <string>) (y <string>)) |
| 9292 |
(string-append x y)) |
| 9293 |
|
| 9294 |
+ will still be as efficient as usual in numerical calculations, but |
| 9295 |
can also be used for concatenating strings. |
| 9296 |
|
| 9297 |
Who will be the first one to extend Guile's numerical tower to |
| 9298 |
rationals? :) [OK, there a few other things to fix before this can |
| 9299 |
be made in a clean way.] |
| 9300 |
|
| 9301 |
*** New snarf macros for defining primitives: SCM_GPROC, SCM_GPROC1 |
| 9302 |
|
| 9303 |
New macro: SCM_GPROC (CNAME, SNAME, REQ, OPT, VAR, CFUNC, GENERIC) |
| 9304 |
|
| 9305 |
New macro: SCM_GPROC1 (CNAME, SNAME, TYPE, CFUNC, GENERIC) |
| 9306 |
|
| 9307 |
These do the same job as SCM_PROC and SCM_PROC1, but they also define |
| 9308 |
a variable GENERIC which can be used by the dispatch macros below. |
| 9309 |
|
| 9310 |
[This is experimental code which may change soon.] |
| 9311 |
|
| 9312 |
*** New macros for forwarding control to a generic on arg type error |
| 9313 |
|
| 9314 |
New macro: SCM_WTA_DISPATCH_1 (GENERIC, ARG1, POS, SUBR) |
| 9315 |
|
| 9316 |
New macro: SCM_WTA_DISPATCH_2 (GENERIC, ARG1, ARG2, POS, SUBR) |
| 9317 |
|
| 9318 |
These correspond to the scm_wta function call, and have the same |
| 9319 |
behaviour until the user has called the GOOPS primitive |
| 9320 |
`enable-primitive-generic!'. After that, these macros will apply the |
| 9321 |
generic function GENERIC to the argument(s) instead of calling |
| 9322 |
scm_wta. |
| 9323 |
|
| 9324 |
[This is experimental code which may change soon.] |
| 9325 |
|
| 9326 |
*** New macros for argument testing with generic dispatch |
| 9327 |
|
| 9328 |
New macro: SCM_GASSERT1 (COND, GENERIC, ARG1, POS, SUBR) |
| 9329 |
|
| 9330 |
New macro: SCM_GASSERT2 (COND, GENERIC, ARG1, ARG2, POS, SUBR) |
| 9331 |
|
| 9332 |
These correspond to the SCM_ASSERT macro, but will defer control to |
| 9333 |
GENERIC on error after `enable-primitive-generic!' has been called. |
| 9334 |
|
| 9335 |
[This is experimental code which may change soon.] |
| 9336 |
|
| 9337 |
** New function: SCM scm_eval_body (SCM body, SCM env) |
| 9338 |
|
| 9339 |
Evaluates the body of a special form. |
| 9340 |
|
| 9341 |
** The internal representation of struct's has changed |
| 9342 |
|
| 9343 |
Previously, four slots were allocated for the procedure(s) of entities |
| 9344 |
and operators. The motivation for this representation had to do with |
| 9345 |
the structure of the evaluator, the wish to support tail-recursive |
| 9346 |
generic functions, and efficiency. Since the generic function |
| 9347 |
dispatch mechanism has changed, there is no longer a need for such an |
| 9348 |
expensive representation, and the representation has been simplified. |
| 9349 |
|
| 9350 |
This should not make any difference for most users. |
| 9351 |
|
| 9352 |
** GOOPS support has been cleaned up. |
| 9353 |
|
| 9354 |
Some code has been moved from eval.c to objects.c and code in both of |
| 9355 |
these compilation units has been cleaned up and better structured. |
| 9356 |
|
| 9357 |
*** New functions for applying generic functions |
| 9358 |
|
| 9359 |
New function: SCM scm_apply_generic (GENERIC, ARGS) |
| 9360 |
New function: SCM scm_call_generic_0 (GENERIC) |
| 9361 |
New function: SCM scm_call_generic_1 (GENERIC, ARG1) |
| 9362 |
New function: SCM scm_call_generic_2 (GENERIC, ARG1, ARG2) |
| 9363 |
New function: SCM scm_call_generic_3 (GENERIC, ARG1, ARG2, ARG3) |
| 9364 |
|
| 9365 |
** Deprecated function: scm_make_named_hook |
| 9366 |
|
| 9367 |
It is now replaced by: |
| 9368 |
|
| 9369 |
** New function: SCM scm_create_hook (const char *name, int arity) |
| 9370 |
|
| 9371 |
Creates a hook in the same way as make-hook above but also |
| 9372 |
binds a variable named NAME to it. |
| 9373 |
|
| 9374 |
This is the typical way of creating a hook from C code. |
| 9375 |
|
| 9376 |
Currently, the variable is created in the "current" module. |
| 9377 |
This might change when we get the new module system. |
| 9378 |
|
| 9379 |
[The behaviour is identical to scm_make_named_hook.] |
| 9380 |
|
| 9381 |
|
| 9382 |
|
| 9383 |
Changes since Guile 1.3: |
| 9384 |
|
| 9385 |
* Changes to mailing lists |
| 9386 |
|
| 9387 |
** Some of the Guile mailing lists have moved to sourceware.cygnus.com. |
| 9388 |
|
| 9389 |
See the README file to find current addresses for all the Guile |
| 9390 |
mailing lists. |
| 9391 |
|
| 9392 |
* Changes to the distribution |
| 9393 |
|
| 9394 |
** Readline support is no longer included with Guile by default. |
| 9395 |
|
| 9396 |
Based on the different license terms of Guile and Readline, we |
| 9397 |
concluded that Guile should not *by default* cause the linking of |
| 9398 |
Readline into an application program. Readline support is now offered |
| 9399 |
as a separate module, which is linked into an application only when |
| 9400 |
you explicitly specify it. |
| 9401 |
|
| 9402 |
Although Guile is GNU software, its distribution terms add a special |
| 9403 |
exception to the usual GNU General Public License (GPL). Guile's |
| 9404 |
license includes a clause that allows you to link Guile with non-free |
| 9405 |
programs. We add this exception so as not to put Guile at a |
| 9406 |
disadvantage vis-a-vis other extensibility packages that support other |
| 9407 |
languages. |
| 9408 |
|
| 9409 |
In contrast, the GNU Readline library is distributed under the GNU |
| 9410 |
General Public License pure and simple. This means that you may not |
| 9411 |
link Readline, even dynamically, into an application unless it is |
| 9412 |
distributed under a free software license that is compatible the GPL. |
| 9413 |
|
| 9414 |
Because of this difference in distribution terms, an application that |
| 9415 |
can use Guile may not be able to use Readline. Now users will be |
| 9416 |
explicitly offered two independent decisions about the use of these |
| 9417 |
two packages. |
| 9418 |
|
| 9419 |
You can activate the readline support by issuing |
| 9420 |
|
| 9421 |
(use-modules (readline-activator)) |
| 9422 |
(activate-readline) |
| 9423 |
|
| 9424 |
from your ".guile" file, for example. |
| 9425 |
|
| 9426 |
* Changes to the stand-alone interpreter |
| 9427 |
|
| 9428 |
** All builtins now print as primitives. |
| 9429 |
Previously builtin procedures not belonging to the fundamental subr |
| 9430 |
types printed as #<compiled closure #<primitive-procedure gsubr-apply>>. |
| 9431 |
Now, they print as #<primitive-procedure NAME>. |
| 9432 |
|
| 9433 |
** Backtraces slightly more intelligible. |
| 9434 |
gsubr-apply and macro transformer application frames no longer appear |
| 9435 |
in backtraces. |
| 9436 |
|
| 9437 |
* Changes to Scheme functions and syntax |
| 9438 |
|
| 9439 |
** Guile now correctly handles internal defines by rewriting them into |
| 9440 |
their equivalent letrec. Previously, internal defines would |
| 9441 |
incrementally add to the innermost environment, without checking |
| 9442 |
whether the restrictions specified in RnRS were met. This lead to the |
| 9443 |
correct behaviour when these restriction actually were met, but didn't |
| 9444 |
catch all illegal uses. Such an illegal use could lead to crashes of |
| 9445 |
the Guile interpreter or other unwanted results. An example of |
| 9446 |
incorrect internal defines that made Guile behave erratically: |
| 9447 |
|
| 9448 |
(let () |
| 9449 |
(define a 1) |
| 9450 |
(define (b) a) |
| 9451 |
(define c (1+ (b))) |
| 9452 |
(define d 3) |
| 9453 |
|
| 9454 |
(b)) |
| 9455 |
|
| 9456 |
=> 2 |
| 9457 |
|
| 9458 |
The problem with this example is that the definition of `c' uses the |
| 9459 |
value of `b' directly. This confuses the meoization machine of Guile |
| 9460 |
so that the second call of `b' (this time in a larger environment that |
| 9461 |
also contains bindings for `c' and `d') refers to the binding of `c' |
| 9462 |
instead of `a'. You could also make Guile crash with a variation on |
| 9463 |
this theme: |
| 9464 |
|
| 9465 |
(define (foo flag) |
| 9466 |
(define a 1) |
| 9467 |
(define (b flag) (if flag a 1)) |
| 9468 |
(define c (1+ (b flag))) |
| 9469 |
(define d 3) |
| 9470 |
|
| 9471 |
(b #t)) |
| 9472 |
|
| 9473 |
(foo #f) |
| 9474 |
(foo #t) |
| 9475 |
|
| 9476 |
From now on, Guile will issue an `Unbound variable: b' error message |
| 9477 |
for both examples. |
| 9478 |
|
| 9479 |
** Hooks |
| 9480 |
|
| 9481 |
A hook contains a list of functions which should be called on |
| 9482 |
particular occasions in an existing program. Hooks are used for |
| 9483 |
customization. |
| 9484 |
|
| 9485 |
A window manager might have a hook before-window-map-hook. The window |
| 9486 |
manager uses the function run-hooks to call all functions stored in |
| 9487 |
before-window-map-hook each time a window is mapped. The user can |
| 9488 |
store functions in the hook using add-hook!. |
| 9489 |
|
| 9490 |
In Guile, hooks are first class objects. |
| 9491 |
|
| 9492 |
*** New function: make-hook [N_ARGS] |
| 9493 |
|
| 9494 |
Return a hook for hook functions which can take N_ARGS arguments. |
| 9495 |
The default value for N_ARGS is 0. |
| 9496 |
|
| 9497 |
(See also scm_make_named_hook below.) |
| 9498 |
|
| 9499 |
*** New function: add-hook! HOOK PROC [APPEND_P] |
| 9500 |
|
| 9501 |
Put PROC at the beginning of the list of functions stored in HOOK. |
| 9502 |
If APPEND_P is supplied, and non-false, put PROC at the end instead. |
| 9503 |
|
| 9504 |
PROC must be able to take the number of arguments specified when the |
| 9505 |
hook was created. |
| 9506 |
|
| 9507 |
If PROC already exists in HOOK, then remove it first. |
| 9508 |
|
| 9509 |
*** New function: remove-hook! HOOK PROC |
| 9510 |
|
| 9511 |
Remove PROC from the list of functions in HOOK. |
| 9512 |
|
| 9513 |
*** New function: reset-hook! HOOK |
| 9514 |
|
| 9515 |
Clear the list of hook functions stored in HOOK. |
| 9516 |
|
| 9517 |
*** New function: run-hook HOOK ARG1 ... |
| 9518 |
|
| 9519 |
Run all hook functions stored in HOOK with arguments ARG1 ... . |
| 9520 |
The number of arguments supplied must correspond to the number given |
| 9521 |
when the hook was created. |
| 9522 |
|
| 9523 |
** The function `dynamic-link' now takes optional keyword arguments. |
| 9524 |
The only keyword argument that is currently defined is `:global |
| 9525 |
BOOL'. With it, you can control whether the shared library will be |
| 9526 |
linked in global mode or not. In global mode, the symbols from the |
| 9527 |
linked library can be used to resolve references from other |
| 9528 |
dynamically linked libraries. In non-global mode, the linked |
| 9529 |
library is essentially invisible and can only be accessed via |
| 9530 |
`dynamic-func', etc. The default is now to link in global mode. |
| 9531 |
Previously, the default has been non-global mode. |
| 9532 |
|
| 9533 |
The `#:global' keyword is only effective on platforms that support |
| 9534 |
the dlopen family of functions. |
| 9535 |
|
| 9536 |
** New function `provided?' |
| 9537 |
|
| 9538 |
- Function: provided? FEATURE |
| 9539 |
Return true iff FEATURE is supported by this installation of |
| 9540 |
Guile. FEATURE must be a symbol naming a feature; the global |
| 9541 |
variable `*features*' is a list of available features. |
| 9542 |
|
| 9543 |
** Changes to the module (ice-9 expect): |
| 9544 |
|
| 9545 |
*** The expect-strings macro now matches `$' in a regular expression |
| 9546 |
only at a line-break or end-of-file by default. Previously it would |
| 9547 |
match the end of the string accumulated so far. The old behaviour |
| 9548 |
can be obtained by setting the variable `expect-strings-exec-flags' |
| 9549 |
to 0. |
| 9550 |
|
| 9551 |
*** The expect-strings macro now uses a variable `expect-strings-exec-flags' |
| 9552 |
for the regexp-exec flags. If `regexp/noteol' is included, then `$' |
| 9553 |
in a regular expression will still match before a line-break or |
| 9554 |
end-of-file. The default is `regexp/noteol'. |
| 9555 |
|
| 9556 |
*** The expect-strings macro now uses a variable |
| 9557 |
`expect-strings-compile-flags' for the flags to be supplied to |
| 9558 |
`make-regexp'. The default is `regexp/newline', which was previously |
| 9559 |
hard-coded. |
| 9560 |
|
| 9561 |
*** The expect macro now supplies two arguments to a match procedure: |
| 9562 |
the current accumulated string and a flag to indicate whether |
| 9563 |
end-of-file has been reached. Previously only the string was supplied. |
| 9564 |
If end-of-file is reached, the match procedure will be called an |
| 9565 |
additional time with the same accumulated string as the previous call |
| 9566 |
but with the flag set. |
| 9567 |
|
| 9568 |
** New module (ice-9 format), implementing the Common Lisp `format' function. |
| 9569 |
|
| 9570 |
This code, and the documentation for it that appears here, was |
| 9571 |
borrowed from SLIB, with minor adaptations for Guile. |
| 9572 |
|
| 9573 |
- Function: format DESTINATION FORMAT-STRING . ARGUMENTS |
| 9574 |
An almost complete implementation of Common LISP format description |
| 9575 |
according to the CL reference book `Common LISP' from Guy L. |
| 9576 |
Steele, Digital Press. Backward compatible to most of the |
| 9577 |
available Scheme format implementations. |
| 9578 |
|
| 9579 |
Returns `#t', `#f' or a string; has side effect of printing |
| 9580 |
according to FORMAT-STRING. If DESTINATION is `#t', the output is |
| 9581 |
to the current output port and `#t' is returned. If DESTINATION |
| 9582 |
is `#f', a formatted string is returned as the result of the call. |
| 9583 |
NEW: If DESTINATION is a string, DESTINATION is regarded as the |
| 9584 |
format string; FORMAT-STRING is then the first argument and the |
| 9585 |
output is returned as a string. If DESTINATION is a number, the |
| 9586 |
output is to the current error port if available by the |
| 9587 |
implementation. Otherwise DESTINATION must be an output port and |
| 9588 |
`#t' is returned. |
| 9589 |
|
| 9590 |
FORMAT-STRING must be a string. In case of a formatting error |
| 9591 |
format returns `#f' and prints a message on the current output or |
| 9592 |
error port. Characters are output as if the string were output by |
| 9593 |
the `display' function with the exception of those prefixed by a |
| 9594 |
tilde (~). For a detailed description of the FORMAT-STRING syntax |
| 9595 |
please consult a Common LISP format reference manual. For a test |
| 9596 |
suite to verify this format implementation load `formatst.scm'. |
| 9597 |
Please send bug reports to `[email protected]'. |
| 9598 |
|
| 9599 |
Note: `format' is not reentrant, i.e. only one `format'-call may |
| 9600 |
be executed at a time. |
| 9601 |
|
| 9602 |
|
| 9603 |
*** Format Specification (Format version 3.0) |
| 9604 |
|
| 9605 |
Please consult a Common LISP format reference manual for a detailed |
| 9606 |
description of the format string syntax. For a demonstration of the |
| 9607 |
implemented directives see `formatst.scm'. |
| 9608 |
|
| 9609 |
This implementation supports directive parameters and modifiers (`:' |
| 9610 |
and `@' characters). Multiple parameters must be separated by a comma |
| 9611 |
(`,'). Parameters can be numerical parameters (positive or negative), |
| 9612 |
character parameters (prefixed by a quote character (`''), variable |
| 9613 |
parameters (`v'), number of rest arguments parameter (`#'), empty and |
| 9614 |
default parameters. Directive characters are case independent. The |
| 9615 |
general form of a directive is: |
| 9616 |
|
| 9617 |
DIRECTIVE ::= ~{DIRECTIVE-PARAMETER,}[:][@]DIRECTIVE-CHARACTER |
| 9618 |
|
| 9619 |
DIRECTIVE-PARAMETER ::= [ [-|+]{0-9}+ | 'CHARACTER | v | # ] |
| 9620 |
|
| 9621 |
*** Implemented CL Format Control Directives |
| 9622 |
|
| 9623 |
Documentation syntax: Uppercase characters represent the |
| 9624 |
corresponding control directive characters. Lowercase characters |
| 9625 |
represent control directive parameter descriptions. |
| 9626 |
|
| 9627 |
`~A' |
| 9628 |
Any (print as `display' does). |
| 9629 |
`~@A' |
| 9630 |
left pad. |
| 9631 |
|
| 9632 |
`~MINCOL,COLINC,MINPAD,PADCHARA' |
| 9633 |
full padding. |
| 9634 |
|
| 9635 |
`~S' |
| 9636 |
S-expression (print as `write' does). |
| 9637 |
`~@S' |
| 9638 |
left pad. |
| 9639 |
|
| 9640 |
`~MINCOL,COLINC,MINPAD,PADCHARS' |
| 9641 |
full padding. |
| 9642 |
|
| 9643 |
`~D' |
| 9644 |
Decimal. |
| 9645 |
`~@D' |
| 9646 |
print number sign always. |
| 9647 |
|
| 9648 |
`~:D' |
| 9649 |
print comma separated. |
| 9650 |
|
| 9651 |
`~MINCOL,PADCHAR,COMMACHARD' |
| 9652 |
padding. |
| 9653 |
|
| 9654 |
`~X' |
| 9655 |
Hexadecimal. |
| 9656 |
`~@X' |
| 9657 |
print number sign always. |
| 9658 |
|
| 9659 |
`~:X' |
| 9660 |
print comma separated. |
| 9661 |
|
| 9662 |
`~MINCOL,PADCHAR,COMMACHARX' |
| 9663 |
padding. |
| 9664 |
|
| 9665 |
`~O' |
| 9666 |
Octal. |
| 9667 |
`~@O' |
| 9668 |
print number sign always. |
| 9669 |
|
| 9670 |
`~:O' |
| 9671 |
print comma separated. |
| 9672 |
|
| 9673 |
`~MINCOL,PADCHAR,COMMACHARO' |
| 9674 |
padding. |
| 9675 |
|
| 9676 |
`~B' |
| 9677 |
Binary. |
| 9678 |
`~@B' |
| 9679 |
print number sign always. |
| 9680 |
|
| 9681 |
`~:B' |
| 9682 |
print comma separated. |
| 9683 |
|
| 9684 |
`~MINCOL,PADCHAR,COMMACHARB' |
| 9685 |
padding. |
| 9686 |
|
| 9687 |
`~NR' |
| 9688 |
Radix N. |
| 9689 |
`~N,MINCOL,PADCHAR,COMMACHARR' |
| 9690 |
padding. |
| 9691 |
|
| 9692 |
`~@R' |
| 9693 |
print a number as a Roman numeral. |
| 9694 |
|
| 9695 |
`~:@R' |
| 9696 |
print a number as an "old fashioned" Roman numeral. |
| 9697 |
|
| 9698 |
`~:R' |
| 9699 |
print a number as an ordinal English number. |
| 9700 |
|
| 9701 |
`~:@R' |
| 9702 |
print a number as a cardinal English number. |
| 9703 |
|
| 9704 |
`~P' |
| 9705 |
Plural. |
| 9706 |
`~@P' |
| 9707 |
prints `y' and `ies'. |
| 9708 |
|
| 9709 |
`~:P' |
| 9710 |
as `~P but jumps 1 argument backward.' |
| 9711 |
|
| 9712 |
`~:@P' |
| 9713 |
as `~@P but jumps 1 argument backward.' |
| 9714 |
|
| 9715 |
`~C' |
| 9716 |
Character. |
| 9717 |
`~@C' |
| 9718 |
prints a character as the reader can understand it (i.e. `#\' |
| 9719 |
prefixing). |
| 9720 |
|
| 9721 |
`~:C' |
| 9722 |
prints a character as emacs does (eg. `^C' for ASCII 03). |
| 9723 |
|
| 9724 |
`~F' |
| 9725 |
Fixed-format floating-point (prints a flonum like MMM.NNN). |
| 9726 |
`~WIDTH,DIGITS,SCALE,OVERFLOWCHAR,PADCHARF' |
| 9727 |
`~@F' |
| 9728 |
If the number is positive a plus sign is printed. |
| 9729 |
|
| 9730 |
`~E' |
| 9731 |
Exponential floating-point (prints a flonum like MMM.NNN`E'EE). |
| 9732 |
`~WIDTH,DIGITS,EXPONENTDIGITS,SCALE,OVERFLOWCHAR,PADCHAR,EXPONENTCHARE' |
| 9733 |
`~@E' |
| 9734 |
If the number is positive a plus sign is printed. |
| 9735 |
|
| 9736 |
`~G' |
| 9737 |
General floating-point (prints a flonum either fixed or |
| 9738 |
exponential). |
| 9739 |
`~WIDTH,DIGITS,EXPONENTDIGITS,SCALE,OVERFLOWCHAR,PADCHAR,EXPONENTCHARG' |
| 9740 |
`~@G' |
| 9741 |
If the number is positive a plus sign is printed. |
| 9742 |
|
| 9743 |
`~$' |
| 9744 |
Dollars floating-point (prints a flonum in fixed with signs |
| 9745 |
separated). |
| 9746 |
`~DIGITS,SCALE,WIDTH,PADCHAR$' |
| 9747 |
`~@$' |
| 9748 |
If the number is positive a plus sign is printed. |
| 9749 |
|
| 9750 |
`~:@$' |
| 9751 |
A sign is always printed and appears before the padding. |
| 9752 |
|
| 9753 |
`~:$' |
| 9754 |
The sign appears before the padding. |
| 9755 |
|
| 9756 |
`~%' |
| 9757 |
Newline. |
| 9758 |
`~N%' |
| 9759 |
print N newlines. |
| 9760 |
|
| 9761 |
`~&' |
| 9762 |
print newline if not at the beginning of the output line. |
| 9763 |
`~N&' |
| 9764 |
prints `~&' and then N-1 newlines. |
| 9765 |
|
| 9766 |
`~|' |
| 9767 |
Page Separator. |
| 9768 |
`~N|' |
| 9769 |
print N page separators. |
| 9770 |
|
| 9771 |
`~~' |
| 9772 |
Tilde. |
| 9773 |
`~N~' |
| 9774 |
print N tildes. |
| 9775 |
|
| 9776 |
`~'<newline> |
| 9777 |
Continuation Line. |
| 9778 |
`~:'<newline> |
| 9779 |
newline is ignored, white space left. |
| 9780 |
|
| 9781 |
`~@'<newline> |
| 9782 |
newline is left, white space ignored. |
| 9783 |
|
| 9784 |
`~T' |
| 9785 |
Tabulation. |
| 9786 |
`~@T' |
| 9787 |
relative tabulation. |
| 9788 |
|
| 9789 |
`~COLNUM,COLINCT' |
| 9790 |
full tabulation. |
| 9791 |
|
| 9792 |
`~?' |
| 9793 |
Indirection (expects indirect arguments as a list). |
| 9794 |
`~@?' |
| 9795 |
extracts indirect arguments from format arguments. |
| 9796 |
|
| 9797 |
`~(STR~)' |
| 9798 |
Case conversion (converts by `string-downcase'). |
| 9799 |
`~:(STR~)' |
| 9800 |
converts by `string-capitalize'. |
| 9801 |
|
| 9802 |
`~@(STR~)' |
| 9803 |
converts by `string-capitalize-first'. |
| 9804 |
|
| 9805 |
`~:@(STR~)' |
| 9806 |
converts by `string-upcase'. |
| 9807 |
|
| 9808 |
`~*' |
| 9809 |
Argument Jumping (jumps 1 argument forward). |
| 9810 |
`~N*' |
| 9811 |
jumps N arguments forward. |
| 9812 |
|
| 9813 |
`~:*' |
| 9814 |
jumps 1 argument backward. |
| 9815 |
|
| 9816 |
`~N:*' |
| 9817 |
jumps N arguments backward. |
| 9818 |
|
| 9819 |
`~@*' |
| 9820 |
jumps to the 0th argument. |
| 9821 |
|
| 9822 |
`~N@*' |
| 9823 |
jumps to the Nth argument (beginning from 0) |
| 9824 |
|
| 9825 |
`~[STR0~;STR1~;...~;STRN~]' |
| 9826 |
Conditional Expression (numerical clause conditional). |
| 9827 |
`~N[' |
| 9828 |
take argument from N. |
| 9829 |
|
| 9830 |
`~@[' |
| 9831 |
true test conditional. |
| 9832 |
|
| 9833 |
`~:[' |
| 9834 |
if-else-then conditional. |
| 9835 |
|
| 9836 |
`~;' |
| 9837 |
clause separator. |
| 9838 |
|
| 9839 |
`~:;' |
| 9840 |
default clause follows. |
| 9841 |
|
| 9842 |
`~{STR~}' |
| 9843 |
Iteration (args come from the next argument (a list)). |
| 9844 |
`~N{' |
| 9845 |
at most N iterations. |
| 9846 |
|
| 9847 |
`~:{' |
| 9848 |
args from next arg (a list of lists). |
| 9849 |
|
| 9850 |
`~@{' |
| 9851 |
args from the rest of arguments. |
| 9852 |
|
| 9853 |
`~:@{' |
| 9854 |
args from the rest args (lists). |
| 9855 |
|
| 9856 |
`~^' |
| 9857 |
Up and out. |
| 9858 |
`~N^' |
| 9859 |
aborts if N = 0 |
| 9860 |
|
| 9861 |
`~N,M^' |
| 9862 |
aborts if N = M |
| 9863 |
|
| 9864 |
`~N,M,K^' |
| 9865 |
aborts if N <= M <= K |
| 9866 |
|
| 9867 |
*** Not Implemented CL Format Control Directives |
| 9868 |
|
| 9869 |
`~:A' |
| 9870 |
print `#f' as an empty list (see below). |
| 9871 |
|
| 9872 |
`~:S' |
| 9873 |
print `#f' as an empty list (see below). |
| 9874 |
|
| 9875 |
`~<~>' |
| 9876 |
Justification. |
| 9877 |
|
| 9878 |
`~:^' |
| 9879 |
(sorry I don't understand its semantics completely) |
| 9880 |
|
| 9881 |
*** Extended, Replaced and Additional Control Directives |
| 9882 |
|
| 9883 |
`~MINCOL,PADCHAR,COMMACHAR,COMMAWIDTHD' |
| 9884 |
`~MINCOL,PADCHAR,COMMACHAR,COMMAWIDTHX' |
| 9885 |
`~MINCOL,PADCHAR,COMMACHAR,COMMAWIDTHO' |
| 9886 |
`~MINCOL,PADCHAR,COMMACHAR,COMMAWIDTHB' |
| 9887 |
`~N,MINCOL,PADCHAR,COMMACHAR,COMMAWIDTHR' |
| 9888 |
COMMAWIDTH is the number of characters between two comma |
| 9889 |
characters. |
| 9890 |
|
| 9891 |
`~I' |
| 9892 |
print a R4RS complex number as `~F~@Fi' with passed parameters for |
| 9893 |
`~F'. |
| 9894 |
|
| 9895 |
`~Y' |
| 9896 |
Pretty print formatting of an argument for scheme code lists. |
| 9897 |
|
| 9898 |
`~K' |
| 9899 |
Same as `~?.' |
| 9900 |
|
| 9901 |
`~!' |
| 9902 |
Flushes the output if format DESTINATION is a port. |
| 9903 |
|
| 9904 |
`~_' |
| 9905 |
Print a `#\space' character |
| 9906 |
`~N_' |
| 9907 |
print N `#\space' characters. |
| 9908 |
|
| 9909 |
`~/' |
| 9910 |
Print a `#\tab' character |
| 9911 |
`~N/' |
| 9912 |
print N `#\tab' characters. |
| 9913 |
|
| 9914 |
`~NC' |
| 9915 |
Takes N as an integer representation for a character. No arguments |
| 9916 |
are consumed. N is converted to a character by `integer->char'. N |
| 9917 |
must be a positive decimal number. |
| 9918 |
|
| 9919 |
`~:S' |
| 9920 |
Print out readproof. Prints out internal objects represented as |
| 9921 |
`#<...>' as strings `"#<...>"' so that the format output can always |
| 9922 |
be processed by `read'. |
| 9923 |
|
| 9924 |
`~:A' |
| 9925 |
Print out readproof. Prints out internal objects represented as |
| 9926 |
`#<...>' as strings `"#<...>"' so that the format output can always |
| 9927 |
be processed by `read'. |
| 9928 |
|
| 9929 |
`~Q' |
| 9930 |
Prints information and a copyright notice on the format |
| 9931 |
implementation. |
| 9932 |
`~:Q' |
| 9933 |
prints format version. |
| 9934 |
|
| 9935 |
`~F, ~E, ~G, ~$' |
| 9936 |
may also print number strings, i.e. passing a number as a string |
| 9937 |
and format it accordingly. |
| 9938 |
|
| 9939 |
*** Configuration Variables |
| 9940 |
|
| 9941 |
The format module exports some configuration variables to suit the |
| 9942 |
systems and users needs. There should be no modification necessary for |
| 9943 |
the configuration that comes with Guile. Format detects automatically |
| 9944 |
if the running scheme system implements floating point numbers and |
| 9945 |
complex numbers. |
| 9946 |
|
| 9947 |
format:symbol-case-conv |
| 9948 |
Symbols are converted by `symbol->string' so the case type of the |
| 9949 |
printed symbols is implementation dependent. |
| 9950 |
`format:symbol-case-conv' is a one arg closure which is either |
| 9951 |
`#f' (no conversion), `string-upcase', `string-downcase' or |
| 9952 |
`string-capitalize'. (default `#f') |
| 9953 |
|
| 9954 |
format:iobj-case-conv |
| 9955 |
As FORMAT:SYMBOL-CASE-CONV but applies for the representation of |
| 9956 |
implementation internal objects. (default `#f') |
| 9957 |
|
| 9958 |
format:expch |
| 9959 |
The character prefixing the exponent value in `~E' printing. |
| 9960 |
(default `#\E') |
| 9961 |
|
| 9962 |
*** Compatibility With Other Format Implementations |
| 9963 |
|
| 9964 |
SLIB format 2.x: |
| 9965 |
See `format.doc'. |
| 9966 |
|
| 9967 |
SLIB format 1.4: |
| 9968 |
Downward compatible except for padding support and `~A', `~S', |
| 9969 |
`~P', `~X' uppercase printing. SLIB format 1.4 uses C-style |
| 9970 |
`printf' padding support which is completely replaced by the CL |
| 9971 |
`format' padding style. |
| 9972 |
|
| 9973 |
MIT C-Scheme 7.1: |
| 9974 |
Downward compatible except for `~', which is not documented |
| 9975 |
(ignores all characters inside the format string up to a newline |
| 9976 |
character). (7.1 implements `~a', `~s', ~NEWLINE, `~~', `~%', |
| 9977 |
numerical and variable parameters and `:/@' modifiers in the CL |
| 9978 |
sense). |
| 9979 |
|
| 9980 |
Elk 1.5/2.0: |
| 9981 |
Downward compatible except for `~A' and `~S' which print in |
| 9982 |
uppercase. (Elk implements `~a', `~s', `~~', and `~%' (no |
| 9983 |
directive parameters or modifiers)). |
| 9984 |
|
| 9985 |
Scheme->C 01nov91: |
| 9986 |
Downward compatible except for an optional destination parameter: |
| 9987 |
S2C accepts a format call without a destination which returns a |
| 9988 |
formatted string. This is equivalent to a #f destination in S2C. |
| 9989 |
(S2C implements `~a', `~s', `~c', `~%', and `~~' (no directive |
| 9990 |
parameters or modifiers)). |
| 9991 |
|
| 9992 |
|
| 9993 |
** Changes to string-handling functions. |
| 9994 |
|
| 9995 |
These functions were added to support the (ice-9 format) module, above. |
| 9996 |
|
| 9997 |
*** New function: string-upcase STRING |
| 9998 |
*** New function: string-downcase STRING |
| 9999 |
|
| 10000 |
These are non-destructive versions of the existing string-upcase! and |
| 10001 |
string-downcase! functions. |
| 10002 |
|
| 10003 |
*** New function: string-capitalize! STRING |
| 10004 |
*** New function: string-capitalize STRING |
| 10005 |
|
| 10006 |
These functions convert the first letter of each word in the string to |
| 10007 |
upper case. Thus: |
| 10008 |
|
| 10009 |
(string-capitalize "howdy there") |
| 10010 |
=> "Howdy There" |
| 10011 |
|
| 10012 |
As with the other functions, string-capitalize! modifies the string in |
| 10013 |
place, while string-capitalize returns a modified copy of its argument. |
| 10014 |
|
| 10015 |
*** New function: string-ci->symbol STRING |
| 10016 |
|
| 10017 |
Return a symbol whose name is STRING, but having the same case as if |
| 10018 |
the symbol had be read by `read'. |
| 10019 |
|
| 10020 |
Guile can be configured to be sensitive or insensitive to case |
| 10021 |
differences in Scheme identifiers. If Guile is case-insensitive, all |
| 10022 |
symbols are converted to lower case on input. The `string-ci->symbol' |
| 10023 |
function returns a symbol whose name in STRING, transformed as Guile |
| 10024 |
would if STRING were input. |
| 10025 |
|
| 10026 |
*** New function: substring-move! STRING1 START END STRING2 START |
| 10027 |
|
| 10028 |
Copy the substring of STRING1 from START (inclusive) to END |
| 10029 |
(exclusive) to STRING2 at START. STRING1 and STRING2 may be the same |
| 10030 |
string, and the source and destination areas may overlap; in all |
| 10031 |
cases, the function behaves as if all the characters were copied |
| 10032 |
simultanously. |
| 10033 |
|
| 10034 |
*** Extended functions: substring-move-left! substring-move-right! |
| 10035 |
|
| 10036 |
These functions now correctly copy arbitrarily overlapping substrings; |
| 10037 |
they are both synonyms for substring-move!. |
| 10038 |
|
| 10039 |
|
| 10040 |
** New module (ice-9 getopt-long), with the function `getopt-long'. |
| 10041 |
|
| 10042 |
getopt-long is a function for parsing command-line arguments in a |
| 10043 |
manner consistent with other GNU programs. |
| 10044 |
|
| 10045 |
(getopt-long ARGS GRAMMAR) |
| 10046 |
Parse the arguments ARGS according to the argument list grammar GRAMMAR. |
| 10047 |
|
| 10048 |
ARGS should be a list of strings. Its first element should be the |
| 10049 |
name of the program; subsequent elements should be the arguments |
| 10050 |
that were passed to the program on the command line. The |
| 10051 |
`program-arguments' procedure returns a list of this form. |
| 10052 |
|
| 10053 |
GRAMMAR is a list of the form: |
| 10054 |
((OPTION (PROPERTY VALUE) ...) ...) |
| 10055 |
|
| 10056 |
Each OPTION should be a symbol. `getopt-long' will accept a |
| 10057 |
command-line option named `--OPTION'. |
| 10058 |
Each option can have the following (PROPERTY VALUE) pairs: |
| 10059 |
|
| 10060 |
(single-char CHAR) --- Accept `-CHAR' as a single-character |
| 10061 |
equivalent to `--OPTION'. This is how to specify traditional |
| 10062 |
Unix-style flags. |
| 10063 |
(required? BOOL) --- If BOOL is true, the option is required. |
| 10064 |
getopt-long will raise an error if it is not found in ARGS. |
| 10065 |
(value BOOL) --- If BOOL is #t, the option accepts a value; if |
| 10066 |
it is #f, it does not; and if it is the symbol |
| 10067 |
`optional', the option may appear in ARGS with or |
| 10068 |
without a value. |
| 10069 |
(predicate FUNC) --- If the option accepts a value (i.e. you |
| 10070 |
specified `(value #t)' for this option), then getopt |
| 10071 |
will apply FUNC to the value, and throw an exception |
| 10072 |
if it returns #f. FUNC should be a procedure which |
| 10073 |
accepts a string and returns a boolean value; you may |
| 10074 |
need to use quasiquotes to get it into GRAMMAR. |
| 10075 |
|
| 10076 |
The (PROPERTY VALUE) pairs may occur in any order, but each |
| 10077 |
property may occur only once. By default, options do not have |
| 10078 |
single-character equivalents, are not required, and do not take |
| 10079 |
values. |
| 10080 |
|
| 10081 |
In ARGS, single-character options may be combined, in the usual |
| 10082 |
Unix fashion: ("-x" "-y") is equivalent to ("-xy"). If an option |
| 10083 |
accepts values, then it must be the last option in the |
| 10084 |
combination; the value is the next argument. So, for example, using |
| 10085 |
the following grammar: |
| 10086 |
((apples (single-char #\a)) |
| 10087 |
(blimps (single-char #\b) (value #t)) |
| 10088 |
(catalexis (single-char #\c) (value #t))) |
| 10089 |
the following argument lists would be acceptable: |
| 10090 |
("-a" "-b" "bang" "-c" "couth") ("bang" and "couth" are the values |
| 10091 |
for "blimps" and "catalexis") |
| 10092 |
("-ab" "bang" "-c" "couth") (same) |
| 10093 |
("-ac" "couth" "-b" "bang") (same) |
| 10094 |
("-abc" "couth" "bang") (an error, since `-b' is not the |
| 10095 |
last option in its combination) |
| 10096 |
|
| 10097 |
If an option's value is optional, then `getopt-long' decides |
| 10098 |
whether it has a value by looking at what follows it in ARGS. If |
| 10099 |
the next element is a string, and it does not appear to be an |
| 10100 |
option itself, then that string is the option's value. |
| 10101 |
|
| 10102 |
The value of a long option can appear as the next element in ARGS, |
| 10103 |
or it can follow the option name, separated by an `=' character. |
| 10104 |
Thus, using the same grammar as above, the following argument lists |
| 10105 |
are equivalent: |
| 10106 |
("--apples" "Braeburn" "--blimps" "Goodyear") |
| 10107 |
("--apples=Braeburn" "--blimps" "Goodyear") |
| 10108 |
("--blimps" "Goodyear" "--apples=Braeburn") |
| 10109 |
|
| 10110 |
If the option "--" appears in ARGS, argument parsing stops there; |
| 10111 |
subsequent arguments are returned as ordinary arguments, even if |
| 10112 |
they resemble options. So, in the argument list: |
| 10113 |
("--apples" "Granny Smith" "--" "--blimp" "Goodyear") |
| 10114 |
`getopt-long' will recognize the `apples' option as having the |
| 10115 |
value "Granny Smith", but it will not recognize the `blimp' |
| 10116 |
option; it will return the strings "--blimp" and "Goodyear" as |
| 10117 |
ordinary argument strings. |
| 10118 |
|
| 10119 |
The `getopt-long' function returns the parsed argument list as an |
| 10120 |
assocation list, mapping option names --- the symbols from GRAMMAR |
| 10121 |
--- onto their values, or #t if the option does not accept a value. |
| 10122 |
Unused options do not appear in the alist. |
| 10123 |
|
| 10124 |
All arguments that are not the value of any option are returned |
| 10125 |
as a list, associated with the empty list. |
| 10126 |
|
| 10127 |
`getopt-long' throws an exception if: |
| 10128 |
- it finds an unrecognized option in ARGS |
| 10129 |
- a required option is omitted |
| 10130 |
- an option that requires an argument doesn't get one |
| 10131 |
- an option that doesn't accept an argument does get one (this can |
| 10132 |
only happen using the long option `--opt=value' syntax) |
| 10133 |
- an option predicate fails |
| 10134 |
|
| 10135 |
So, for example: |
| 10136 |
|
| 10137 |
(define grammar |
| 10138 |
`((lockfile-dir (required? #t) |
| 10139 |
(value #t) |
| 10140 |
(single-char #\k) |
| 10141 |
(predicate ,file-is-directory?)) |
| 10142 |
(verbose (required? #f) |
| 10143 |
(single-char #\v) |
| 10144 |
(value #f)) |
| 10145 |
(x-includes (single-char #\x)) |
| 10146 |
(rnet-server (single-char #\y) |
| 10147 |
(predicate ,string?)))) |
| 10148 |
|
| 10149 |
(getopt-long '("my-prog" "-vk" "/tmp" "foo1" "--x-includes=/usr/include" |
| 10150 |
"--rnet-server=lamprod" "--" "-fred" "foo2" "foo3") |
| 10151 |
grammar) |
| 10152 |
=> ((() "foo1" "-fred" "foo2" "foo3") |
| 10153 |
(rnet-server . "lamprod") |
| 10154 |
(x-includes . "/usr/include") |
| 10155 |
(lockfile-dir . "/tmp") |
| 10156 |
(verbose . #t)) |
| 10157 |
|
| 10158 |
** The (ice-9 getopt-gnu-style) module is obsolete; use (ice-9 getopt-long). |
| 10159 |
|
| 10160 |
It will be removed in a few releases. |
| 10161 |
|
| 10162 |
** New syntax: lambda* |
| 10163 |
** New syntax: define* |
| 10164 |
** New syntax: define*-public |
| 10165 |
** New syntax: defmacro* |
| 10166 |
** New syntax: defmacro*-public |
| 10167 |
Guile now supports optional arguments. |
| 10168 |
|
| 10169 |
`lambda*', `define*', `define*-public', `defmacro*' and |
| 10170 |
`defmacro*-public' are identical to the non-* versions except that |
| 10171 |
they use an extended type of parameter list that has the following BNF |
| 10172 |
syntax (parentheses are literal, square brackets indicate grouping, |
| 10173 |
and `*', `+' and `?' have the usual meaning): |
| 10174 |
|
| 10175 |
ext-param-list ::= ( [identifier]* [#&optional [ext-var-decl]+]? |
| 10176 |
[#&key [ext-var-decl]+ [#&allow-other-keys]?]? |
| 10177 |
[[#&rest identifier]|[. identifier]]? ) | [identifier] |
| 10178 |
|
| 10179 |
ext-var-decl ::= identifier | ( identifier expression ) |
| 10180 |
|
| 10181 |
The semantics are best illustrated with the following documentation |
| 10182 |
and examples for `lambda*': |
| 10183 |
|
| 10184 |
lambda* args . body |
| 10185 |
lambda extended for optional and keyword arguments |
| 10186 |
|
| 10187 |
lambda* creates a procedure that takes optional arguments. These |
| 10188 |
are specified by putting them inside brackets at the end of the |
| 10189 |
paramater list, but before any dotted rest argument. For example, |
| 10190 |
(lambda* (a b #&optional c d . e) '()) |
| 10191 |
creates a procedure with fixed arguments a and b, optional arguments c |
| 10192 |
and d, and rest argument e. If the optional arguments are omitted |
| 10193 |
in a call, the variables for them are unbound in the procedure. This |
| 10194 |
can be checked with the bound? macro. |
| 10195 |
|
| 10196 |
lambda* can also take keyword arguments. For example, a procedure |
| 10197 |
defined like this: |
| 10198 |
(lambda* (#&key xyzzy larch) '()) |
| 10199 |
can be called with any of the argument lists (#:xyzzy 11) |
| 10200 |
(#:larch 13) (#:larch 42 #:xyzzy 19) (). Whichever arguments |
| 10201 |
are given as keywords are bound to values. |
| 10202 |
|
| 10203 |
Optional and keyword arguments can also be given default values |
| 10204 |
which they take on when they are not present in a call, by giving a |
| 10205 |
two-item list in place of an optional argument, for example in: |
| 10206 |
(lambda* (foo #&optional (bar 42) #&key (baz 73)) (list foo bar baz)) |
| 10207 |
foo is a fixed argument, bar is an optional argument with default |
| 10208 |
value 42, and baz is a keyword argument with default value 73. |
| 10209 |
Default value expressions are not evaluated unless they are needed |
| 10210 |
and until the procedure is called. |
| 10211 |
|
| 10212 |
lambda* now supports two more special parameter list keywords. |
| 10213 |
|
| 10214 |
lambda*-defined procedures now throw an error by default if a |
| 10215 |
keyword other than one of those specified is found in the actual |
| 10216 |
passed arguments. However, specifying #&allow-other-keys |
| 10217 |
immediately after the kyword argument declarations restores the |
| 10218 |
previous behavior of ignoring unknown keywords. lambda* also now |
| 10219 |
guarantees that if the same keyword is passed more than once, the |
| 10220 |
last one passed is the one that takes effect. For example, |
| 10221 |
((lambda* (#&key (heads 0) (tails 0)) (display (list heads tails))) |
| 10222 |
#:heads 37 #:tails 42 #:heads 99) |
| 10223 |
would result in (99 47) being displayed. |
| 10224 |
|
| 10225 |
#&rest is also now provided as a synonym for the dotted syntax rest |
| 10226 |
argument. The argument lists (a . b) and (a #&rest b) are equivalent in |
| 10227 |
all respects to lambda*. This is provided for more similarity to DSSSL, |
| 10228 |
MIT-Scheme and Kawa among others, as well as for refugees from other |
| 10229 |
Lisp dialects. |
| 10230 |
|
| 10231 |
Further documentation may be found in the optargs.scm file itself. |
| 10232 |
|
| 10233 |
The optional argument module also exports the macros `let-optional', |
| 10234 |
`let-optional*', `let-keywords', `let-keywords*' and `bound?'. These |
| 10235 |
are not documented here because they may be removed in the future, but |
| 10236 |
full documentation is still available in optargs.scm. |
| 10237 |
|
| 10238 |
** New syntax: and-let* |
| 10239 |
Guile now supports the `and-let*' form, described in the draft SRFI-2. |
| 10240 |
|
| 10241 |
Syntax: (land* (<clause> ...) <body> ...) |
| 10242 |
Each <clause> should have one of the following forms: |
| 10243 |
(<variable> <expression>) |
| 10244 |
(<expression>) |
| 10245 |
<bound-variable> |
| 10246 |
Each <variable> or <bound-variable> should be an identifier. Each |
| 10247 |
<expression> should be a valid expression. The <body> should be a |
| 10248 |
possibly empty sequence of expressions, like the <body> of a |
| 10249 |
lambda form. |
| 10250 |
|
| 10251 |
Semantics: A LAND* expression is evaluated by evaluating the |
| 10252 |
<expression> or <bound-variable> of each of the <clause>s from |
| 10253 |
left to right. The value of the first <expression> or |
| 10254 |
<bound-variable> that evaluates to a false value is returned; the |
| 10255 |
remaining <expression>s and <bound-variable>s are not evaluated. |
| 10256 |
The <body> forms are evaluated iff all the <expression>s and |
| 10257 |
<bound-variable>s evaluate to true values. |
| 10258 |
|
| 10259 |
The <expression>s and the <body> are evaluated in an environment |
| 10260 |
binding each <variable> of the preceding (<variable> <expression>) |
| 10261 |
clauses to the value of the <expression>. Later bindings |
| 10262 |
shadow earlier bindings. |
| 10263 |
|
| 10264 |
Guile's and-let* macro was contributed by Michael Livshin. |
| 10265 |
|
| 10266 |
** New sorting functions |
| 10267 |
|
| 10268 |
*** New function: sorted? SEQUENCE LESS? |
| 10269 |
Returns `#t' when the sequence argument is in non-decreasing order |
| 10270 |
according to LESS? (that is, there is no adjacent pair `... x y |
| 10271 |
...' for which `(less? y x)'). |
| 10272 |
|
| 10273 |
Returns `#f' when the sequence contains at least one out-of-order |
| 10274 |
pair. It is an error if the sequence is neither a list nor a |
| 10275 |
vector. |
| 10276 |
|
| 10277 |
*** New function: merge LIST1 LIST2 LESS? |
| 10278 |
LIST1 and LIST2 are sorted lists. |
| 10279 |
Returns the sorted list of all elements in LIST1 and LIST2. |
| 10280 |
|
| 10281 |
Assume that the elements a and b1 in LIST1 and b2 in LIST2 are "equal" |
| 10282 |
in the sense that (LESS? x y) --> #f for x, y in {a, b1, b2}, |
| 10283 |
and that a < b1 in LIST1. Then a < b1 < b2 in the result. |
| 10284 |
(Here "<" should read "comes before".) |
| 10285 |
|
| 10286 |
*** New procedure: merge! LIST1 LIST2 LESS? |
| 10287 |
Merges two lists, re-using the pairs of LIST1 and LIST2 to build |
| 10288 |
the result. If the code is compiled, and LESS? constructs no new |
| 10289 |
pairs, no pairs at all will be allocated. The first pair of the |
| 10290 |
result will be either the first pair of LIST1 or the first pair of |
| 10291 |
LIST2. |
| 10292 |
|
| 10293 |
*** New function: sort SEQUENCE LESS? |
| 10294 |
Accepts either a list or a vector, and returns a new sequence |
| 10295 |
which is sorted. The new sequence is the same type as the input. |
| 10296 |
Always `(sorted? (sort sequence less?) less?)'. The original |
| 10297 |
sequence is not altered in any way. The new sequence shares its |
| 10298 |
elements with the old one; no elements are copied. |
| 10299 |
|
| 10300 |
*** New procedure: sort! SEQUENCE LESS |
| 10301 |
Returns its sorted result in the original boxes. No new storage is |
| 10302 |
allocated at all. Proper usage: (set! slist (sort! slist <)) |
| 10303 |
|
| 10304 |
*** New function: stable-sort SEQUENCE LESS? |
| 10305 |
Similar to `sort' but stable. That is, if "equal" elements are |
| 10306 |
ordered a < b in the original sequence, they will have the same order |
| 10307 |
in the result. |
| 10308 |
|
| 10309 |
*** New function: stable-sort! SEQUENCE LESS? |
| 10310 |
Similar to `sort!' but stable. |
| 10311 |
Uses temporary storage when sorting vectors. |
| 10312 |
|
| 10313 |
*** New functions: sort-list, sort-list! |
| 10314 |
Added for compatibility with scsh. |
| 10315 |
|
| 10316 |
** New built-in random number support |
| 10317 |
|
| 10318 |
*** New function: random N [STATE] |
| 10319 |
Accepts a positive integer or real N and returns a number of the |
| 10320 |
same type between zero (inclusive) and N (exclusive). The values |
| 10321 |
returned have a uniform distribution. |
| 10322 |
|
| 10323 |
The optional argument STATE must be of the type produced by |
| 10324 |
`copy-random-state' or `seed->random-state'. It defaults to the value |
| 10325 |
of the variable `*random-state*'. This object is used to maintain the |
| 10326 |
state of the pseudo-random-number generator and is altered as a side |
| 10327 |
effect of the `random' operation. |
| 10328 |
|
| 10329 |
*** New variable: *random-state* |
| 10330 |
Holds a data structure that encodes the internal state of the |
| 10331 |
random-number generator that `random' uses by default. The nature |
| 10332 |
of this data structure is implementation-dependent. It may be |
| 10333 |
printed out and successfully read back in, but may or may not |
| 10334 |
function correctly as a random-number state object in another |
| 10335 |
implementation. |
| 10336 |
|
| 10337 |
*** New function: copy-random-state [STATE] |
| 10338 |
Returns a new object of type suitable for use as the value of the |
| 10339 |
variable `*random-state*' and as a second argument to `random'. |
| 10340 |
If argument STATE is given, a copy of it is returned. Otherwise a |
| 10341 |
copy of `*random-state*' is returned. |
| 10342 |
|
| 10343 |
*** New function: seed->random-state SEED |
| 10344 |
Returns a new object of type suitable for use as the value of the |
| 10345 |
variable `*random-state*' and as a second argument to `random'. |
| 10346 |
SEED is a string or a number. A new state is generated and |
| 10347 |
initialized using SEED. |
| 10348 |
|
| 10349 |
*** New function: random:uniform [STATE] |
| 10350 |
Returns an uniformly distributed inexact real random number in the |
| 10351 |
range between 0 and 1. |
| 10352 |
|
| 10353 |
*** New procedure: random:solid-sphere! VECT [STATE] |
| 10354 |
Fills VECT with inexact real random numbers the sum of whose |
| 10355 |
squares is less than 1.0. Thinking of VECT as coordinates in |
| 10356 |
space of dimension N = `(vector-length VECT)', the coordinates are |
| 10357 |
uniformly distributed within the unit N-shere. The sum of the |
| 10358 |
squares of the numbers is returned. VECT can be either a vector |
| 10359 |
or a uniform vector of doubles. |
| 10360 |
|
| 10361 |
*** New procedure: random:hollow-sphere! VECT [STATE] |
| 10362 |
Fills VECT with inexact real random numbers the sum of whose squares |
| 10363 |
is equal to 1.0. Thinking of VECT as coordinates in space of |
| 10364 |
dimension n = `(vector-length VECT)', the coordinates are uniformly |
| 10365 |
distributed over the surface of the unit n-shere. VECT can be either |
| 10366 |
a vector or a uniform vector of doubles. |
| 10367 |
|
| 10368 |
*** New function: random:normal [STATE] |
| 10369 |
Returns an inexact real in a normal distribution with mean 0 and |
| 10370 |
standard deviation 1. For a normal distribution with mean M and |
| 10371 |
standard deviation D use `(+ M (* D (random:normal)))'. |
| 10372 |
|
| 10373 |
*** New procedure: random:normal-vector! VECT [STATE] |
| 10374 |
Fills VECT with inexact real random numbers which are independent and |
| 10375 |
standard normally distributed (i.e., with mean 0 and variance 1). |
| 10376 |
VECT can be either a vector or a uniform vector of doubles. |
| 10377 |
|
| 10378 |
*** New function: random:exp STATE |
| 10379 |
Returns an inexact real in an exponential distribution with mean 1. |
| 10380 |
For an exponential distribution with mean U use (* U (random:exp)). |
| 10381 |
|
| 10382 |
** The range of logand, logior, logxor, logtest, and logbit? have changed. |
| 10383 |
|
| 10384 |
These functions now operate on numbers in the range of a C unsigned |
| 10385 |
long. |
| 10386 |
|
| 10387 |
These functions used to operate on numbers in the range of a C signed |
| 10388 |
long; however, this seems inappropriate, because Guile integers don't |
| 10389 |
overflow. |
| 10390 |
|
| 10391 |
** New function: make-guardian |
| 10392 |
This is an implementation of guardians as described in |
| 10393 |
R. Kent Dybvig, Carl Bruggeman, and David Eby (1993) "Guardians in a |
| 10394 |
Generation-Based Garbage Collector" ACM SIGPLAN Conference on |
| 10395 |
Programming Language Design and Implementation, June 1993 |
| 10396 |
ftp://ftp.cs.indiana.edu/pub/scheme-repository/doc/pubs/guardians.ps.gz |
| 10397 |
|
| 10398 |
** New functions: delq1!, delv1!, delete1! |
| 10399 |
These procedures behave similar to delq! and friends but delete only |
| 10400 |
one object if at all. |
| 10401 |
|
| 10402 |
** New function: unread-string STRING PORT |
| 10403 |
Unread STRING to PORT, that is, push it back onto the port so that |
| 10404 |
next read operation will work on the pushed back characters. |
| 10405 |
|
| 10406 |
** unread-char can now be called multiple times |
| 10407 |
If unread-char is called multiple times, the unread characters will be |
| 10408 |
read again in last-in first-out order. |
| 10409 |
|
| 10410 |
** the procedures uniform-array-read! and uniform-array-write! now |
| 10411 |
work on any kind of port, not just ports which are open on a file. |
| 10412 |
|
| 10413 |
** Now 'l' in a port mode requests line buffering. |
| 10414 |
|
| 10415 |
** The procedure truncate-file now works on string ports as well |
| 10416 |
as file ports. If the size argument is omitted, the current |
| 10417 |
file position is used. |
| 10418 |
|
| 10419 |
** new procedure: seek PORT/FDES OFFSET WHENCE |
| 10420 |
The arguments are the same as for the old fseek procedure, but it |
| 10421 |
works on string ports as well as random-access file ports. |
| 10422 |
|
| 10423 |
** the fseek procedure now works on string ports, since it has been |
| 10424 |
redefined using seek. |
| 10425 |
|
| 10426 |
** the setvbuf procedure now uses a default size if mode is _IOFBF and |
| 10427 |
size is not supplied. |
| 10428 |
|
| 10429 |
** the newline procedure no longer flushes the port if it's not |
| 10430 |
line-buffered: previously it did if it was the current output port. |
| 10431 |
|
| 10432 |
** open-pipe and close-pipe are no longer primitive procedures, but |
| 10433 |
an emulation can be obtained using `(use-modules (ice-9 popen))'. |
| 10434 |
|
| 10435 |
** the freopen procedure has been removed. |
| 10436 |
|
| 10437 |
** new procedure: drain-input PORT |
| 10438 |
Drains PORT's read buffers (including any pushed-back characters) |
| 10439 |
and returns the contents as a single string. |
| 10440 |
|
| 10441 |
** New function: map-in-order PROC LIST1 LIST2 ... |
| 10442 |
Version of `map' which guarantees that the procedure is applied to the |
| 10443 |
lists in serial order. |
| 10444 |
|
| 10445 |
** Renamed `serial-array-copy!' and `serial-array-map!' to |
| 10446 |
`array-copy-in-order!' and `array-map-in-order!'. The old names are |
| 10447 |
now obsolete and will go away in release 1.5. |
| 10448 |
|
| 10449 |
** New syntax: collect BODY1 ... |
| 10450 |
Version of `begin' which returns a list of the results of the body |
| 10451 |
forms instead of the result of the last body form. In contrast to |
| 10452 |
`begin', `collect' allows an empty body. |
| 10453 |
|
| 10454 |
** New functions: read-history FILENAME, write-history FILENAME |
| 10455 |
Read/write command line history from/to file. Returns #t on success |
| 10456 |
and #f if an error occured. |
| 10457 |
|
| 10458 |
** `ls' and `lls' in module (ice-9 ls) now handle no arguments. |
| 10459 |
|
| 10460 |
These procedures return a list of definitions available in the specified |
| 10461 |
argument, a relative module reference. In the case of no argument, |
| 10462 |
`(current-module)' is now consulted for definitions to return, instead |
| 10463 |
of simply returning #f, the former behavior. |
| 10464 |
|
| 10465 |
** The #/ syntax for lists is no longer supported. |
| 10466 |
|
| 10467 |
Earlier versions of Scheme accepted this syntax, but printed a |
| 10468 |
warning. |
| 10469 |
|
| 10470 |
** Guile no longer consults the SCHEME_LOAD_PATH environment variable. |
| 10471 |
|
| 10472 |
Instead, you should set GUILE_LOAD_PATH to tell Guile where to find |
| 10473 |
modules. |
| 10474 |
|
| 10475 |
* Changes to the gh_ interface |
| 10476 |
|
| 10477 |
** gh_scm2doubles |
| 10478 |
|
| 10479 |
Now takes a second argument which is the result array. If this |
| 10480 |
pointer is NULL, a new array is malloced (the old behaviour). |
| 10481 |
|
| 10482 |
** gh_chars2byvect, gh_shorts2svect, gh_floats2fvect, gh_scm2chars, |
| 10483 |
gh_scm2shorts, gh_scm2longs, gh_scm2floats |
| 10484 |
|
| 10485 |
New functions. |
| 10486 |
|
| 10487 |
* Changes to the scm_ interface |
| 10488 |
|
| 10489 |
** Function: scm_make_named_hook (char* name, int n_args) |
| 10490 |
|
| 10491 |
Creates a hook in the same way as make-hook above but also |
| 10492 |
binds a variable named NAME to it. |
| 10493 |
|
| 10494 |
This is the typical way of creating a hook from C code. |
| 10495 |
|
| 10496 |
Currently, the variable is created in the "current" module. This |
| 10497 |
might change when we get the new module system. |
| 10498 |
|
| 10499 |
** The smob interface |
| 10500 |
|
| 10501 |
The interface for creating smobs has changed. For documentation, see |
| 10502 |
data-rep.info (made from guile-core/doc/data-rep.texi). |
| 10503 |
|
| 10504 |
*** Deprecated function: SCM scm_newsmob (scm_smobfuns *) |
| 10505 |
|
| 10506 |
>>> This function will be removed in 1.3.4. <<< |
| 10507 |
|
| 10508 |
It is replaced by: |
| 10509 |
|
| 10510 |
*** Function: SCM scm_make_smob_type (const char *name, scm_sizet size) |
| 10511 |
This function adds a new smob type, named NAME, with instance size |
| 10512 |
SIZE to the system. The return value is a tag that is used in |
| 10513 |
creating instances of the type. If SIZE is 0, then no memory will |
| 10514 |
be allocated when instances of the smob are created, and nothing |
| 10515 |
will be freed by the default free function. |
| 10516 |
|
| 10517 |
*** Function: void scm_set_smob_mark (long tc, SCM (*mark) (SCM)) |
| 10518 |
This function sets the smob marking procedure for the smob type |
| 10519 |
specified by the tag TC. TC is the tag returned by |
| 10520 |
`scm_make_smob_type'. |
| 10521 |
|
| 10522 |
*** Function: void scm_set_smob_free (long tc, SCM (*mark) (SCM)) |
| 10523 |
This function sets the smob freeing procedure for the smob type |
| 10524 |
specified by the tag TC. TC is the tag returned by |
| 10525 |
`scm_make_smob_type'. |
| 10526 |
|
| 10527 |
*** Function: void scm_set_smob_print (tc, print) |
| 10528 |
|
| 10529 |
- Function: void scm_set_smob_print (long tc, |
| 10530 |
scm_sizet (*print) (SCM, |
| 10531 |
SCM, |
| 10532 |
scm_print_state *)) |
| 10533 |
|
| 10534 |
This function sets the smob printing procedure for the smob type |
| 10535 |
specified by the tag TC. TC is the tag returned by |
| 10536 |
`scm_make_smob_type'. |
| 10537 |
|
| 10538 |
*** Function: void scm_set_smob_equalp (long tc, SCM (*equalp) (SCM, SCM)) |
| 10539 |
This function sets the smob equality-testing predicate for the |
| 10540 |
smob type specified by the tag TC. TC is the tag returned by |
| 10541 |
`scm_make_smob_type'. |
| 10542 |
|
| 10543 |
*** Macro: void SCM_NEWSMOB (SCM var, long tc, void *data) |
| 10544 |
Make VALUE contain a smob instance of the type with type code TC and |
| 10545 |
smob data DATA. VALUE must be previously declared as C type `SCM'. |
| 10546 |
|
| 10547 |
*** Macro: fn_returns SCM_RETURN_NEWSMOB (long tc, void *data) |
| 10548 |
This macro expands to a block of code that creates a smob instance |
| 10549 |
of the type with type code TC and smob data DATA, and returns that |
| 10550 |
`SCM' value. It should be the last piece of code in a block. |
| 10551 |
|
| 10552 |
** The interfaces for using I/O ports and implementing port types |
| 10553 |
(ptobs) have changed significantly. The new interface is based on |
| 10554 |
shared access to buffers and a new set of ptob procedures. |
| 10555 |
|
| 10556 |
*** scm_newptob has been removed |
| 10557 |
|
| 10558 |
It is replaced by: |
| 10559 |
|
| 10560 |
*** Function: SCM scm_make_port_type (type_name, fill_buffer, write_flush) |
| 10561 |
|
| 10562 |
- Function: SCM scm_make_port_type (char *type_name, |
| 10563 |
int (*fill_buffer) (SCM port), |
| 10564 |
void (*write_flush) (SCM port)); |
| 10565 |
|
| 10566 |
Similarly to the new smob interface, there is a set of function |
| 10567 |
setters by which the user can customize the behaviour of his port |
| 10568 |
type. See ports.h (scm_set_port_XXX). |
| 10569 |
|
| 10570 |
** scm_strport_to_string: New function: creates a new string from |
| 10571 |
a string port's buffer. |
| 10572 |
|
| 10573 |
** Plug in interface for random number generators |
| 10574 |
The variable `scm_the_rng' in random.c contains a value and three |
| 10575 |
function pointers which together define the current random number |
| 10576 |
generator being used by the Scheme level interface and the random |
| 10577 |
number library functions. |
| 10578 |
|
| 10579 |
The user is free to replace the default generator with the generator |
| 10580 |
of his own choice. |
| 10581 |
|
| 10582 |
*** Variable: size_t scm_the_rng.rstate_size |
| 10583 |
The size of the random state type used by the current RNG |
| 10584 |
measured in chars. |
| 10585 |
|
| 10586 |
*** Function: unsigned long scm_the_rng.random_bits (scm_rstate *STATE) |
| 10587 |
Given the random STATE, return 32 random bits. |
| 10588 |
|
| 10589 |
*** Function: void scm_the_rng.init_rstate (scm_rstate *STATE, chars *S, int N) |
| 10590 |
Seed random state STATE using string S of length N. |
| 10591 |
|
| 10592 |
*** Function: scm_rstate *scm_the_rng.copy_rstate (scm_rstate *STATE) |
| 10593 |
Given random state STATE, return a malloced copy. |
| 10594 |
|
| 10595 |
** Default RNG |
| 10596 |
The default RNG is the MWC (Multiply With Carry) random number |
| 10597 |
generator described by George Marsaglia at the Department of |
| 10598 |
Statistics and Supercomputer Computations Research Institute, The |
| 10599 |
Florida State University (http://stat.fsu.edu/~geo). |
| 10600 |
|
| 10601 |
It uses 64 bits, has a period of 4578426017172946943 (4.6e18), and |
| 10602 |
passes all tests in the DIEHARD test suite |
| 10603 |
(http://stat.fsu.edu/~geo/diehard.html). The generation of 32 bits |
| 10604 |
costs one multiply and one add on platforms which either supports long |
| 10605 |
longs (gcc does this on most systems) or have 64 bit longs. The cost |
| 10606 |
is four multiply on other systems but this can be optimized by writing |
| 10607 |
scm_i_uniform32 in assembler. |
| 10608 |
|
| 10609 |
These functions are provided through the scm_the_rng interface for use |
| 10610 |
by libguile and the application. |
| 10611 |
|
| 10612 |
*** Function: unsigned long scm_i_uniform32 (scm_i_rstate *STATE) |
| 10613 |
Given the random STATE, return 32 random bits. |
| 10614 |
Don't use this function directly. Instead go through the plugin |
| 10615 |
interface (see "Plug in interface" above). |
| 10616 |
|
| 10617 |
*** Function: void scm_i_init_rstate (scm_i_rstate *STATE, char *SEED, int N) |
| 10618 |
Initialize STATE using SEED of length N. |
| 10619 |
|
| 10620 |
*** Function: scm_i_rstate *scm_i_copy_rstate (scm_i_rstate *STATE) |
| 10621 |
Return a malloc:ed copy of STATE. This function can easily be re-used |
| 10622 |
in the interfaces to other RNGs. |
| 10623 |
|
| 10624 |
** Random number library functions |
| 10625 |
These functions use the current RNG through the scm_the_rng interface. |
| 10626 |
It might be a good idea to use these functions from your C code so |
| 10627 |
that only one random generator is used by all code in your program. |
| 10628 |
|
| 10629 |
The default random state is stored in: |
| 10630 |
|
| 10631 |
*** Variable: SCM scm_var_random_state |
| 10632 |
Contains the vcell of the Scheme variable "*random-state*" which is |
| 10633 |
used as default state by all random number functions in the Scheme |
| 10634 |
level interface. |
| 10635 |
|
| 10636 |
Example: |
| 10637 |
|
| 10638 |
double x = scm_c_uniform01 (SCM_RSTATE (SCM_CDR (scm_var_random_state))); |
| 10639 |
|
| 10640 |
*** Function: scm_rstate *scm_c_default_rstate (void) |
| 10641 |
This is a convenience function which returns the value of |
| 10642 |
scm_var_random_state. An error message is generated if this value |
| 10643 |
isn't a random state. |
| 10644 |
|
| 10645 |
*** Function: scm_rstate *scm_c_make_rstate (char *SEED, int LENGTH) |
| 10646 |
Make a new random state from the string SEED of length LENGTH. |
| 10647 |
|
| 10648 |
It is generally not a good idea to use multiple random states in a |
| 10649 |
program. While subsequent random numbers generated from one random |
| 10650 |
state are guaranteed to be reasonably independent, there is no such |
| 10651 |
guarantee for numbers generated from different random states. |
| 10652 |
|
| 10653 |
*** Macro: unsigned long scm_c_uniform32 (scm_rstate *STATE) |
| 10654 |
Return 32 random bits. |
| 10655 |
|
| 10656 |
*** Function: double scm_c_uniform01 (scm_rstate *STATE) |
| 10657 |
Return a sample from the uniform(0,1) distribution. |
| 10658 |
|
| 10659 |
*** Function: double scm_c_normal01 (scm_rstate *STATE) |
| 10660 |
Return a sample from the normal(0,1) distribution. |
| 10661 |
|
| 10662 |
*** Function: double scm_c_exp1 (scm_rstate *STATE) |
| 10663 |
Return a sample from the exp(1) distribution. |
| 10664 |
|
| 10665 |
*** Function: unsigned long scm_c_random (scm_rstate *STATE, unsigned long M) |
| 10666 |
Return a sample from the discrete uniform(0,M) distribution. |
| 10667 |
|
| 10668 |
*** Function: SCM scm_c_random_bignum (scm_rstate *STATE, SCM M) |
| 10669 |
Return a sample from the discrete uniform(0,M) distribution. |
| 10670 |
M must be a bignum object. The returned value may be an INUM. |
| 10671 |
|
| 10672 |
|
| 10673 |
|
| 10674 |
Changes in Guile 1.3 (released Monday, October 19, 1998): |
| 10675 |
|
| 10676 |
* Changes to the distribution |
| 10677 |
|
| 10678 |
** We renamed the SCHEME_LOAD_PATH environment variable to GUILE_LOAD_PATH. |
| 10679 |
To avoid conflicts, programs should name environment variables after |
| 10680 |
themselves, except when there's a common practice establishing some |
| 10681 |
other convention. |
| 10682 |
|
| 10683 |
For now, Guile supports both GUILE_LOAD_PATH and SCHEME_LOAD_PATH, |
| 10684 |
giving the former precedence, and printing a warning message if the |
| 10685 |
latter is set. Guile 1.4 will not recognize SCHEME_LOAD_PATH at all. |
| 10686 |
|
| 10687 |
** The header files related to multi-byte characters have been removed. |
| 10688 |
They were: libguile/extchrs.h and libguile/mbstrings.h. Any C code |
| 10689 |
which referred to these explicitly will probably need to be rewritten, |
| 10690 |
since the support for the variant string types has been removed; see |
| 10691 |
below. |
| 10692 |
|
| 10693 |
** The header files append.h and sequences.h have been removed. These |
| 10694 |
files implemented non-R4RS operations which would encourage |
| 10695 |
non-portable programming style and less easy-to-read code. |
| 10696 |
|
| 10697 |
* Changes to the stand-alone interpreter |
| 10698 |
|
| 10699 |
** New procedures have been added to implement a "batch mode": |
| 10700 |
|
| 10701 |
*** Function: batch-mode? |
| 10702 |
|
| 10703 |
Returns a boolean indicating whether the interpreter is in batch |
| 10704 |
mode. |
| 10705 |
|
| 10706 |
*** Function: set-batch-mode?! ARG |
| 10707 |
|
| 10708 |
If ARG is true, switches the interpreter to batch mode. The `#f' |
| 10709 |
case has not been implemented. |
| 10710 |
|
| 10711 |
** Guile now provides full command-line editing, when run interactively. |
| 10712 |
To use this feature, you must have the readline library installed. |
| 10713 |
The Guile build process will notice it, and automatically include |
| 10714 |
support for it. |
| 10715 |
|
| 10716 |
The readline library is available via anonymous FTP from any GNU |
| 10717 |
mirror site; the canonical location is "ftp://prep.ai.mit.edu/pub/gnu". |
| 10718 |
|
| 10719 |
** the-last-stack is now a fluid. |
| 10720 |
|
| 10721 |
* Changes to the procedure for linking libguile with your programs |
| 10722 |
|
| 10723 |
** You can now use the `guile-config' utility to build programs that use Guile. |
| 10724 |
|
| 10725 |
Guile now includes a command-line utility called `guile-config', which |
| 10726 |
can provide information about how to compile and link programs that |
| 10727 |
use Guile. |
| 10728 |
|
| 10729 |
*** `guile-config compile' prints any C compiler flags needed to use Guile. |
| 10730 |
You should include this command's output on the command line you use |
| 10731 |
to compile C or C++ code that #includes the Guile header files. It's |
| 10732 |
usually just a `-I' flag to help the compiler find the Guile headers. |
| 10733 |
|
| 10734 |
|
| 10735 |
*** `guile-config link' prints any linker flags necessary to link with Guile. |
| 10736 |
|
| 10737 |
This command writes to its standard output a list of flags which you |
| 10738 |
must pass to the linker to link your code against the Guile library. |
| 10739 |
The flags include '-lguile' itself, any other libraries the Guile |
| 10740 |
library depends upon, and any `-L' flags needed to help the linker |
| 10741 |
find those libraries. |
| 10742 |
|
| 10743 |
For example, here is a Makefile rule that builds a program named 'foo' |
| 10744 |
from the object files ${FOO_OBJECTS}, and links them against Guile: |
| 10745 |
|
| 10746 |
foo: ${FOO_OBJECTS} |
| 10747 |
${CC} ${CFLAGS} ${FOO_OBJECTS} `guile-config link` -o foo |
| 10748 |
|
| 10749 |
Previous Guile releases recommended that you use autoconf to detect |
| 10750 |
which of a predefined set of libraries were present on your system. |
| 10751 |
It is more robust to use `guile-config', since it records exactly which |
| 10752 |
libraries the installed Guile library requires. |
| 10753 |
|
| 10754 |
This was originally called `build-guile', but was renamed to |
| 10755 |
`guile-config' before Guile 1.3 was released, to be consistent with |
| 10756 |
the analogous script for the GTK+ GUI toolkit, which is called |
| 10757 |
`gtk-config'. |
| 10758 |
|
| 10759 |
|
| 10760 |
** Use the GUILE_FLAGS macro in your configure.in file to find Guile. |
| 10761 |
|
| 10762 |
If you are using the GNU autoconf package to configure your program, |
| 10763 |
you can use the GUILE_FLAGS autoconf macro to call `guile-config' |
| 10764 |
(described above) and gather the necessary values for use in your |
| 10765 |
Makefiles. |
| 10766 |
|
| 10767 |
The GUILE_FLAGS macro expands to configure script code which runs the |
| 10768 |
`guile-config' script, to find out where Guile's header files and |
| 10769 |
libraries are installed. It sets two variables, marked for |
| 10770 |
substitution, as by AC_SUBST. |
| 10771 |
|
| 10772 |
GUILE_CFLAGS --- flags to pass to a C or C++ compiler to build |
| 10773 |
code that uses Guile header files. This is almost always just a |
| 10774 |
-I flag. |
| 10775 |
|
| 10776 |
GUILE_LDFLAGS --- flags to pass to the linker to link a |
| 10777 |
program against Guile. This includes `-lguile' for the Guile |
| 10778 |
library itself, any libraries that Guile itself requires (like |
| 10779 |
-lqthreads), and so on. It may also include a -L flag to tell the |
| 10780 |
compiler where to find the libraries. |
| 10781 |
|
| 10782 |
GUILE_FLAGS is defined in the file guile.m4, in the top-level |
| 10783 |
directory of the Guile distribution. You can copy it into your |
| 10784 |
package's aclocal.m4 file, and then use it in your configure.in file. |
| 10785 |
|
| 10786 |
If you are using the `aclocal' program, distributed with GNU automake, |
| 10787 |
to maintain your aclocal.m4 file, the Guile installation process |
| 10788 |
installs guile.m4 where aclocal will find it. All you need to do is |
| 10789 |
use GUILE_FLAGS in your configure.in file, and then run `aclocal'; |
| 10790 |
this will copy the definition of GUILE_FLAGS into your aclocal.m4 |
| 10791 |
file. |
| 10792 |
|
| 10793 |
|
| 10794 |
* Changes to Scheme functions and syntax |
| 10795 |
|
| 10796 |
** Multi-byte strings have been removed, as have multi-byte and wide |
| 10797 |
ports. We felt that these were the wrong approach to |
| 10798 |
internationalization support. |
| 10799 |
|
| 10800 |
** New function: readline [PROMPT] |
| 10801 |
Read a line from the terminal, and allow the user to edit it, |
| 10802 |
prompting with PROMPT. READLINE provides a large set of Emacs-like |
| 10803 |
editing commands, lets the user recall previously typed lines, and |
| 10804 |
works on almost every kind of terminal, including dumb terminals. |
| 10805 |
|
| 10806 |
READLINE assumes that the cursor is at the beginning of the line when |
| 10807 |
it is invoked. Thus, you can't print a prompt yourself, and then call |
| 10808 |
READLINE; you need to package up your prompt as a string, pass it to |
| 10809 |
the function, and let READLINE print the prompt itself. This is |
| 10810 |
because READLINE needs to know the prompt's screen width. |
| 10811 |
|
| 10812 |
For Guile to provide this function, you must have the readline |
| 10813 |
library, version 2.1 or later, installed on your system. Readline is |
| 10814 |
available via anonymous FTP from prep.ai.mit.edu in pub/gnu, or from |
| 10815 |
any GNU mirror site. |
| 10816 |
|
| 10817 |
See also ADD-HISTORY function. |
| 10818 |
|
| 10819 |
** New function: add-history STRING |
| 10820 |
Add STRING as the most recent line in the history used by the READLINE |
| 10821 |
command. READLINE does not add lines to the history itself; you must |
| 10822 |
call ADD-HISTORY to make previous input available to the user. |
| 10823 |
|
| 10824 |
** The behavior of the read-line function has changed. |
| 10825 |
|
| 10826 |
This function now uses standard C library functions to read the line, |
| 10827 |
for speed. This means that it doesn not respect the value of |
| 10828 |
scm-line-incrementors; it assumes that lines are delimited with |
| 10829 |
#\newline. |
| 10830 |
|
| 10831 |
(Note that this is read-line, the function that reads a line of text |
| 10832 |
from a port, not readline, the function that reads a line from a |
| 10833 |
terminal, providing full editing capabilities.) |
| 10834 |
|
| 10835 |
** New module (ice-9 getopt-gnu-style): Parse command-line arguments. |
| 10836 |
|
| 10837 |
This module provides some simple argument parsing. It exports one |
| 10838 |
function: |
| 10839 |
|
| 10840 |
Function: getopt-gnu-style ARG-LS |
| 10841 |
Parse a list of program arguments into an alist of option |
| 10842 |
descriptions. |
| 10843 |
|
| 10844 |
Each item in the list of program arguments is examined to see if |
| 10845 |
it meets the syntax of a GNU long-named option. An argument like |
| 10846 |
`--MUMBLE' produces an element of the form (MUMBLE . #t) in the |
| 10847 |
returned alist, where MUMBLE is a keyword object with the same |
| 10848 |
name as the argument. An argument like `--MUMBLE=FROB' produces |
| 10849 |
an element of the form (MUMBLE . FROB), where FROB is a string. |
| 10850 |
|
| 10851 |
As a special case, the returned alist also contains a pair whose |
| 10852 |
car is the symbol `rest'. The cdr of this pair is a list |
| 10853 |
containing all the items in the argument list that are not options |
| 10854 |
of the form mentioned above. |
| 10855 |
|
| 10856 |
The argument `--' is treated specially: all items in the argument |
| 10857 |
list appearing after such an argument are not examined, and are |
| 10858 |
returned in the special `rest' list. |
| 10859 |
|
| 10860 |
This function does not parse normal single-character switches. |
| 10861 |
You will need to parse them out of the `rest' list yourself. |
| 10862 |
|
| 10863 |
** The read syntax for byte vectors and short vectors has changed. |
| 10864 |
|
| 10865 |
Instead of #bytes(...), write #y(...). |
| 10866 |
|
| 10867 |
Instead of #short(...), write #h(...). |
| 10868 |
|
| 10869 |
This may seem nutty, but, like the other uniform vectors, byte vectors |
| 10870 |
and short vectors want to have the same print and read syntax (and, |
| 10871 |
more basic, want to have read syntax!). Changing the read syntax to |
| 10872 |
use multiple characters after the hash sign breaks with the |
| 10873 |
conventions used in R5RS and the conventions used for the other |
| 10874 |
uniform vectors. It also introduces complexity in the current reader, |
| 10875 |
both on the C and Scheme levels. (The Right solution is probably to |
| 10876 |
change the syntax and prototypes for uniform vectors entirely.) |
| 10877 |
|
| 10878 |
|
| 10879 |
** The new module (ice-9 session) provides useful interactive functions. |
| 10880 |
|
| 10881 |
*** New procedure: (apropos REGEXP OPTION ...) |
| 10882 |
|
| 10883 |
Display a list of top-level variables whose names match REGEXP, and |
| 10884 |
the modules they are imported from. Each OPTION should be one of the |
| 10885 |
following symbols: |
| 10886 |
|
| 10887 |
value --- Show the value of each matching variable. |
| 10888 |
shadow --- Show bindings shadowed by subsequently imported modules. |
| 10889 |
full --- Same as both `shadow' and `value'. |
| 10890 |
|
| 10891 |
For example: |
| 10892 |
|
| 10893 |
guile> (apropos "trace" 'full) |
| 10894 |
debug: trace #<procedure trace args> |
| 10895 |
debug: untrace #<procedure untrace args> |
| 10896 |
the-scm-module: display-backtrace #<compiled-closure #<primitive-procedure gsubr-apply>> |
| 10897 |
the-scm-module: before-backtrace-hook () |
| 10898 |
the-scm-module: backtrace #<primitive-procedure backtrace> |
| 10899 |
the-scm-module: after-backtrace-hook () |
| 10900 |
the-scm-module: has-shown-backtrace-hint? #f |
| 10901 |
guile> |
| 10902 |
|
| 10903 |
** There are new functions and syntax for working with macros. |
| 10904 |
|
| 10905 |
Guile implements macros as a special object type. Any variable whose |
| 10906 |
top-level binding is a macro object acts as a macro. The macro object |
| 10907 |
specifies how the expression should be transformed before evaluation. |
| 10908 |
|
| 10909 |
*** Macro objects now print in a reasonable way, resembling procedures. |
| 10910 |
|
| 10911 |
*** New function: (macro? OBJ) |
| 10912 |
True iff OBJ is a macro object. |
| 10913 |
|
| 10914 |
*** New function: (primitive-macro? OBJ) |
| 10915 |
Like (macro? OBJ), but true only if OBJ is one of the Guile primitive |
| 10916 |
macro transformers, implemented in eval.c rather than Scheme code. |
| 10917 |
|
| 10918 |
Why do we have this function? |
| 10919 |
- For symmetry with procedure? and primitive-procedure?, |
| 10920 |
- to allow custom print procedures to tell whether a macro is |
| 10921 |
primitive, and display it differently, and |
| 10922 |
- to allow compilers and user-written evaluators to distinguish |
| 10923 |
builtin special forms from user-defined ones, which could be |
| 10924 |
compiled. |
| 10925 |
|
| 10926 |
*** New function: (macro-type OBJ) |
| 10927 |
Return a value indicating what kind of macro OBJ is. Possible return |
| 10928 |
values are: |
| 10929 |
|
| 10930 |
The symbol `syntax' --- a macro created by procedure->syntax. |
| 10931 |
The symbol `macro' --- a macro created by procedure->macro. |
| 10932 |
The symbol `macro!' --- a macro created by procedure->memoizing-macro. |
| 10933 |
The boolean #f --- if OBJ is not a macro object. |
| 10934 |
|
| 10935 |
*** New function: (macro-name MACRO) |
| 10936 |
Return the name of the macro object MACRO's procedure, as returned by |
| 10937 |
procedure-name. |
| 10938 |
|
| 10939 |
*** New function: (macro-transformer MACRO) |
| 10940 |
Return the transformer procedure for MACRO. |
| 10941 |
|
| 10942 |
*** New syntax: (use-syntax MODULE ... TRANSFORMER) |
| 10943 |
|
| 10944 |
Specify a new macro expander to use in the current module. Each |
| 10945 |
MODULE is a module name, with the same meaning as in the `use-modules' |
| 10946 |
form; each named module's exported bindings are added to the current |
| 10947 |
top-level environment. TRANSFORMER is an expression evaluated in the |
| 10948 |
resulting environment which must yield a procedure to use as the |
| 10949 |
module's eval transformer: every expression evaluated in this module |
| 10950 |
is passed to this function, and the result passed to the Guile |
| 10951 |
interpreter. |
| 10952 |
|
| 10953 |
*** macro-eval! is removed. Use local-eval instead. |
| 10954 |
|
| 10955 |
** Some magic has been added to the printer to better handle user |
| 10956 |
written printing routines (like record printers, closure printers). |
| 10957 |
|
| 10958 |
The problem is that these user written routines must have access to |
| 10959 |
the current `print-state' to be able to handle fancy things like |
| 10960 |
detection of circular references. These print-states have to be |
| 10961 |
passed to the builtin printing routines (display, write, etc) to |
| 10962 |
properly continue the print chain. |
| 10963 |
|
| 10964 |
We didn't want to change all existing print code so that it |
| 10965 |
explicitly passes thru a print state in addition to a port. Instead, |
| 10966 |
we extented the possible values that the builtin printing routines |
| 10967 |
accept as a `port'. In addition to a normal port, they now also take |
| 10968 |
a pair of a normal port and a print-state. Printing will go to the |
| 10969 |
port and the print-state will be used to control the detection of |
| 10970 |
circular references, etc. If the builtin function does not care for a |
| 10971 |
print-state, it is simply ignored. |
| 10972 |
|
| 10973 |
User written callbacks are now called with such a pair as their |
| 10974 |
`port', but because every function now accepts this pair as a PORT |
| 10975 |
argument, you don't have to worry about that. In fact, it is probably |
| 10976 |
safest to not check for these pairs. |
| 10977 |
|
| 10978 |
However, it is sometimes necessary to continue a print chain on a |
| 10979 |
different port, for example to get a intermediate string |
| 10980 |
representation of the printed value, mangle that string somehow, and |
| 10981 |
then to finally print the mangled string. Use the new function |
| 10982 |
|
| 10983 |
inherit-print-state OLD-PORT NEW-PORT |
| 10984 |
|
| 10985 |
for this. It constructs a new `port' that prints to NEW-PORT but |
| 10986 |
inherits the print-state of OLD-PORT. |
| 10987 |
|
| 10988 |
** struct-vtable-offset renamed to vtable-offset-user |
| 10989 |
|
| 10990 |
** New constants: vtable-index-layout, vtable-index-vtable, vtable-index-printer |
| 10991 |
|
| 10992 |
** There is now a third optional argument to make-vtable-vtable |
| 10993 |
(and fourth to make-struct) when constructing new types (vtables). |
| 10994 |
This argument initializes field vtable-index-printer of the vtable. |
| 10995 |
|
| 10996 |
** The detection of circular references has been extended to structs. |
| 10997 |
That is, a structure that -- in the process of being printed -- prints |
| 10998 |
itself does not lead to infinite recursion. |
| 10999 |
|
| 11000 |
** There is now some basic support for fluids. Please read |
| 11001 |
"libguile/fluid.h" to find out more. It is accessible from Scheme with |
| 11002 |
the following functions and macros: |
| 11003 |
|
| 11004 |
Function: make-fluid |
| 11005 |
|
| 11006 |
Create a new fluid object. Fluids are not special variables or |
| 11007 |
some other extension to the semantics of Scheme, but rather |
| 11008 |
ordinary Scheme objects. You can store them into variables (that |
| 11009 |
are still lexically scoped, of course) or into any other place you |
| 11010 |
like. Every fluid has a initial value of `#f'. |
| 11011 |
|
| 11012 |
Function: fluid? OBJ |
| 11013 |
|
| 11014 |
Test whether OBJ is a fluid. |
| 11015 |
|
| 11016 |
Function: fluid-ref FLUID |
| 11017 |
Function: fluid-set! FLUID VAL |
| 11018 |
|
| 11019 |
Access/modify the fluid FLUID. Modifications are only visible |
| 11020 |
within the current dynamic root (that includes threads). |
| 11021 |
|
| 11022 |
Function: with-fluids* FLUIDS VALUES THUNK |
| 11023 |
|
| 11024 |
FLUIDS is a list of fluids and VALUES a corresponding list of |
| 11025 |
values for these fluids. Before THUNK gets called the values are |
| 11026 |
installed in the fluids and the old values of the fluids are |
| 11027 |
saved in the VALUES list. When the flow of control leaves THUNK |
| 11028 |
or reenters it, the values get swapped again. You might think of |
| 11029 |
this as a `safe-fluid-excursion'. Note that the VALUES list is |
| 11030 |
modified by `with-fluids*'. |
| 11031 |
|
| 11032 |
Macro: with-fluids ((FLUID VALUE) ...) FORM ... |
| 11033 |
|
| 11034 |
The same as `with-fluids*' but with a different syntax. It looks |
| 11035 |
just like `let', but both FLUID and VALUE are evaluated. Remember, |
| 11036 |
fluids are not special variables but ordinary objects. FLUID |
| 11037 |
should evaluate to a fluid. |
| 11038 |
|
| 11039 |
** Changes to system call interfaces: |
| 11040 |
|
| 11041 |
*** close-port, close-input-port and close-output-port now return a |
| 11042 |
boolean instead of an `unspecified' object. #t means that the port |
| 11043 |
was successfully closed, while #f means it was already closed. It is |
| 11044 |
also now possible for these procedures to raise an exception if an |
| 11045 |
error occurs (some errors from write can be delayed until close.) |
| 11046 |
|
| 11047 |
*** the first argument to chmod, fcntl, ftell and fseek can now be a |
| 11048 |
file descriptor. |
| 11049 |
|
| 11050 |
*** the third argument to fcntl is now optional. |
| 11051 |
|
| 11052 |
*** the first argument to chown can now be a file descriptor or a port. |
| 11053 |
|
| 11054 |
*** the argument to stat can now be a port. |
| 11055 |
|
| 11056 |
*** The following new procedures have been added (most use scsh |
| 11057 |
interfaces): |
| 11058 |
|
| 11059 |
*** procedure: close PORT/FD |
| 11060 |
Similar to close-port (*note close-port: Closing Ports.), but also |
| 11061 |
works on file descriptors. A side effect of closing a file |
| 11062 |
descriptor is that any ports using that file descriptor are moved |
| 11063 |
to a different file descriptor and have their revealed counts set |
| 11064 |
to zero. |
| 11065 |
|
| 11066 |
*** procedure: port->fdes PORT |
| 11067 |
Returns the integer file descriptor underlying PORT. As a side |
| 11068 |
effect the revealed count of PORT is incremented. |
| 11069 |
|
| 11070 |
*** procedure: fdes->ports FDES |
| 11071 |
Returns a list of existing ports which have FDES as an underlying |
| 11072 |
file descriptor, without changing their revealed counts. |
| 11073 |
|
| 11074 |
*** procedure: fdes->inport FDES |
| 11075 |
Returns an existing input port which has FDES as its underlying |
| 11076 |
file descriptor, if one exists, and increments its revealed count. |
| 11077 |
Otherwise, returns a new input port with a revealed count of 1. |
| 11078 |
|
| 11079 |
*** procedure: fdes->outport FDES |
| 11080 |
Returns an existing output port which has FDES as its underlying |
| 11081 |
file descriptor, if one exists, and increments its revealed count. |
| 11082 |
Otherwise, returns a new output port with a revealed count of 1. |
| 11083 |
|
| 11084 |
The next group of procedures perform a `dup2' system call, if NEWFD |
| 11085 |
(an integer) is supplied, otherwise a `dup'. The file descriptor to be |
| 11086 |
duplicated can be supplied as an integer or contained in a port. The |
| 11087 |
type of value returned varies depending on which procedure is used. |
| 11088 |
|
| 11089 |
All procedures also have the side effect when performing `dup2' that |
| 11090 |
any ports using NEWFD are moved to a different file descriptor and have |
| 11091 |
their revealed counts set to zero. |
| 11092 |
|
| 11093 |
*** procedure: dup->fdes PORT/FD [NEWFD] |
| 11094 |
Returns an integer file descriptor. |
| 11095 |
|
| 11096 |
*** procedure: dup->inport PORT/FD [NEWFD] |
| 11097 |
Returns a new input port using the new file descriptor. |
| 11098 |
|
| 11099 |
*** procedure: dup->outport PORT/FD [NEWFD] |
| 11100 |
Returns a new output port using the new file descriptor. |
| 11101 |
|
| 11102 |
*** procedure: dup PORT/FD [NEWFD] |
| 11103 |
Returns a new port if PORT/FD is a port, with the same mode as the |
| 11104 |
supplied port, otherwise returns an integer file descriptor. |
| 11105 |
|
| 11106 |
*** procedure: dup->port PORT/FD MODE [NEWFD] |
| 11107 |
Returns a new port using the new file descriptor. MODE supplies a |
| 11108 |
mode string for the port (*note open-file: File Ports.). |
| 11109 |
|
| 11110 |
*** procedure: setenv NAME VALUE |
| 11111 |
Modifies the environment of the current process, which is also the |
| 11112 |
default environment inherited by child processes. |
| 11113 |
|
| 11114 |
If VALUE is `#f', then NAME is removed from the environment. |
| 11115 |
Otherwise, the string NAME=VALUE is added to the environment, |
| 11116 |
replacing any existing string with name matching NAME. |
| 11117 |
|
| 11118 |
The return value is unspecified. |
| 11119 |
|
| 11120 |
*** procedure: truncate-file OBJ SIZE |
| 11121 |
Truncates the file referred to by OBJ to at most SIZE bytes. OBJ |
| 11122 |
can be a string containing a file name or an integer file |
| 11123 |
descriptor or port open for output on the file. The underlying |
| 11124 |
system calls are `truncate' and `ftruncate'. |
| 11125 |
|
| 11126 |
The return value is unspecified. |
| 11127 |
|
| 11128 |
*** procedure: setvbuf PORT MODE [SIZE] |
| 11129 |
Set the buffering mode for PORT. MODE can be: |
| 11130 |
`_IONBF' |
| 11131 |
non-buffered |
| 11132 |
|
| 11133 |
`_IOLBF' |
| 11134 |
line buffered |
| 11135 |
|
| 11136 |
`_IOFBF' |
| 11137 |
block buffered, using a newly allocated buffer of SIZE bytes. |
| 11138 |
However if SIZE is zero or unspecified, the port will be made |
| 11139 |
non-buffered. |
| 11140 |
|
| 11141 |
This procedure should not be used after I/O has been performed with |
| 11142 |
the port. |
| 11143 |
|
| 11144 |
Ports are usually block buffered by default, with a default buffer |
| 11145 |
size. Procedures e.g., *Note open-file: File Ports, which accept a |
| 11146 |
mode string allow `0' to be added to request an unbuffered port. |
| 11147 |
|
| 11148 |
*** procedure: fsync PORT/FD |
| 11149 |
Copies any unwritten data for the specified output file descriptor |
| 11150 |
to disk. If PORT/FD is a port, its buffer is flushed before the |
| 11151 |
underlying file descriptor is fsync'd. The return value is |
| 11152 |
unspecified. |
| 11153 |
|
| 11154 |
*** procedure: open-fdes PATH FLAGS [MODES] |
| 11155 |
Similar to `open' but returns a file descriptor instead of a port. |
| 11156 |
|
| 11157 |
*** procedure: execle PATH ENV [ARG] ... |
| 11158 |
Similar to `execl', but the environment of the new process is |
| 11159 |
specified by ENV, which must be a list of strings as returned by |
| 11160 |
the `environ' procedure. |
| 11161 |
|
| 11162 |
This procedure is currently implemented using the `execve' system |
| 11163 |
call, but we call it `execle' because of its Scheme calling |
| 11164 |
interface. |
| 11165 |
|
| 11166 |
*** procedure: strerror ERRNO |
| 11167 |
Returns the Unix error message corresponding to ERRNO, an integer. |
| 11168 |
|
| 11169 |
*** procedure: primitive-exit [STATUS] |
| 11170 |
Terminate the current process without unwinding the Scheme stack. |
| 11171 |
This is would typically be useful after a fork. The exit status |
| 11172 |
is STATUS if supplied, otherwise zero. |
| 11173 |
|
| 11174 |
*** procedure: times |
| 11175 |
Returns an object with information about real and processor time. |
| 11176 |
The following procedures accept such an object as an argument and |
| 11177 |
return a selected component: |
| 11178 |
|
| 11179 |
`tms:clock' |
| 11180 |
The current real time, expressed as time units relative to an |
| 11181 |
arbitrary base. |
| 11182 |
|
| 11183 |
`tms:utime' |
| 11184 |
The CPU time units used by the calling process. |
| 11185 |
|
| 11186 |
`tms:stime' |
| 11187 |
The CPU time units used by the system on behalf of the |
| 11188 |
calling process. |
| 11189 |
|
| 11190 |
`tms:cutime' |
| 11191 |
The CPU time units used by terminated child processes of the |
| 11192 |
calling process, whose status has been collected (e.g., using |
| 11193 |
`waitpid'). |
| 11194 |
|
| 11195 |
`tms:cstime' |
| 11196 |
Similarly, the CPU times units used by the system on behalf of |
| 11197 |
terminated child processes. |
| 11198 |
|
| 11199 |
** Removed: list-length |
| 11200 |
** Removed: list-append, list-append! |
| 11201 |
** Removed: list-reverse, list-reverse! |
| 11202 |
|
| 11203 |
** array-map renamed to array-map! |
| 11204 |
|
| 11205 |
** serial-array-map renamed to serial-array-map! |
| 11206 |
|
| 11207 |
** catch doesn't take #f as first argument any longer |
| 11208 |
|
| 11209 |
Previously, it was possible to pass #f instead of a key to `catch'. |
| 11210 |
That would cause `catch' to pass a jump buffer object to the procedure |
| 11211 |
passed as second argument. The procedure could then use this jump |
| 11212 |
buffer objekt as an argument to throw. |
| 11213 |
|
| 11214 |
This mechanism has been removed since its utility doesn't motivate the |
| 11215 |
extra complexity it introduces. |
| 11216 |
|
| 11217 |
** The `#/' notation for lists now provokes a warning message from Guile. |
| 11218 |
This syntax will be removed from Guile in the near future. |
| 11219 |
|
| 11220 |
To disable the warning message, set the GUILE_HUSH environment |
| 11221 |
variable to any non-empty value. |
| 11222 |
|
| 11223 |
** The newline character now prints as `#\newline', following the |
| 11224 |
normal Scheme notation, not `#\nl'. |
| 11225 |
|
| 11226 |
* Changes to the gh_ interface |
| 11227 |
|
| 11228 |
** The gh_enter function now takes care of loading the Guile startup files. |
| 11229 |
gh_enter works by calling scm_boot_guile; see the remarks below. |
| 11230 |
|
| 11231 |
** Function: void gh_write (SCM x) |
| 11232 |
|
| 11233 |
Write the printed representation of the scheme object x to the current |
| 11234 |
output port. Corresponds to the scheme level `write'. |
| 11235 |
|
| 11236 |
** gh_list_length renamed to gh_length. |
| 11237 |
|
| 11238 |
** vector handling routines |
| 11239 |
|
| 11240 |
Several major changes. In particular, gh_vector() now resembles |
| 11241 |
(vector ...) (with a caveat -- see manual), and gh_make_vector() now |
| 11242 |
exists and behaves like (make-vector ...). gh_vset() and gh_vref() |
| 11243 |
have been renamed gh_vector_set_x() and gh_vector_ref(). Some missing |
| 11244 |
vector-related gh_ functions have been implemented. |
| 11245 |
|
| 11246 |
** pair and list routines |
| 11247 |
|
| 11248 |
Implemented several of the R4RS pair and list functions that were |
| 11249 |
missing. |
| 11250 |
|
| 11251 |
** gh_scm2doubles, gh_doubles2scm, gh_doubles2dvect |
| 11252 |
|
| 11253 |
New function. Converts double arrays back and forth between Scheme |
| 11254 |
and C. |
| 11255 |
|
| 11256 |
* Changes to the scm_ interface |
| 11257 |
|
| 11258 |
** The function scm_boot_guile now takes care of loading the startup files. |
| 11259 |
|
| 11260 |
Guile's primary initialization function, scm_boot_guile, now takes |
| 11261 |
care of loading `boot-9.scm', in the `ice-9' module, to initialize |
| 11262 |
Guile, define the module system, and put together some standard |
| 11263 |
bindings. It also loads `init.scm', which is intended to hold |
| 11264 |
site-specific initialization code. |
| 11265 |
|
| 11266 |
Since Guile cannot operate properly until boot-9.scm is loaded, there |
| 11267 |
is no reason to separate loading boot-9.scm from Guile's other |
| 11268 |
initialization processes. |
| 11269 |
|
| 11270 |
This job used to be done by scm_compile_shell_switches, which didn't |
| 11271 |
make much sense; in particular, it meant that people using Guile for |
| 11272 |
non-shell-like applications had to jump through hoops to get Guile |
| 11273 |
initialized properly. |
| 11274 |
|
| 11275 |
** The function scm_compile_shell_switches no longer loads the startup files. |
| 11276 |
Now, Guile always loads the startup files, whenever it is initialized; |
| 11277 |
see the notes above for scm_boot_guile and scm_load_startup_files. |
| 11278 |
|
| 11279 |
** Function: scm_load_startup_files |
| 11280 |
This new function takes care of loading Guile's initialization file |
| 11281 |
(`boot-9.scm'), and the site initialization file, `init.scm'. Since |
| 11282 |
this is always called by the Guile initialization process, it's |
| 11283 |
probably not too useful to call this yourself, but it's there anyway. |
| 11284 |
|
| 11285 |
** The semantics of smob marking have changed slightly. |
| 11286 |
|
| 11287 |
The smob marking function (the `mark' member of the scm_smobfuns |
| 11288 |
structure) is no longer responsible for setting the mark bit on the |
| 11289 |
smob. The generic smob handling code in the garbage collector will |
| 11290 |
set this bit. The mark function need only ensure that any other |
| 11291 |
objects the smob refers to get marked. |
| 11292 |
|
| 11293 |
Note that this change means that the smob's GC8MARK bit is typically |
| 11294 |
already set upon entry to the mark function. Thus, marking functions |
| 11295 |
which look like this: |
| 11296 |
|
| 11297 |
{ |
| 11298 |
if (SCM_GC8MARKP (ptr)) |
| 11299 |
return SCM_BOOL_F; |
| 11300 |
SCM_SETGC8MARK (ptr); |
| 11301 |
... mark objects to which the smob refers ... |
| 11302 |
} |
| 11303 |
|
| 11304 |
are now incorrect, since they will return early, and fail to mark any |
| 11305 |
other objects the smob refers to. Some code in the Guile library used |
| 11306 |
to work this way. |
| 11307 |
|
| 11308 |
** The semantics of the I/O port functions in scm_ptobfuns have changed. |
| 11309 |
|
| 11310 |
If you have implemented your own I/O port type, by writing the |
| 11311 |
functions required by the scm_ptobfuns and then calling scm_newptob, |
| 11312 |
you will need to change your functions slightly. |
| 11313 |
|
| 11314 |
The functions in a scm_ptobfuns structure now expect the port itself |
| 11315 |
as their argument; they used to expect the `stream' member of the |
| 11316 |
port's scm_port_table structure. This allows functions in an |
| 11317 |
scm_ptobfuns structure to easily access the port's cell (and any flags |
| 11318 |
it its CAR), and the port's scm_port_table structure. |
| 11319 |
|
| 11320 |
Guile now passes the I/O port itself as the `port' argument in the |
| 11321 |
following scm_ptobfuns functions: |
| 11322 |
|
| 11323 |
int (*free) (SCM port); |
| 11324 |
int (*fputc) (int, SCM port); |
| 11325 |
int (*fputs) (char *, SCM port); |
| 11326 |
scm_sizet (*fwrite) SCM_P ((char *ptr, |
| 11327 |
scm_sizet size, |
| 11328 |
scm_sizet nitems, |
| 11329 |
SCM port)); |
| 11330 |
int (*fflush) (SCM port); |
| 11331 |
int (*fgetc) (SCM port); |
| 11332 |
int (*fclose) (SCM port); |
| 11333 |
|
| 11334 |
The interfaces to the `mark', `print', `equalp', and `fgets' methods |
| 11335 |
are unchanged. |
| 11336 |
|
| 11337 |
If you have existing code which defines its own port types, it is easy |
| 11338 |
to convert your code to the new interface; simply apply SCM_STREAM to |
| 11339 |
the port argument to yield the value you code used to expect. |
| 11340 |
|
| 11341 |
Note that since both the port and the stream have the same type in the |
| 11342 |
C code --- they are both SCM values --- the C compiler will not remind |
| 11343 |
you if you forget to update your scm_ptobfuns functions. |
| 11344 |
|
| 11345 |
|
| 11346 |
** Function: int scm_internal_select (int fds, |
| 11347 |
SELECT_TYPE *rfds, |
| 11348 |
SELECT_TYPE *wfds, |
| 11349 |
SELECT_TYPE *efds, |
| 11350 |
struct timeval *timeout); |
| 11351 |
|
| 11352 |
This is a replacement for the `select' function provided by the OS. |
| 11353 |
It enables I/O blocking and sleeping to happen for one cooperative |
| 11354 |
thread without blocking other threads. It also avoids busy-loops in |
| 11355 |
these situations. It is intended that all I/O blocking and sleeping |
| 11356 |
will finally go through this function. Currently, this function is |
| 11357 |
only available on systems providing `gettimeofday' and `select'. |
| 11358 |
|
| 11359 |
** Function: SCM scm_internal_stack_catch (SCM tag, |
| 11360 |
scm_catch_body_t body, |
| 11361 |
void *body_data, |
| 11362 |
scm_catch_handler_t handler, |
| 11363 |
void *handler_data) |
| 11364 |
|
| 11365 |
A new sibling to the other two C level `catch' functions |
| 11366 |
scm_internal_catch and scm_internal_lazy_catch. Use it if you want |
| 11367 |
the stack to be saved automatically into the variable `the-last-stack' |
| 11368 |
(scm_the_last_stack_var) on error. This is necessary if you want to |
| 11369 |
use advanced error reporting, such as calling scm_display_error and |
| 11370 |
scm_display_backtrace. (They both take a stack object as argument.) |
| 11371 |
|
| 11372 |
** Function: SCM scm_spawn_thread (scm_catch_body_t body, |
| 11373 |
void *body_data, |
| 11374 |
scm_catch_handler_t handler, |
| 11375 |
void *handler_data) |
| 11376 |
|
| 11377 |
Spawns a new thread. It does a job similar to |
| 11378 |
scm_call_with_new_thread but takes arguments more suitable when |
| 11379 |
spawning threads from application C code. |
| 11380 |
|
| 11381 |
** The hook scm_error_callback has been removed. It was originally |
| 11382 |
intended as a way for the user to install his own error handler. But |
| 11383 |
that method works badly since it intervenes between throw and catch, |
| 11384 |
thereby changing the semantics of expressions like (catch #t ...). |
| 11385 |
The correct way to do it is to use one of the C level catch functions |
| 11386 |
in throw.c: scm_internal_catch/lazy_catch/stack_catch. |
| 11387 |
|
| 11388 |
** Removed functions: |
| 11389 |
|
| 11390 |
scm_obj_length, scm_list_length, scm_list_append, scm_list_append_x, |
| 11391 |
scm_list_reverse, scm_list_reverse_x |
| 11392 |
|
| 11393 |
** New macros: SCM_LISTn where n is one of the integers 0-9. |
| 11394 |
|
| 11395 |
These can be used for pretty list creation from C. The idea is taken |
| 11396 |
from Erick Gallesio's STk. |
| 11397 |
|
| 11398 |
** scm_array_map renamed to scm_array_map_x |
| 11399 |
|
| 11400 |
** mbstrings are now removed |
| 11401 |
|
| 11402 |
This means that the type codes scm_tc7_mb_string and |
| 11403 |
scm_tc7_mb_substring has been removed. |
| 11404 |
|
| 11405 |
** scm_gen_putc, scm_gen_puts, scm_gen_write, and scm_gen_getc have changed. |
| 11406 |
|
| 11407 |
Since we no longer support multi-byte strings, these I/O functions |
| 11408 |
have been simplified, and renamed. Here are their old names, and |
| 11409 |
their new names and arguments: |
| 11410 |
|
| 11411 |
scm_gen_putc -> void scm_putc (int c, SCM port); |
| 11412 |
scm_gen_puts -> void scm_puts (char *s, SCM port); |
| 11413 |
scm_gen_write -> void scm_lfwrite (char *ptr, scm_sizet size, SCM port); |
| 11414 |
scm_gen_getc -> void scm_getc (SCM port); |
| 11415 |
|
| 11416 |
|
| 11417 |
** The macros SCM_TYP7D and SCM_TYP7SD has been removed. |
| 11418 |
|
| 11419 |
** The macro SCM_TYP7S has taken the role of the old SCM_TYP7D |
| 11420 |
|
| 11421 |
SCM_TYP7S now masks away the bit which distinguishes substrings from |
| 11422 |
strings. |
| 11423 |
|
| 11424 |
** scm_catch_body_t: Backward incompatible change! |
| 11425 |
|
| 11426 |
Body functions to scm_internal_catch and friends do not any longer |
| 11427 |
take a second argument. This is because it is no longer possible to |
| 11428 |
pass a #f arg to catch. |
| 11429 |
|
| 11430 |
** Calls to scm_protect_object and scm_unprotect now nest properly. |
| 11431 |
|
| 11432 |
The function scm_protect_object protects its argument from being freed |
| 11433 |
by the garbage collector. scm_unprotect_object removes that |
| 11434 |
protection. |
| 11435 |
|
| 11436 |
These functions now nest properly. That is, for every object O, there |
| 11437 |
is a counter which scm_protect_object(O) increments and |
| 11438 |
scm_unprotect_object(O) decrements, if the counter is greater than |
| 11439 |
zero. Every object's counter is zero when it is first created. If an |
| 11440 |
object's counter is greater than zero, the garbage collector will not |
| 11441 |
reclaim its storage. |
| 11442 |
|
| 11443 |
This allows you to use scm_protect_object in your code without |
| 11444 |
worrying that some other function you call will call |
| 11445 |
scm_unprotect_object, and allow it to be freed. Assuming that the |
| 11446 |
functions you call are well-behaved, and unprotect only those objects |
| 11447 |
they protect, you can follow the same rule and have confidence that |
| 11448 |
objects will be freed only at appropriate times. |
| 11449 |
|
| 11450 |
|
| 11451 |
Changes in Guile 1.2 (released Tuesday, June 24 1997): |
| 11452 |
|
| 11453 |
* Changes to the distribution |
| 11454 |
|
| 11455 |
** Nightly snapshots are now available from ftp.red-bean.com. |
| 11456 |
The old server, ftp.cyclic.com, has been relinquished to its rightful |
| 11457 |
owner. |
| 11458 |
|
| 11459 |
Nightly snapshots of the Guile development sources are now available via |
| 11460 |
anonymous FTP from ftp.red-bean.com, as /pub/guile/guile-snap.tar.gz. |
| 11461 |
|
| 11462 |
Via the web, that's: ftp://ftp.red-bean.com/pub/guile/guile-snap.tar.gz |
| 11463 |
For getit, that's: ftp.red-bean.com:/pub/guile/guile-snap.tar.gz |
| 11464 |
|
| 11465 |
** To run Guile without installing it, the procedure has changed a bit. |
| 11466 |
|
| 11467 |
If you used a separate build directory to compile Guile, you'll need |
| 11468 |
to include the build directory in SCHEME_LOAD_PATH, as well as the |
| 11469 |
source directory. See the `INSTALL' file for examples. |
| 11470 |
|
| 11471 |
* Changes to the procedure for linking libguile with your programs |
| 11472 |
|
| 11473 |
** The standard Guile load path for Scheme code now includes |
| 11474 |
$(datadir)/guile (usually /usr/local/share/guile). This means that |
| 11475 |
you can install your own Scheme files there, and Guile will find them. |
| 11476 |
(Previous versions of Guile only checked a directory whose name |
| 11477 |
contained the Guile version number, so you had to re-install or move |
| 11478 |
your Scheme sources each time you installed a fresh version of Guile.) |
| 11479 |
|
| 11480 |
The load path also includes $(datadir)/guile/site; we recommend |
| 11481 |
putting individual Scheme files there. If you want to install a |
| 11482 |
package with multiple source files, create a directory for them under |
| 11483 |
$(datadir)/guile. |
| 11484 |
|
| 11485 |
** Guile 1.2 will now use the Rx regular expression library, if it is |
| 11486 |
installed on your system. When you are linking libguile into your own |
| 11487 |
programs, this means you will have to link against -lguile, -lqt (if |
| 11488 |
you configured Guile with thread support), and -lrx. |
| 11489 |
|
| 11490 |
If you are using autoconf to generate configuration scripts for your |
| 11491 |
application, the following lines should suffice to add the appropriate |
| 11492 |
libraries to your link command: |
| 11493 |
|
| 11494 |
### Find Rx, quickthreads and libguile. |
| 11495 |
AC_CHECK_LIB(rx, main) |
| 11496 |
AC_CHECK_LIB(qt, main) |
| 11497 |
AC_CHECK_LIB(guile, scm_shell) |
| 11498 |
|
| 11499 |
The Guile 1.2 distribution does not contain sources for the Rx |
| 11500 |
library, as Guile 1.0 did. If you want to use Rx, you'll need to |
| 11501 |
retrieve it from a GNU FTP site and install it separately. |
| 11502 |
|
| 11503 |
* Changes to Scheme functions and syntax |
| 11504 |
|
| 11505 |
** The dynamic linking features of Guile are now enabled by default. |
| 11506 |
You can disable them by giving the `--disable-dynamic-linking' option |
| 11507 |
to configure. |
| 11508 |
|
| 11509 |
(dynamic-link FILENAME) |
| 11510 |
|
| 11511 |
Find the object file denoted by FILENAME (a string) and link it |
| 11512 |
into the running Guile application. When everything works out, |
| 11513 |
return a Scheme object suitable for representing the linked object |
| 11514 |
file. Otherwise an error is thrown. How object files are |
| 11515 |
searched is system dependent. |
| 11516 |
|
| 11517 |
(dynamic-object? VAL) |
| 11518 |
|
| 11519 |
Determine whether VAL represents a dynamically linked object file. |
| 11520 |
|
| 11521 |
(dynamic-unlink DYNOBJ) |
| 11522 |
|
| 11523 |
Unlink the indicated object file from the application. DYNOBJ |
| 11524 |
should be one of the values returned by `dynamic-link'. |
| 11525 |
|
| 11526 |
(dynamic-func FUNCTION DYNOBJ) |
| 11527 |
|
| 11528 |
Search the C function indicated by FUNCTION (a string or symbol) |
| 11529 |
in DYNOBJ and return some Scheme object that can later be used |
| 11530 |
with `dynamic-call' to actually call this function. Right now, |
| 11531 |
these Scheme objects are formed by casting the address of the |
| 11532 |
function to `long' and converting this number to its Scheme |
| 11533 |
representation. |
| 11534 |
|
| 11535 |
(dynamic-call FUNCTION DYNOBJ) |
| 11536 |
|
| 11537 |
Call the C function indicated by FUNCTION and DYNOBJ. The |
| 11538 |
function is passed no arguments and its return value is ignored. |
| 11539 |
When FUNCTION is something returned by `dynamic-func', call that |
| 11540 |
function and ignore DYNOBJ. When FUNCTION is a string (or symbol, |
| 11541 |
etc.), look it up in DYNOBJ; this is equivalent to |
| 11542 |
|
| 11543 |
(dynamic-call (dynamic-func FUNCTION DYNOBJ) #f) |
| 11544 |
|
| 11545 |
Interrupts are deferred while the C function is executing (with |
| 11546 |
SCM_DEFER_INTS/SCM_ALLOW_INTS). |
| 11547 |
|
| 11548 |
(dynamic-args-call FUNCTION DYNOBJ ARGS) |
| 11549 |
|
| 11550 |
Call the C function indicated by FUNCTION and DYNOBJ, but pass it |
| 11551 |
some arguments and return its return value. The C function is |
| 11552 |
expected to take two arguments and return an `int', just like |
| 11553 |
`main': |
| 11554 |
|
| 11555 |
int c_func (int argc, char **argv); |
| 11556 |
|
| 11557 |
ARGS must be a list of strings and is converted into an array of |
| 11558 |
`char *'. The array is passed in ARGV and its size in ARGC. The |
| 11559 |
return value is converted to a Scheme number and returned from the |
| 11560 |
call to `dynamic-args-call'. |
| 11561 |
|
| 11562 |
When dynamic linking is disabled or not supported on your system, |
| 11563 |
the above functions throw errors, but they are still available. |
| 11564 |
|
| 11565 |
Here is a small example that works on GNU/Linux: |
| 11566 |
|
| 11567 |
(define libc-obj (dynamic-link "libc.so")) |
| 11568 |
(dynamic-args-call 'rand libc-obj '()) |
| 11569 |
|
| 11570 |
See the file `libguile/DYNAMIC-LINKING' for additional comments. |
| 11571 |
|
| 11572 |
** The #/ syntax for module names is depreciated, and will be removed |
| 11573 |
in a future version of Guile. Instead of |
| 11574 |
|
| 11575 |
#/foo/bar/baz |
| 11576 |
|
| 11577 |
instead write |
| 11578 |
|
| 11579 |
(foo bar baz) |
| 11580 |
|
| 11581 |
The latter syntax is more consistent with existing Lisp practice. |
| 11582 |
|
| 11583 |
** Guile now does fancier printing of structures. Structures are the |
| 11584 |
underlying implementation for records, which in turn are used to |
| 11585 |
implement modules, so all of these object now print differently and in |
| 11586 |
a more informative way. |
| 11587 |
|
| 11588 |
The Scheme printer will examine the builtin variable *struct-printer* |
| 11589 |
whenever it needs to print a structure object. When this variable is |
| 11590 |
not `#f' it is deemed to be a procedure and will be applied to the |
| 11591 |
structure object and the output port. When *struct-printer* is `#f' |
| 11592 |
or the procedure return `#f' the structure object will be printed in |
| 11593 |
the boring #<struct 80458270> form. |
| 11594 |
|
| 11595 |
This hook is used by some routines in ice-9/boot-9.scm to implement |
| 11596 |
type specific printing routines. Please read the comments there about |
| 11597 |
"printing structs". |
| 11598 |
|
| 11599 |
One of the more specific uses of structs are records. The printing |
| 11600 |
procedure that could be passed to MAKE-RECORD-TYPE is now actually |
| 11601 |
called. It should behave like a *struct-printer* procedure (described |
| 11602 |
above). |
| 11603 |
|
| 11604 |
** Guile now supports a new R4RS-compliant syntax for keywords. A |
| 11605 |
token of the form #:NAME, where NAME has the same syntax as a Scheme |
| 11606 |
symbol, is the external representation of the keyword named NAME. |
| 11607 |
Keyword objects print using this syntax as well, so values containing |
| 11608 |
keyword objects can be read back into Guile. When used in an |
| 11609 |
expression, keywords are self-quoting objects. |
| 11610 |
|
| 11611 |
Guile suports this read syntax, and uses this print syntax, regardless |
| 11612 |
of the current setting of the `keyword' read option. The `keyword' |
| 11613 |
read option only controls whether Guile recognizes the `:NAME' syntax, |
| 11614 |
which is incompatible with R4RS. (R4RS says such token represent |
| 11615 |
symbols.) |
| 11616 |
|
| 11617 |
** Guile has regular expression support again. Guile 1.0 included |
| 11618 |
functions for matching regular expressions, based on the Rx library. |
| 11619 |
In Guile 1.1, the Guile/Rx interface was removed to simplify the |
| 11620 |
distribution, and thus Guile had no regular expression support. Guile |
| 11621 |
1.2 again supports the most commonly used functions, and supports all |
| 11622 |
of SCSH's regular expression functions. |
| 11623 |
|
| 11624 |
If your system does not include a POSIX regular expression library, |
| 11625 |
and you have not linked Guile with a third-party regexp library such as |
| 11626 |
Rx, these functions will not be available. You can tell whether your |
| 11627 |
Guile installation includes regular expression support by checking |
| 11628 |
whether the `*features*' list includes the `regex' symbol. |
| 11629 |
|
| 11630 |
*** regexp functions |
| 11631 |
|
| 11632 |
By default, Guile supports POSIX extended regular expressions. That |
| 11633 |
means that the characters `(', `)', `+' and `?' are special, and must |
| 11634 |
be escaped if you wish to match the literal characters. |
| 11635 |
|
| 11636 |
This regular expression interface was modeled after that implemented |
| 11637 |
by SCSH, the Scheme Shell. It is intended to be upwardly compatible |
| 11638 |
with SCSH regular expressions. |
| 11639 |
|
| 11640 |
**** Function: string-match PATTERN STR [START] |
| 11641 |
Compile the string PATTERN into a regular expression and compare |
| 11642 |
it with STR. The optional numeric argument START specifies the |
| 11643 |
position of STR at which to begin matching. |
| 11644 |
|
| 11645 |
`string-match' returns a "match structure" which describes what, |
| 11646 |
if anything, was matched by the regular expression. *Note Match |
| 11647 |
Structures::. If STR does not match PATTERN at all, |
| 11648 |
`string-match' returns `#f'. |
| 11649 |
|
| 11650 |
Each time `string-match' is called, it must compile its PATTERN |
| 11651 |
argument into a regular expression structure. This operation is |
| 11652 |
expensive, which makes `string-match' inefficient if the same regular |
| 11653 |
expression is used several times (for example, in a loop). For better |
| 11654 |
performance, you can compile a regular expression in advance and then |
| 11655 |
match strings against the compiled regexp. |
| 11656 |
|
| 11657 |
**** Function: make-regexp STR [FLAGS] |
| 11658 |
Compile the regular expression described by STR, and return the |
| 11659 |
compiled regexp structure. If STR does not describe a legal |
| 11660 |
regular expression, `make-regexp' throws a |
| 11661 |
`regular-expression-syntax' error. |
| 11662 |
|
| 11663 |
FLAGS may be the bitwise-or of one or more of the following: |
| 11664 |
|
| 11665 |
**** Constant: regexp/extended |
| 11666 |
Use POSIX Extended Regular Expression syntax when interpreting |
| 11667 |
STR. If not set, POSIX Basic Regular Expression syntax is used. |
| 11668 |
If the FLAGS argument is omitted, we assume regexp/extended. |
| 11669 |
|
| 11670 |
**** Constant: regexp/icase |
| 11671 |
Do not differentiate case. Subsequent searches using the |
| 11672 |
returned regular expression will be case insensitive. |
| 11673 |
|
| 11674 |
**** Constant: regexp/newline |
| 11675 |
Match-any-character operators don't match a newline. |
| 11676 |
|
| 11677 |
A non-matching list ([^...]) not containing a newline matches a |
| 11678 |
newline. |
| 11679 |
|
| 11680 |
Match-beginning-of-line operator (^) matches the empty string |
| 11681 |
immediately after a newline, regardless of whether the FLAGS |
| 11682 |
passed to regexp-exec contain regexp/notbol. |
| 11683 |
|
| 11684 |
Match-end-of-line operator ($) matches the empty string |
| 11685 |
immediately before a newline, regardless of whether the FLAGS |
| 11686 |
passed to regexp-exec contain regexp/noteol. |
| 11687 |
|
| 11688 |
**** Function: regexp-exec REGEXP STR [START [FLAGS]] |
| 11689 |
Match the compiled regular expression REGEXP against `str'. If |
| 11690 |
the optional integer START argument is provided, begin matching |
| 11691 |
from that position in the string. Return a match structure |
| 11692 |
describing the results of the match, or `#f' if no match could be |
| 11693 |
found. |
| 11694 |
|
| 11695 |
FLAGS may be the bitwise-or of one or more of the following: |
| 11696 |
|
| 11697 |
**** Constant: regexp/notbol |
| 11698 |
The match-beginning-of-line operator always fails to match (but |
| 11699 |
see the compilation flag regexp/newline above) This flag may be |
| 11700 |
used when different portions of a string are passed to |
| 11701 |
regexp-exec and the beginning of the string should not be |
| 11702 |
interpreted as the beginning of the line. |
| 11703 |
|
| 11704 |
**** Constant: regexp/noteol |
| 11705 |
The match-end-of-line operator always fails to match (but see the |
| 11706 |
compilation flag regexp/newline above) |
| 11707 |
|
| 11708 |
**** Function: regexp? OBJ |
| 11709 |
Return `#t' if OBJ is a compiled regular expression, or `#f' |
| 11710 |
otherwise. |
| 11711 |
|
| 11712 |
Regular expressions are commonly used to find patterns in one string |
| 11713 |
and replace them with the contents of another string. |
| 11714 |
|
| 11715 |
**** Function: regexp-substitute PORT MATCH [ITEM...] |
| 11716 |
Write to the output port PORT selected contents of the match |
| 11717 |
structure MATCH. Each ITEM specifies what should be written, and |
| 11718 |
may be one of the following arguments: |
| 11719 |
|
| 11720 |
* A string. String arguments are written out verbatim. |
| 11721 |
|
| 11722 |
* An integer. The submatch with that number is written. |
| 11723 |
|
| 11724 |
* The symbol `pre'. The portion of the matched string preceding |
| 11725 |
the regexp match is written. |
| 11726 |
|
| 11727 |
* The symbol `post'. The portion of the matched string |
| 11728 |
following the regexp match is written. |
| 11729 |
|
| 11730 |
PORT may be `#f', in which case nothing is written; instead, |
| 11731 |
`regexp-substitute' constructs a string from the specified ITEMs |
| 11732 |
and returns that. |
| 11733 |
|
| 11734 |
**** Function: regexp-substitute/global PORT REGEXP TARGET [ITEM...] |
| 11735 |
Similar to `regexp-substitute', but can be used to perform global |
| 11736 |
substitutions on STR. Instead of taking a match structure as an |
| 11737 |
argument, `regexp-substitute/global' takes two string arguments: a |
| 11738 |
REGEXP string describing a regular expression, and a TARGET string |
| 11739 |
which should be matched against this regular expression. |
| 11740 |
|
| 11741 |
Each ITEM behaves as in REGEXP-SUBSTITUTE, with the following |
| 11742 |
exceptions: |
| 11743 |
|
| 11744 |
* A function may be supplied. When this function is called, it |
| 11745 |
will be passed one argument: a match structure for a given |
| 11746 |
regular expression match. It should return a string to be |
| 11747 |
written out to PORT. |
| 11748 |
|
| 11749 |
* The `post' symbol causes `regexp-substitute/global' to recurse |
| 11750 |
on the unmatched portion of STR. This *must* be supplied in |
| 11751 |
order to perform global search-and-replace on STR; if it is |
| 11752 |
not present among the ITEMs, then `regexp-substitute/global' |
| 11753 |
will return after processing a single match. |
| 11754 |
|
| 11755 |
*** Match Structures |
| 11756 |
|
| 11757 |
A "match structure" is the object returned by `string-match' and |
| 11758 |
`regexp-exec'. It describes which portion of a string, if any, matched |
| 11759 |
the given regular expression. Match structures include: a reference to |
| 11760 |
the string that was checked for matches; the starting and ending |
| 11761 |
positions of the regexp match; and, if the regexp included any |
| 11762 |
parenthesized subexpressions, the starting and ending positions of each |
| 11763 |
submatch. |
| 11764 |
|
| 11765 |
In each of the regexp match functions described below, the `match' |
| 11766 |
argument must be a match structure returned by a previous call to |
| 11767 |
`string-match' or `regexp-exec'. Most of these functions return some |
| 11768 |
information about the original target string that was matched against a |
| 11769 |
regular expression; we will call that string TARGET for easy reference. |
| 11770 |
|
| 11771 |
**** Function: regexp-match? OBJ |
| 11772 |
Return `#t' if OBJ is a match structure returned by a previous |
| 11773 |
call to `regexp-exec', or `#f' otherwise. |
| 11774 |
|
| 11775 |
**** Function: match:substring MATCH [N] |
| 11776 |
Return the portion of TARGET matched by subexpression number N. |
| 11777 |
Submatch 0 (the default) represents the entire regexp match. If |
| 11778 |
the regular expression as a whole matched, but the subexpression |
| 11779 |
number N did not match, return `#f'. |
| 11780 |
|
| 11781 |
**** Function: match:start MATCH [N] |
| 11782 |
Return the starting position of submatch number N. |
| 11783 |
|
| 11784 |
**** Function: match:end MATCH [N] |
| 11785 |
Return the ending position of submatch number N. |
| 11786 |
|
| 11787 |
**** Function: match:prefix MATCH |
| 11788 |
Return the unmatched portion of TARGET preceding the regexp match. |
| 11789 |
|
| 11790 |
**** Function: match:suffix MATCH |
| 11791 |
Return the unmatched portion of TARGET following the regexp match. |
| 11792 |
|
| 11793 |
**** Function: match:count MATCH |
| 11794 |
Return the number of parenthesized subexpressions from MATCH. |
| 11795 |
Note that the entire regular expression match itself counts as a |
| 11796 |
subexpression, and failed submatches are included in the count. |
| 11797 |
|
| 11798 |
**** Function: match:string MATCH |
| 11799 |
Return the original TARGET string. |
| 11800 |
|
| 11801 |
*** Backslash Escapes |
| 11802 |
|
| 11803 |
Sometimes you will want a regexp to match characters like `*' or `$' |
| 11804 |
exactly. For example, to check whether a particular string represents |
| 11805 |
a menu entry from an Info node, it would be useful to match it against |
| 11806 |
a regexp like `^* [^:]*::'. However, this won't work; because the |
| 11807 |
asterisk is a metacharacter, it won't match the `*' at the beginning of |
| 11808 |
the string. In this case, we want to make the first asterisk un-magic. |
| 11809 |
|
| 11810 |
You can do this by preceding the metacharacter with a backslash |
| 11811 |
character `\'. (This is also called "quoting" the metacharacter, and |
| 11812 |
is known as a "backslash escape".) When Guile sees a backslash in a |
| 11813 |
regular expression, it considers the following glyph to be an ordinary |
| 11814 |
character, no matter what special meaning it would ordinarily have. |
| 11815 |
Therefore, we can make the above example work by changing the regexp to |
| 11816 |
`^\* [^:]*::'. The `\*' sequence tells the regular expression engine |
| 11817 |
to match only a single asterisk in the target string. |
| 11818 |
|
| 11819 |
Since the backslash is itself a metacharacter, you may force a |
| 11820 |
regexp to match a backslash in the target string by preceding the |
| 11821 |
backslash with itself. For example, to find variable references in a |
| 11822 |
TeX program, you might want to find occurrences of the string `\let\' |
| 11823 |
followed by any number of alphabetic characters. The regular expression |
| 11824 |
`\\let\\[A-Za-z]*' would do this: the double backslashes in the regexp |
| 11825 |
each match a single backslash in the target string. |
| 11826 |
|
| 11827 |
**** Function: regexp-quote STR |
| 11828 |
Quote each special character found in STR with a backslash, and |
| 11829 |
return the resulting string. |
| 11830 |
|
| 11831 |
*Very important:* Using backslash escapes in Guile source code (as |
| 11832 |
in Emacs Lisp or C) can be tricky, because the backslash character has |
| 11833 |
special meaning for the Guile reader. For example, if Guile encounters |
| 11834 |
the character sequence `\n' in the middle of a string while processing |
| 11835 |
Scheme code, it replaces those characters with a newline character. |
| 11836 |
Similarly, the character sequence `\t' is replaced by a horizontal tab. |
| 11837 |
Several of these "escape sequences" are processed by the Guile reader |
| 11838 |
before your code is executed. Unrecognized escape sequences are |
| 11839 |
ignored: if the characters `\*' appear in a string, they will be |
| 11840 |
translated to the single character `*'. |
| 11841 |
|
| 11842 |
This translation is obviously undesirable for regular expressions, |
| 11843 |
since we want to be able to include backslashes in a string in order to |
| 11844 |
escape regexp metacharacters. Therefore, to make sure that a backslash |
| 11845 |
is preserved in a string in your Guile program, you must use *two* |
| 11846 |
consecutive backslashes: |
| 11847 |
|
| 11848 |
(define Info-menu-entry-pattern (make-regexp "^\\* [^:]*")) |
| 11849 |
|
| 11850 |
The string in this example is preprocessed by the Guile reader before |
| 11851 |
any code is executed. The resulting argument to `make-regexp' is the |
| 11852 |
string `^\* [^:]*', which is what we really want. |
| 11853 |
|
| 11854 |
This also means that in order to write a regular expression that |
| 11855 |
matches a single backslash character, the regular expression string in |
| 11856 |
the source code must include *four* backslashes. Each consecutive pair |
| 11857 |
of backslashes gets translated by the Guile reader to a single |
| 11858 |
backslash, and the resulting double-backslash is interpreted by the |
| 11859 |
regexp engine as matching a single backslash character. Hence: |
| 11860 |
|
| 11861 |
(define tex-variable-pattern (make-regexp "\\\\let\\\\=[A-Za-z]*")) |
| 11862 |
|
| 11863 |
The reason for the unwieldiness of this syntax is historical. Both |
| 11864 |
regular expression pattern matchers and Unix string processing systems |
| 11865 |
have traditionally used backslashes with the special meanings described |
| 11866 |
above. The POSIX regular expression specification and ANSI C standard |
| 11867 |
both require these semantics. Attempting to abandon either convention |
| 11868 |
would cause other kinds of compatibility problems, possibly more severe |
| 11869 |
ones. Therefore, without extending the Scheme reader to support |
| 11870 |
strings with different quoting conventions (an ungainly and confusing |
| 11871 |
extension when implemented in other languages), we must adhere to this |
| 11872 |
cumbersome escape syntax. |
| 11873 |
|
| 11874 |
* Changes to the gh_ interface |
| 11875 |
|
| 11876 |
* Changes to the scm_ interface |
| 11877 |
|
| 11878 |
* Changes to system call interfaces: |
| 11879 |
|
| 11880 |
** The value returned by `raise' is now unspecified. It throws an exception |
| 11881 |
if an error occurs. |
| 11882 |
|
| 11883 |
*** A new procedure `sigaction' can be used to install signal handlers |
| 11884 |
|
| 11885 |
(sigaction signum [action] [flags]) |
| 11886 |
|
| 11887 |
signum is the signal number, which can be specified using the value |
| 11888 |
of SIGINT etc. |
| 11889 |
|
| 11890 |
If action is omitted, sigaction returns a pair: the CAR is the current |
| 11891 |
signal hander, which will be either an integer with the value SIG_DFL |
| 11892 |
(default action) or SIG_IGN (ignore), or the Scheme procedure which |
| 11893 |
handles the signal, or #f if a non-Scheme procedure handles the |
| 11894 |
signal. The CDR contains the current sigaction flags for the handler. |
| 11895 |
|
| 11896 |
If action is provided, it is installed as the new handler for signum. |
| 11897 |
action can be a Scheme procedure taking one argument, or the value of |
| 11898 |
SIG_DFL (default action) or SIG_IGN (ignore), or #f to restore |
| 11899 |
whatever signal handler was installed before sigaction was first used. |
| 11900 |
Flags can optionally be specified for the new handler (SA_RESTART is |
| 11901 |
always used if the system provides it, so need not be specified.) The |
| 11902 |
return value is a pair with information about the old handler as |
| 11903 |
described above. |
| 11904 |
|
| 11905 |
This interface does not provide access to the "signal blocking" |
| 11906 |
facility. Maybe this is not needed, since the thread support may |
| 11907 |
provide solutions to the problem of consistent access to data |
| 11908 |
structures. |
| 11909 |
|
| 11910 |
*** A new procedure `flush-all-ports' is equivalent to running |
| 11911 |
`force-output' on every port open for output. |
| 11912 |
|
| 11913 |
** Guile now provides information on how it was built, via the new |
| 11914 |
global variable, %guile-build-info. This variable records the values |
| 11915 |
of the standard GNU makefile directory variables as an assocation |
| 11916 |
list, mapping variable names (symbols) onto directory paths (strings). |
| 11917 |
For example, to find out where the Guile link libraries were |
| 11918 |
installed, you can say: |
| 11919 |
|
| 11920 |
guile -c "(display (assq-ref %guile-build-info 'libdir)) (newline)" |
| 11921 |
|
| 11922 |
|
| 11923 |
* Changes to the scm_ interface |
| 11924 |
|
| 11925 |
** The new function scm_handle_by_message_noexit is just like the |
| 11926 |
existing scm_handle_by_message function, except that it doesn't call |
| 11927 |
exit to terminate the process. Instead, it prints a message and just |
| 11928 |
returns #f. This might be a more appropriate catch-all handler for |
| 11929 |
new dynamic roots and threads. |
| 11930 |
|
| 11931 |
|
| 11932 |
Changes in Guile 1.1 (released Friday, May 16 1997): |
| 11933 |
|
| 11934 |
* Changes to the distribution. |
| 11935 |
|
| 11936 |
The Guile 1.0 distribution has been split up into several smaller |
| 11937 |
pieces: |
| 11938 |
guile-core --- the Guile interpreter itself. |
| 11939 |
guile-tcltk --- the interface between the Guile interpreter and |
| 11940 |
Tcl/Tk; Tcl is an interpreter for a stringy language, and Tk |
| 11941 |
is a toolkit for building graphical user interfaces. |
| 11942 |
guile-rgx-ctax --- the interface between Guile and the Rx regular |
| 11943 |
expression matcher, and the translator for the Ctax |
| 11944 |
programming language. These are packaged together because the |
| 11945 |
Ctax translator uses Rx to parse Ctax source code. |
| 11946 |
|
| 11947 |
This NEWS file describes the changes made to guile-core since the 1.0 |
| 11948 |
release. |
| 11949 |
|
| 11950 |
We no longer distribute the documentation, since it was either out of |
| 11951 |
date, or incomplete. As soon as we have current documentation, we |
| 11952 |
will distribute it. |
| 11953 |
|
| 11954 |
|
| 11955 |
|
| 11956 |
* Changes to the stand-alone interpreter |
| 11957 |
|
| 11958 |
** guile now accepts command-line arguments compatible with SCSH, Olin |
| 11959 |
Shivers' Scheme Shell. |
| 11960 |
|
| 11961 |
In general, arguments are evaluated from left to right, but there are |
| 11962 |
exceptions. The following switches stop argument processing, and |
| 11963 |
stash all remaining command-line arguments as the value returned by |
| 11964 |
the (command-line) function. |
| 11965 |
-s SCRIPT load Scheme source code from FILE, and exit |
| 11966 |
-c EXPR evalute Scheme expression EXPR, and exit |
| 11967 |
-- stop scanning arguments; run interactively |
| 11968 |
|
| 11969 |
The switches below are processed as they are encountered. |
| 11970 |
-l FILE load Scheme source code from FILE |
| 11971 |
-e FUNCTION after reading script, apply FUNCTION to |
| 11972 |
command line arguments |
| 11973 |
-ds do -s script at this point |
| 11974 |
--emacs enable Emacs protocol (experimental) |
| 11975 |
-h, --help display this help and exit |
| 11976 |
-v, --version display version information and exit |
| 11977 |
\ read arguments from following script lines |
| 11978 |
|
| 11979 |
So, for example, here is a Guile script named `ekko' (thanks, Olin) |
| 11980 |
which re-implements the traditional "echo" command: |
| 11981 |
|
| 11982 |
#!/usr/local/bin/guile -s |
| 11983 |
!# |
| 11984 |
(define (main args) |
| 11985 |
(map (lambda (arg) (display arg) (display " ")) |
| 11986 |
(cdr args)) |
| 11987 |
(newline)) |
| 11988 |
|
| 11989 |
(main (command-line)) |
| 11990 |
|
| 11991 |
Suppose we invoke this script as follows: |
| 11992 |
|
| 11993 |
ekko a speckled gecko |
| 11994 |
|
| 11995 |
Through the magic of Unix script processing (triggered by the `#!' |
| 11996 |
token at the top of the file), /usr/local/bin/guile receives the |
| 11997 |
following list of command-line arguments: |
| 11998 |
|
| 11999 |
("-s" "./ekko" "a" "speckled" "gecko") |
| 12000 |
|
| 12001 |
Unix inserts the name of the script after the argument specified on |
| 12002 |
the first line of the file (in this case, "-s"), and then follows that |
| 12003 |
with the arguments given to the script. Guile loads the script, which |
| 12004 |
defines the `main' function, and then applies it to the list of |
| 12005 |
remaining command-line arguments, ("a" "speckled" "gecko"). |
| 12006 |
|
| 12007 |
In Unix, the first line of a script file must take the following form: |
| 12008 |
|
| 12009 |
#!INTERPRETER ARGUMENT |
| 12010 |
|
| 12011 |
where INTERPRETER is the absolute filename of the interpreter |
| 12012 |
executable, and ARGUMENT is a single command-line argument to pass to |
| 12013 |
the interpreter. |
| 12014 |
|
| 12015 |
You may only pass one argument to the interpreter, and its length is |
| 12016 |
limited. These restrictions can be annoying to work around, so Guile |
| 12017 |
provides a general mechanism (borrowed from, and compatible with, |
| 12018 |
SCSH) for circumventing them. |
| 12019 |
|
| 12020 |
If the ARGUMENT in a Guile script is a single backslash character, |
| 12021 |
`\', Guile will open the script file, parse arguments from its second |
| 12022 |
and subsequent lines, and replace the `\' with them. So, for example, |
| 12023 |
here is another implementation of the `ekko' script: |
| 12024 |
|
| 12025 |
#!/usr/local/bin/guile \ |
| 12026 |
-e main -s |
| 12027 |
!# |
| 12028 |
(define (main args) |
| 12029 |
(for-each (lambda (arg) (display arg) (display " ")) |
| 12030 |
(cdr args)) |
| 12031 |
(newline)) |
| 12032 |
|
| 12033 |
If the user invokes this script as follows: |
| 12034 |
|
| 12035 |
ekko a speckled gecko |
| 12036 |
|
| 12037 |
Unix expands this into |
| 12038 |
|
| 12039 |
/usr/local/bin/guile \ ekko a speckled gecko |
| 12040 |
|
| 12041 |
When Guile sees the `\' argument, it replaces it with the arguments |
| 12042 |
read from the second line of the script, producing: |
| 12043 |
|
| 12044 |
/usr/local/bin/guile -e main -s ekko a speckled gecko |
| 12045 |
|
| 12046 |
This tells Guile to load the `ekko' script, and apply the function |
| 12047 |
`main' to the argument list ("a" "speckled" "gecko"). |
| 12048 |
|
| 12049 |
Here is how Guile parses the command-line arguments: |
| 12050 |
- Each space character terminates an argument. This means that two |
| 12051 |
spaces in a row introduce an empty-string argument. |
| 12052 |
- The tab character is not permitted (unless you quote it with the |
| 12053 |
backslash character, as described below), to avoid confusion. |
| 12054 |
- The newline character terminates the sequence of arguments, and will |
| 12055 |
also terminate a final non-empty argument. (However, a newline |
| 12056 |
following a space will not introduce a final empty-string argument; |
| 12057 |
it only terminates the argument list.) |
| 12058 |
- The backslash character is the escape character. It escapes |
| 12059 |
backslash, space, tab, and newline. The ANSI C escape sequences |
| 12060 |
like \n and \t are also supported. These produce argument |
| 12061 |
constituents; the two-character combination \n doesn't act like a |
| 12062 |
terminating newline. The escape sequence \NNN for exactly three |
| 12063 |
octal digits reads as the character whose ASCII code is NNN. As |
| 12064 |
above, characters produced this way are argument constituents. |
| 12065 |
Backslash followed by other characters is not allowed. |
| 12066 |
|
| 12067 |
* Changes to the procedure for linking libguile with your programs |
| 12068 |
|
| 12069 |
** Guile now builds and installs a shared guile library, if your |
| 12070 |
system support shared libraries. (It still builds a static library on |
| 12071 |
all systems.) Guile automatically detects whether your system |
| 12072 |
supports shared libraries. To prevent Guile from buildisg shared |
| 12073 |
libraries, pass the `--disable-shared' flag to the configure script. |
| 12074 |
|
| 12075 |
Guile takes longer to compile when it builds shared libraries, because |
| 12076 |
it must compile every file twice --- once to produce position- |
| 12077 |
independent object code, and once to produce normal object code. |
| 12078 |
|
| 12079 |
** The libthreads library has been merged into libguile. |
| 12080 |
|
| 12081 |
To link a program against Guile, you now need only link against |
| 12082 |
-lguile and -lqt; -lthreads is no longer needed. If you are using |
| 12083 |
autoconf to generate configuration scripts for your application, the |
| 12084 |
following lines should suffice to add the appropriate libraries to |
| 12085 |
your link command: |
| 12086 |
|
| 12087 |
### Find quickthreads and libguile. |
| 12088 |
AC_CHECK_LIB(qt, main) |
| 12089 |
AC_CHECK_LIB(guile, scm_shell) |
| 12090 |
|
| 12091 |
* Changes to Scheme functions |
| 12092 |
|
| 12093 |
** Guile Scheme's special syntax for keyword objects is now optional, |
| 12094 |
and disabled by default. |
| 12095 |
|
| 12096 |
The syntax variation from R4RS made it difficult to port some |
| 12097 |
interesting packages to Guile. The routines which accepted keyword |
| 12098 |
arguments (mostly in the module system) have been modified to also |
| 12099 |
accept symbols whose names begin with `:'. |
| 12100 |
|
| 12101 |
To change the keyword syntax, you must first import the (ice-9 debug) |
| 12102 |
module: |
| 12103 |
(use-modules (ice-9 debug)) |
| 12104 |
|
| 12105 |
Then you can enable the keyword syntax as follows: |
| 12106 |
(read-set! keywords 'prefix) |
| 12107 |
|
| 12108 |
To disable keyword syntax, do this: |
| 12109 |
(read-set! keywords #f) |
| 12110 |
|
| 12111 |
** Many more primitive functions accept shared substrings as |
| 12112 |
arguments. In the past, these functions required normal, mutable |
| 12113 |
strings as arguments, although they never made use of this |
| 12114 |
restriction. |
| 12115 |
|
| 12116 |
** The uniform array functions now operate on byte vectors. These |
| 12117 |
functions are `array-fill!', `serial-array-copy!', `array-copy!', |
| 12118 |
`serial-array-map', `array-map', `array-for-each', and |
| 12119 |
`array-index-map!'. |
| 12120 |
|
| 12121 |
** The new functions `trace' and `untrace' implement simple debugging |
| 12122 |
support for Scheme functions. |
| 12123 |
|
| 12124 |
The `trace' function accepts any number of procedures as arguments, |
| 12125 |
and tells the Guile interpreter to display each procedure's name and |
| 12126 |
arguments each time the procedure is invoked. When invoked with no |
| 12127 |
arguments, `trace' returns the list of procedures currently being |
| 12128 |
traced. |
| 12129 |
|
| 12130 |
The `untrace' function accepts any number of procedures as arguments, |
| 12131 |
and tells the Guile interpreter not to trace them any more. When |
| 12132 |
invoked with no arguments, `untrace' untraces all curretly traced |
| 12133 |
procedures. |
| 12134 |
|
| 12135 |
The tracing in Guile has an advantage over most other systems: we |
| 12136 |
don't create new procedure objects, but mark the procedure objects |
| 12137 |
themselves. This means that anonymous and internal procedures can be |
| 12138 |
traced. |
| 12139 |
|
| 12140 |
** The function `assert-repl-prompt' has been renamed to |
| 12141 |
`set-repl-prompt!'. It takes one argument, PROMPT. |
| 12142 |
- If PROMPT is #f, the Guile read-eval-print loop will not prompt. |
| 12143 |
- If PROMPT is a string, we use it as a prompt. |
| 12144 |
- If PROMPT is a procedure accepting no arguments, we call it, and |
| 12145 |
display the result as a prompt. |
| 12146 |
- Otherwise, we display "> ". |
| 12147 |
|
| 12148 |
** The new function `eval-string' reads Scheme expressions from a |
| 12149 |
string and evaluates them, returning the value of the last expression |
| 12150 |
in the string. If the string contains no expressions, it returns an |
| 12151 |
unspecified value. |
| 12152 |
|
| 12153 |
** The new function `thunk?' returns true iff its argument is a |
| 12154 |
procedure of zero arguments. |
| 12155 |
|
| 12156 |
** `defined?' is now a builtin function, instead of syntax. This |
| 12157 |
means that its argument should be quoted. It returns #t iff its |
| 12158 |
argument is bound in the current module. |
| 12159 |
|
| 12160 |
** The new syntax `use-modules' allows you to add new modules to your |
| 12161 |
environment without re-typing a complete `define-module' form. It |
| 12162 |
accepts any number of module names as arguments, and imports their |
| 12163 |
public bindings into the current module. |
| 12164 |
|
| 12165 |
** The new function (module-defined? NAME MODULE) returns true iff |
| 12166 |
NAME, a symbol, is defined in MODULE, a module object. |
| 12167 |
|
| 12168 |
** The new function `builtin-bindings' creates and returns a hash |
| 12169 |
table containing copies of all the root module's bindings. |
| 12170 |
|
| 12171 |
** The new function `builtin-weak-bindings' does the same as |
| 12172 |
`builtin-bindings', but creates a doubly-weak hash table. |
| 12173 |
|
| 12174 |
** The `equal?' function now considers variable objects to be |
| 12175 |
equivalent if they have the same name and the same value. |
| 12176 |
|
| 12177 |
** The new function `command-line' returns the command-line arguments |
| 12178 |
given to Guile, as a list of strings. |
| 12179 |
|
| 12180 |
When using guile as a script interpreter, `command-line' returns the |
| 12181 |
script's arguments; those processed by the interpreter (like `-s' or |
| 12182 |
`-c') are omitted. (In other words, you get the normal, expected |
| 12183 |
behavior.) Any application that uses scm_shell to process its |
| 12184 |
command-line arguments gets this behavior as well. |
| 12185 |
|
| 12186 |
** The new function `load-user-init' looks for a file called `.guile' |
| 12187 |
in the user's home directory, and loads it if it exists. This is |
| 12188 |
mostly for use by the code generated by scm_compile_shell_switches, |
| 12189 |
but we thought it might also be useful in other circumstances. |
| 12190 |
|
| 12191 |
** The new function `log10' returns the base-10 logarithm of its |
| 12192 |
argument. |
| 12193 |
|
| 12194 |
** Changes to I/O functions |
| 12195 |
|
| 12196 |
*** The functions `read', `primitive-load', `read-and-eval!', and |
| 12197 |
`primitive-load-path' no longer take optional arguments controlling |
| 12198 |
case insensitivity and a `#' parser. |
| 12199 |
|
| 12200 |
Case sensitivity is now controlled by a read option called |
| 12201 |
`case-insensitive'. The user can add new `#' syntaxes with the |
| 12202 |
`read-hash-extend' function (see below). |
| 12203 |
|
| 12204 |
*** The new function `read-hash-extend' allows the user to change the |
| 12205 |
syntax of Guile Scheme in a somewhat controlled way. |
| 12206 |
|
| 12207 |
(read-hash-extend CHAR PROC) |
| 12208 |
When parsing S-expressions, if we read a `#' character followed by |
| 12209 |
the character CHAR, use PROC to parse an object from the stream. |
| 12210 |
If PROC is #f, remove any parsing procedure registered for CHAR. |
| 12211 |
|
| 12212 |
The reader applies PROC to two arguments: CHAR and an input port. |
| 12213 |
|
| 12214 |
*** The new functions read-delimited and read-delimited! provide a |
| 12215 |
general mechanism for doing delimited input on streams. |
| 12216 |
|
| 12217 |
(read-delimited DELIMS [PORT HANDLE-DELIM]) |
| 12218 |
Read until we encounter one of the characters in DELIMS (a string), |
| 12219 |
or end-of-file. PORT is the input port to read from; it defaults to |
| 12220 |
the current input port. The HANDLE-DELIM parameter determines how |
| 12221 |
the terminating character is handled; it should be one of the |
| 12222 |
following symbols: |
| 12223 |
|
| 12224 |
'trim omit delimiter from result |
| 12225 |
'peek leave delimiter character in input stream |
| 12226 |
'concat append delimiter character to returned value |
| 12227 |
'split return a pair: (RESULT . TERMINATOR) |
| 12228 |
|
| 12229 |
HANDLE-DELIM defaults to 'peek. |
| 12230 |
|
| 12231 |
(read-delimited! DELIMS BUF [PORT HANDLE-DELIM START END]) |
| 12232 |
A side-effecting variant of `read-delimited'. |
| 12233 |
|
| 12234 |
The data is written into the string BUF at the indices in the |
| 12235 |
half-open interval [START, END); the default interval is the whole |
| 12236 |
string: START = 0 and END = (string-length BUF). The values of |
| 12237 |
START and END must specify a well-defined interval in BUF, i.e. |
| 12238 |
0 <= START <= END <= (string-length BUF). |
| 12239 |
|
| 12240 |
It returns NBYTES, the number of bytes read. If the buffer filled |
| 12241 |
up without a delimiter character being found, it returns #f. If the |
| 12242 |
port is at EOF when the read starts, it returns the EOF object. |
| 12243 |
|
| 12244 |
If an integer is returned (i.e., the read is successfully terminated |
| 12245 |
by reading a delimiter character), then the HANDLE-DELIM parameter |
| 12246 |
determines how to handle the terminating character. It is described |
| 12247 |
above, and defaults to 'peek. |
| 12248 |
|
| 12249 |
(The descriptions of these functions were borrowed from the SCSH |
| 12250 |
manual, by Olin Shivers and Brian Carlstrom.) |
| 12251 |
|
| 12252 |
*** The `%read-delimited!' function is the primitive used to implement |
| 12253 |
`read-delimited' and `read-delimited!'. |
| 12254 |
|
| 12255 |
(%read-delimited! DELIMS BUF GOBBLE? [PORT START END]) |
| 12256 |
|
| 12257 |
This returns a pair of values: (TERMINATOR . NUM-READ). |
| 12258 |
- TERMINATOR describes why the read was terminated. If it is a |
| 12259 |
character or the eof object, then that is the value that terminated |
| 12260 |
the read. If it is #f, the function filled the buffer without finding |
| 12261 |
a delimiting character. |
| 12262 |
- NUM-READ is the number of characters read into BUF. |
| 12263 |
|
| 12264 |
If the read is successfully terminated by reading a delimiter |
| 12265 |
character, then the gobble? parameter determines what to do with the |
| 12266 |
terminating character. If true, the character is removed from the |
| 12267 |
input stream; if false, the character is left in the input stream |
| 12268 |
where a subsequent read operation will retrieve it. In either case, |
| 12269 |
the character is also the first value returned by the procedure call. |
| 12270 |
|
| 12271 |
(The descriptions of this function was borrowed from the SCSH manual, |
| 12272 |
by Olin Shivers and Brian Carlstrom.) |
| 12273 |
|
| 12274 |
*** The `read-line' and `read-line!' functions have changed; they now |
| 12275 |
trim the terminator by default; previously they appended it to the |
| 12276 |
returned string. For the old behavior, use (read-line PORT 'concat). |
| 12277 |
|
| 12278 |
*** The functions `uniform-array-read!' and `uniform-array-write!' now |
| 12279 |
take new optional START and END arguments, specifying the region of |
| 12280 |
the array to read and write. |
| 12281 |
|
| 12282 |
*** The `ungetc-char-ready?' function has been removed. We feel it's |
| 12283 |
inappropriate for an interface to expose implementation details this |
| 12284 |
way. |
| 12285 |
|
| 12286 |
** Changes to the Unix library and system call interface |
| 12287 |
|
| 12288 |
*** The new fcntl function provides access to the Unix `fcntl' system |
| 12289 |
call. |
| 12290 |
|
| 12291 |
(fcntl PORT COMMAND VALUE) |
| 12292 |
Apply COMMAND to PORT's file descriptor, with VALUE as an argument. |
| 12293 |
Values for COMMAND are: |
| 12294 |
|
| 12295 |
F_DUPFD duplicate a file descriptor |
| 12296 |
F_GETFD read the descriptor's close-on-exec flag |
| 12297 |
F_SETFD set the descriptor's close-on-exec flag to VALUE |
| 12298 |
F_GETFL read the descriptor's flags, as set on open |
| 12299 |
F_SETFL set the descriptor's flags, as set on open to VALUE |
| 12300 |
F_GETOWN return the process ID of a socket's owner, for SIGIO |
| 12301 |
F_SETOWN set the process that owns a socket to VALUE, for SIGIO |
| 12302 |
FD_CLOEXEC not sure what this is |
| 12303 |
|
| 12304 |
For details, see the documentation for the fcntl system call. |
| 12305 |
|
| 12306 |
*** The arguments to `select' have changed, for compatibility with |
| 12307 |
SCSH. The TIMEOUT parameter may now be non-integral, yielding the |
| 12308 |
expected behavior. The MILLISECONDS parameter has been changed to |
| 12309 |
MICROSECONDS, to more closely resemble the underlying system call. |
| 12310 |
The RVEC, WVEC, and EVEC arguments can now be vectors; the type of the |
| 12311 |
corresponding return set will be the same. |
| 12312 |
|
| 12313 |
*** The arguments to the `mknod' system call have changed. They are |
| 12314 |
now: |
| 12315 |
|
| 12316 |
(mknod PATH TYPE PERMS DEV) |
| 12317 |
Create a new file (`node') in the file system. PATH is the name of |
| 12318 |
the file to create. TYPE is the kind of file to create; it should |
| 12319 |
be 'fifo, 'block-special, or 'char-special. PERMS specifies the |
| 12320 |
permission bits to give the newly created file. If TYPE is |
| 12321 |
'block-special or 'char-special, DEV specifies which device the |
| 12322 |
special file refers to; its interpretation depends on the kind of |
| 12323 |
special file being created. |
| 12324 |
|
| 12325 |
*** The `fork' function has been renamed to `primitive-fork', to avoid |
| 12326 |
clashing with various SCSH forks. |
| 12327 |
|
| 12328 |
*** The `recv' and `recvfrom' functions have been renamed to `recv!' |
| 12329 |
and `recvfrom!'. They no longer accept a size for a second argument; |
| 12330 |
you must pass a string to hold the received value. They no longer |
| 12331 |
return the buffer. Instead, `recv' returns the length of the message |
| 12332 |
received, and `recvfrom' returns a pair containing the packet's length |
| 12333 |
and originating address. |
| 12334 |
|
| 12335 |
*** The file descriptor datatype has been removed, as have the |
| 12336 |
`read-fd', `write-fd', `close', `lseek', and `dup' functions. |
| 12337 |
We plan to replace these functions with a SCSH-compatible interface. |
| 12338 |
|
| 12339 |
*** The `create' function has been removed; it's just a special case |
| 12340 |
of `open'. |
| 12341 |
|
| 12342 |
*** There are new functions to break down process termination status |
| 12343 |
values. In the descriptions below, STATUS is a value returned by |
| 12344 |
`waitpid'. |
| 12345 |
|
| 12346 |
(status:exit-val STATUS) |
| 12347 |
If the child process exited normally, this function returns the exit |
| 12348 |
code for the child process (i.e., the value passed to exit, or |
| 12349 |
returned from main). If the child process did not exit normally, |
| 12350 |
this function returns #f. |
| 12351 |
|
| 12352 |
(status:stop-sig STATUS) |
| 12353 |
If the child process was suspended by a signal, this function |
| 12354 |
returns the signal that suspended the child. Otherwise, it returns |
| 12355 |
#f. |
| 12356 |
|
| 12357 |
(status:term-sig STATUS) |
| 12358 |
If the child process terminated abnormally, this function returns |
| 12359 |
the signal that terminated the child. Otherwise, this function |
| 12360 |
returns false. |
| 12361 |
|
| 12362 |
POSIX promises that exactly one of these functions will return true on |
| 12363 |
a valid STATUS value. |
| 12364 |
|
| 12365 |
These functions are compatible with SCSH. |
| 12366 |
|
| 12367 |
*** There are new accessors and setters for the broken-out time vectors |
| 12368 |
returned by `localtime', `gmtime', and that ilk. They are: |
| 12369 |
|
| 12370 |
Component Accessor Setter |
| 12371 |
========================= ============ ============ |
| 12372 |
seconds tm:sec set-tm:sec |
| 12373 |
minutes tm:min set-tm:min |
| 12374 |
hours tm:hour set-tm:hour |
| 12375 |
day of the month tm:mday set-tm:mday |
| 12376 |
month tm:mon set-tm:mon |
| 12377 |
year tm:year set-tm:year |
| 12378 |
day of the week tm:wday set-tm:wday |
| 12379 |
day in the year tm:yday set-tm:yday |
| 12380 |
daylight saving time tm:isdst set-tm:isdst |
| 12381 |
GMT offset, seconds tm:gmtoff set-tm:gmtoff |
| 12382 |
name of time zone tm:zone set-tm:zone |
| 12383 |
|
| 12384 |
*** There are new accessors for the vectors returned by `uname', |
| 12385 |
describing the host system: |
| 12386 |
|
| 12387 |
Component Accessor |
| 12388 |
============================================== ================ |
| 12389 |
name of the operating system implementation utsname:sysname |
| 12390 |
network name of this machine utsname:nodename |
| 12391 |
release level of the operating system utsname:release |
| 12392 |
version level of the operating system utsname:version |
| 12393 |
machine hardware platform utsname:machine |
| 12394 |
|
| 12395 |
*** There are new accessors for the vectors returned by `getpw', |
| 12396 |
`getpwnam', `getpwuid', and `getpwent', describing entries from the |
| 12397 |
system's user database: |
| 12398 |
|
| 12399 |
Component Accessor |
| 12400 |
====================== ================= |
| 12401 |
user name passwd:name |
| 12402 |
user password passwd:passwd |
| 12403 |
user id passwd:uid |
| 12404 |
group id passwd:gid |
| 12405 |
real name passwd:gecos |
| 12406 |
home directory passwd:dir |
| 12407 |
shell program passwd:shell |
| 12408 |
|
| 12409 |
*** There are new accessors for the vectors returned by `getgr', |
| 12410 |
`getgrnam', `getgrgid', and `getgrent', describing entries from the |
| 12411 |
system's group database: |
| 12412 |
|
| 12413 |
Component Accessor |
| 12414 |
======================= ============ |
| 12415 |
group name group:name |
| 12416 |
group password group:passwd |
| 12417 |
group id group:gid |
| 12418 |
group members group:mem |
| 12419 |
|
| 12420 |
*** There are new accessors for the vectors returned by `gethost', |
| 12421 |
`gethostbyaddr', `gethostbyname', and `gethostent', describing |
| 12422 |
internet hosts: |
| 12423 |
|
| 12424 |
Component Accessor |
| 12425 |
========================= =============== |
| 12426 |
official name of host hostent:name |
| 12427 |
alias list hostent:aliases |
| 12428 |
host address type hostent:addrtype |
| 12429 |
length of address hostent:length |
| 12430 |
list of addresses hostent:addr-list |
| 12431 |
|
| 12432 |
*** There are new accessors for the vectors returned by `getnet', |
| 12433 |
`getnetbyaddr', `getnetbyname', and `getnetent', describing internet |
| 12434 |
networks: |
| 12435 |
|
| 12436 |
Component Accessor |
| 12437 |
========================= =============== |
| 12438 |
official name of net netent:name |
| 12439 |
alias list netent:aliases |
| 12440 |
net number type netent:addrtype |
| 12441 |
net number netent:net |
| 12442 |
|
| 12443 |
*** There are new accessors for the vectors returned by `getproto', |
| 12444 |
`getprotobyname', `getprotobynumber', and `getprotoent', describing |
| 12445 |
internet protocols: |
| 12446 |
|
| 12447 |
Component Accessor |
| 12448 |
========================= =============== |
| 12449 |
official protocol name protoent:name |
| 12450 |
alias list protoent:aliases |
| 12451 |
protocol number protoent:proto |
| 12452 |
|
| 12453 |
*** There are new accessors for the vectors returned by `getserv', |
| 12454 |
`getservbyname', `getservbyport', and `getservent', describing |
| 12455 |
internet protocols: |
| 12456 |
|
| 12457 |
Component Accessor |
| 12458 |
========================= =============== |
| 12459 |
official service name servent:name |
| 12460 |
alias list servent:aliases |
| 12461 |
port number servent:port |
| 12462 |
protocol to use servent:proto |
| 12463 |
|
| 12464 |
*** There are new accessors for the sockaddr structures returned by |
| 12465 |
`accept', `getsockname', `getpeername', `recvfrom!': |
| 12466 |
|
| 12467 |
Component Accessor |
| 12468 |
======================================== =============== |
| 12469 |
address format (`family') sockaddr:fam |
| 12470 |
path, for file domain addresses sockaddr:path |
| 12471 |
address, for internet domain addresses sockaddr:addr |
| 12472 |
TCP or UDP port, for internet sockaddr:port |
| 12473 |
|
| 12474 |
*** The `getpwent', `getgrent', `gethostent', `getnetent', |
| 12475 |
`getprotoent', and `getservent' functions now return #f at the end of |
| 12476 |
the user database. (They used to throw an exception.) |
| 12477 |
|
| 12478 |
Note that calling MUMBLEent function is equivalent to calling the |
| 12479 |
corresponding MUMBLE function with no arguments. |
| 12480 |
|
| 12481 |
*** The `setpwent', `setgrent', `sethostent', `setnetent', |
| 12482 |
`setprotoent', and `setservent' routines now take no arguments. |
| 12483 |
|
| 12484 |
*** The `gethost', `getproto', `getnet', and `getserv' functions now |
| 12485 |
provide more useful information when they throw an exception. |
| 12486 |
|
| 12487 |
*** The `lnaof' function has been renamed to `inet-lnaof'. |
| 12488 |
|
| 12489 |
*** Guile now claims to have the `current-time' feature. |
| 12490 |
|
| 12491 |
*** The `mktime' function now takes an optional second argument ZONE, |
| 12492 |
giving the time zone to use for the conversion. ZONE should be a |
| 12493 |
string, in the same format as expected for the "TZ" environment variable. |
| 12494 |
|
| 12495 |
*** The `strptime' function now returns a pair (TIME . COUNT), where |
| 12496 |
TIME is the parsed time as a vector, and COUNT is the number of |
| 12497 |
characters from the string left unparsed. This function used to |
| 12498 |
return the remaining characters as a string. |
| 12499 |
|
| 12500 |
*** The `gettimeofday' function has replaced the old `time+ticks' function. |
| 12501 |
The return value is now (SECONDS . MICROSECONDS); the fractional |
| 12502 |
component is no longer expressed in "ticks". |
| 12503 |
|
| 12504 |
*** The `ticks/sec' constant has been removed, in light of the above change. |
| 12505 |
|
| 12506 |
* Changes to the gh_ interface |
| 12507 |
|
| 12508 |
** gh_eval_str() now returns an SCM object which is the result of the |
| 12509 |
evaluation |
| 12510 |
|
| 12511 |
** gh_scm2str() now copies the Scheme data to a caller-provided C |
| 12512 |
array |
| 12513 |
|
| 12514 |
** gh_scm2newstr() now makes a C array, copies the Scheme data to it, |
| 12515 |
and returns the array |
| 12516 |
|
| 12517 |
** gh_scm2str0() is gone: there is no need to distinguish |
| 12518 |
null-terminated from non-null-terminated, since gh_scm2newstr() allows |
| 12519 |
the user to interpret the data both ways. |
| 12520 |
|
| 12521 |
* Changes to the scm_ interface |
| 12522 |
|
| 12523 |
** The new function scm_symbol_value0 provides an easy way to get a |
| 12524 |
symbol's value from C code: |
| 12525 |
|
| 12526 |
SCM scm_symbol_value0 (char *NAME) |
| 12527 |
Return the value of the symbol named by the null-terminated string |
| 12528 |
NAME in the current module. If the symbol named NAME is unbound in |
| 12529 |
the current module, return SCM_UNDEFINED. |
| 12530 |
|
| 12531 |
** The new function scm_sysintern0 creates new top-level variables, |
| 12532 |
without assigning them a value. |
| 12533 |
|
| 12534 |
SCM scm_sysintern0 (char *NAME) |
| 12535 |
Create a new Scheme top-level variable named NAME. NAME is a |
| 12536 |
null-terminated string. Return the variable's value cell. |
| 12537 |
|
| 12538 |
** The function scm_internal_catch is the guts of catch. It handles |
| 12539 |
all the mechanics of setting up a catch target, invoking the catch |
| 12540 |
body, and perhaps invoking the handler if the body does a throw. |
| 12541 |
|
| 12542 |
The function is designed to be usable from C code, but is general |
| 12543 |
enough to implement all the semantics Guile Scheme expects from throw. |
| 12544 |
|
| 12545 |
TAG is the catch tag. Typically, this is a symbol, but this function |
| 12546 |
doesn't actually care about that. |
| 12547 |
|
| 12548 |
BODY is a pointer to a C function which runs the body of the catch; |
| 12549 |
this is the code you can throw from. We call it like this: |
| 12550 |
BODY (BODY_DATA, JMPBUF) |
| 12551 |
where: |
| 12552 |
BODY_DATA is just the BODY_DATA argument we received; we pass it |
| 12553 |
through to BODY as its first argument. The caller can make |
| 12554 |
BODY_DATA point to anything useful that BODY might need. |
| 12555 |
JMPBUF is the Scheme jmpbuf object corresponding to this catch, |
| 12556 |
which we have just created and initialized. |
| 12557 |
|
| 12558 |
HANDLER is a pointer to a C function to deal with a throw to TAG, |
| 12559 |
should one occur. We call it like this: |
| 12560 |
HANDLER (HANDLER_DATA, THROWN_TAG, THROW_ARGS) |
| 12561 |
where |
| 12562 |
HANDLER_DATA is the HANDLER_DATA argument we recevied; it's the |
| 12563 |
same idea as BODY_DATA above. |
| 12564 |
THROWN_TAG is the tag that the user threw to; usually this is |
| 12565 |
TAG, but it could be something else if TAG was #t (i.e., a |
| 12566 |
catch-all), or the user threw to a jmpbuf. |
| 12567 |
THROW_ARGS is the list of arguments the user passed to the THROW |
| 12568 |
function. |
| 12569 |
|
| 12570 |
BODY_DATA is just a pointer we pass through to BODY. HANDLER_DATA |
| 12571 |
is just a pointer we pass through to HANDLER. We don't actually |
| 12572 |
use either of those pointers otherwise ourselves. The idea is |
| 12573 |
that, if our caller wants to communicate something to BODY or |
| 12574 |
HANDLER, it can pass a pointer to it as MUMBLE_DATA, which BODY and |
| 12575 |
HANDLER can then use. Think of it as a way to make BODY and |
| 12576 |
HANDLER closures, not just functions; MUMBLE_DATA points to the |
| 12577 |
enclosed variables. |
| 12578 |
|
| 12579 |
Of course, it's up to the caller to make sure that any data a |
| 12580 |
MUMBLE_DATA needs is protected from GC. A common way to do this is |
| 12581 |
to make MUMBLE_DATA a pointer to data stored in an automatic |
| 12582 |
structure variable; since the collector must scan the stack for |
| 12583 |
references anyway, this assures that any references in MUMBLE_DATA |
| 12584 |
will be found. |
| 12585 |
|
| 12586 |
** The new function scm_internal_lazy_catch is exactly like |
| 12587 |
scm_internal_catch, except: |
| 12588 |
|
| 12589 |
- It does not unwind the stack (this is the major difference). |
| 12590 |
- If handler returns, its value is returned from the throw. |
| 12591 |
- BODY always receives #f as its JMPBUF argument (since there's no |
| 12592 |
jmpbuf associated with a lazy catch, because we don't unwind the |
| 12593 |
stack.) |
| 12594 |
|
| 12595 |
** scm_body_thunk is a new body function you can pass to |
| 12596 |
scm_internal_catch if you want the body to be like Scheme's `catch' |
| 12597 |
--- a thunk, or a function of one argument if the tag is #f. |
| 12598 |
|
| 12599 |
BODY_DATA is a pointer to a scm_body_thunk_data structure, which |
| 12600 |
contains the Scheme procedure to invoke as the body, and the tag |
| 12601 |
we're catching. If the tag is #f, then we pass JMPBUF (created by |
| 12602 |
scm_internal_catch) to the body procedure; otherwise, the body gets |
| 12603 |
no arguments. |
| 12604 |
|
| 12605 |
** scm_handle_by_proc is a new handler function you can pass to |
| 12606 |
scm_internal_catch if you want the handler to act like Scheme's catch |
| 12607 |
--- call a procedure with the tag and the throw arguments. |
| 12608 |
|
| 12609 |
If the user does a throw to this catch, this function runs a handler |
| 12610 |
procedure written in Scheme. HANDLER_DATA is a pointer to an SCM |
| 12611 |
variable holding the Scheme procedure object to invoke. It ought to |
| 12612 |
be a pointer to an automatic variable (i.e., one living on the stack), |
| 12613 |
or the procedure object should be otherwise protected from GC. |
| 12614 |
|
| 12615 |
** scm_handle_by_message is a new handler function to use with |
| 12616 |
`scm_internal_catch' if you want Guile to print a message and die. |
| 12617 |
It's useful for dealing with throws to uncaught keys at the top level. |
| 12618 |
|
| 12619 |
HANDLER_DATA, if non-zero, is assumed to be a char * pointing to a |
| 12620 |
message header to print; if zero, we use "guile" instead. That |
| 12621 |
text is followed by a colon, then the message described by ARGS. |
| 12622 |
|
| 12623 |
** The return type of scm_boot_guile is now void; the function does |
| 12624 |
not return a value, and indeed, never returns at all. |
| 12625 |
|
| 12626 |
** The new function scm_shell makes it easy for user applications to |
| 12627 |
process command-line arguments in a way that is compatible with the |
| 12628 |
stand-alone guile interpreter (which is in turn compatible with SCSH, |
| 12629 |
the Scheme shell). |
| 12630 |
|
| 12631 |
To use the scm_shell function, first initialize any guile modules |
| 12632 |
linked into your application, and then call scm_shell with the values |
| 12633 |
of ARGC and ARGV your `main' function received. scm_shell will add |
| 12634 |
any SCSH-style meta-arguments from the top of the script file to the |
| 12635 |
argument vector, and then process the command-line arguments. This |
| 12636 |
generally means loading a script file or starting up an interactive |
| 12637 |
command interpreter. For details, see "Changes to the stand-alone |
| 12638 |
interpreter" above. |
| 12639 |
|
| 12640 |
** The new functions scm_get_meta_args and scm_count_argv help you |
| 12641 |
implement the SCSH-style meta-argument, `\'. |
| 12642 |
|
| 12643 |
char **scm_get_meta_args (int ARGC, char **ARGV) |
| 12644 |
If the second element of ARGV is a string consisting of a single |
| 12645 |
backslash character (i.e. "\\" in Scheme notation), open the file |
| 12646 |
named by the following argument, parse arguments from it, and return |
| 12647 |
the spliced command line. The returned array is terminated by a |
| 12648 |
null pointer. |
| 12649 |
|
| 12650 |
For details of argument parsing, see above, under "guile now accepts |
| 12651 |
command-line arguments compatible with SCSH..." |
| 12652 |
|
| 12653 |
int scm_count_argv (char **ARGV) |
| 12654 |
Count the arguments in ARGV, assuming it is terminated by a null |
| 12655 |
pointer. |
| 12656 |
|
| 12657 |
For an example of how these functions might be used, see the source |
| 12658 |
code for the function scm_shell in libguile/script.c. |
| 12659 |
|
| 12660 |
You will usually want to use scm_shell instead of calling this |
| 12661 |
function yourself. |
| 12662 |
|
| 12663 |
** The new function scm_compile_shell_switches turns an array of |
| 12664 |
command-line arguments into Scheme code to carry out the actions they |
| 12665 |
describe. Given ARGC and ARGV, it returns a Scheme expression to |
| 12666 |
evaluate, and calls scm_set_program_arguments to make any remaining |
| 12667 |
command-line arguments available to the Scheme code. For example, |
| 12668 |
given the following arguments: |
| 12669 |
|
| 12670 |
-e main -s ekko a speckled gecko |
| 12671 |
|
| 12672 |
scm_set_program_arguments will return the following expression: |
| 12673 |
|
| 12674 |
(begin (load "ekko") (main (command-line)) (quit)) |
| 12675 |
|
| 12676 |
You will usually want to use scm_shell instead of calling this |
| 12677 |
function yourself. |
| 12678 |
|
| 12679 |
** The function scm_shell_usage prints a usage message appropriate for |
| 12680 |
an interpreter that uses scm_compile_shell_switches to handle its |
| 12681 |
command-line arguments. |
| 12682 |
|
| 12683 |
void scm_shell_usage (int FATAL, char *MESSAGE) |
| 12684 |
Print a usage message to the standard error output. If MESSAGE is |
| 12685 |
non-zero, write it before the usage message, followed by a newline. |
| 12686 |
If FATAL is non-zero, exit the process, using FATAL as the |
| 12687 |
termination status. (If you want to be compatible with Guile, |
| 12688 |
always use 1 as the exit status when terminating due to command-line |
| 12689 |
usage problems.) |
| 12690 |
|
| 12691 |
You will usually want to use scm_shell instead of calling this |
| 12692 |
function yourself. |
| 12693 |
|
| 12694 |
** scm_eval_0str now returns SCM_UNSPECIFIED if the string contains no |
| 12695 |
expressions. It used to return SCM_EOL. Earth-shattering. |
| 12696 |
|
| 12697 |
** The macros for declaring scheme objects in C code have been |
| 12698 |
rearranged slightly. They are now: |
| 12699 |
|
| 12700 |
SCM_SYMBOL (C_NAME, SCHEME_NAME) |
| 12701 |
Declare a static SCM variable named C_NAME, and initialize it to |
| 12702 |
point to the Scheme symbol whose name is SCHEME_NAME. C_NAME should |
| 12703 |
be a C identifier, and SCHEME_NAME should be a C string. |
| 12704 |
|
| 12705 |
SCM_GLOBAL_SYMBOL (C_NAME, SCHEME_NAME) |
| 12706 |
Just like SCM_SYMBOL, but make C_NAME globally visible. |
| 12707 |
|
| 12708 |
SCM_VCELL (C_NAME, SCHEME_NAME) |
| 12709 |
Create a global variable at the Scheme level named SCHEME_NAME. |
| 12710 |
Declare a static SCM variable named C_NAME, and initialize it to |
| 12711 |
point to the Scheme variable's value cell. |
| 12712 |
|
| 12713 |
SCM_GLOBAL_VCELL (C_NAME, SCHEME_NAME) |
| 12714 |
Just like SCM_VCELL, but make C_NAME globally visible. |
| 12715 |
|
| 12716 |
The `guile-snarf' script writes initialization code for these macros |
| 12717 |
to its standard output, given C source code as input. |
| 12718 |
|
| 12719 |
The SCM_GLOBAL macro is gone. |
| 12720 |
|
| 12721 |
** The scm_read_line and scm_read_line_x functions have been replaced |
| 12722 |
by Scheme code based on the %read-delimited! procedure (known to C |
| 12723 |
code as scm_read_delimited_x). See its description above for more |
| 12724 |
information. |
| 12725 |
|
| 12726 |
** The function scm_sys_open has been renamed to scm_open. It now |
| 12727 |
returns a port instead of an FD object. |
| 12728 |
|
| 12729 |
* The dynamic linking support has changed. For more information, see |
| 12730 |
libguile/DYNAMIC-LINKING. |
| 12731 |
|
| 12732 |
|
| 12733 |
Guile 1.0b3 |
| 12734 |
|
| 12735 |
User-visible changes from Thursday, September 5, 1996 until Guile 1.0 |
| 12736 |
(Sun 5 Jan 1997): |
| 12737 |
|
| 12738 |
* Changes to the 'guile' program: |
| 12739 |
|
| 12740 |
** Guile now loads some new files when it starts up. Guile first |
| 12741 |
searches the load path for init.scm, and loads it if found. Then, if |
| 12742 |
Guile is not being used to execute a script, and the user's home |
| 12743 |
directory contains a file named `.guile', Guile loads that. |
| 12744 |
|
| 12745 |
** You can now use Guile as a shell script interpreter. |
| 12746 |
|
| 12747 |
To paraphrase the SCSH manual: |
| 12748 |
|
| 12749 |
When Unix tries to execute an executable file whose first two |
| 12750 |
characters are the `#!', it treats the file not as machine code to |
| 12751 |
be directly executed by the native processor, but as source code |
| 12752 |
to be executed by some interpreter. The interpreter to use is |
| 12753 |
specified immediately after the #! sequence on the first line of |
| 12754 |
the source file. The kernel reads in the name of the interpreter, |
| 12755 |
and executes that instead. It passes the interpreter the source |
| 12756 |
filename as its first argument, with the original arguments |
| 12757 |
following. Consult the Unix man page for the `exec' system call |
| 12758 |
for more information. |
| 12759 |
|
| 12760 |
Now you can use Guile as an interpreter, using a mechanism which is a |
| 12761 |
compatible subset of that provided by SCSH. |
| 12762 |
|
| 12763 |
Guile now recognizes a '-s' command line switch, whose argument is the |
| 12764 |
name of a file of Scheme code to load. It also treats the two |
| 12765 |
characters `#!' as the start of a comment, terminated by `!#'. Thus, |
| 12766 |
to make a file of Scheme code directly executable by Unix, insert the |
| 12767 |
following two lines at the top of the file: |
| 12768 |
|
| 12769 |
#!/usr/local/bin/guile -s |
| 12770 |
!# |
| 12771 |
|
| 12772 |
Guile treats the argument of the `-s' command-line switch as the name |
| 12773 |
of a file of Scheme code to load, and treats the sequence `#!' as the |
| 12774 |
start of a block comment, terminated by `!#'. |
| 12775 |
|
| 12776 |
For example, here's a version of 'echo' written in Scheme: |
| 12777 |
|
| 12778 |
#!/usr/local/bin/guile -s |
| 12779 |
!# |
| 12780 |
(let loop ((args (cdr (program-arguments)))) |
| 12781 |
(if (pair? args) |
| 12782 |
(begin |
| 12783 |
(display (car args)) |
| 12784 |
(if (pair? (cdr args)) |
| 12785 |
(display " ")) |
| 12786 |
(loop (cdr args))))) |
| 12787 |
(newline) |
| 12788 |
|
| 12789 |
Why does `#!' start a block comment terminated by `!#', instead of the |
| 12790 |
end of the line? That is the notation SCSH uses, and although we |
| 12791 |
don't yet support the other SCSH features that motivate that choice, |
| 12792 |
we would like to be backward-compatible with any existing Guile |
| 12793 |
scripts once we do. Furthermore, if the path to Guile on your system |
| 12794 |
is too long for your kernel, you can start the script with this |
| 12795 |
horrible hack: |
| 12796 |
|
| 12797 |
#!/bin/sh |
| 12798 |
exec /really/long/path/to/guile -s "$0" ${1+"$@"} |
| 12799 |
!# |
| 12800 |
|
| 12801 |
Note that some very old Unix systems don't support the `#!' syntax. |
| 12802 |
|
| 12803 |
|
| 12804 |
** You can now run Guile without installing it. |
| 12805 |
|
| 12806 |
Previous versions of the interactive Guile interpreter (`guile') |
| 12807 |
couldn't start up unless Guile's Scheme library had been installed; |
| 12808 |
they used the value of the environment variable `SCHEME_LOAD_PATH' |
| 12809 |
later on in the startup process, but not to find the startup code |
| 12810 |
itself. Now Guile uses `SCHEME_LOAD_PATH' in all searches for Scheme |
| 12811 |
code. |
| 12812 |
|
| 12813 |
To run Guile without installing it, build it in the normal way, and |
| 12814 |
then set the environment variable `SCHEME_LOAD_PATH' to a |
| 12815 |
colon-separated list of directories, including the top-level directory |
| 12816 |
of the Guile sources. For example, if you unpacked Guile so that the |
| 12817 |
full filename of this NEWS file is /home/jimb/guile-1.0b3/NEWS, then |
| 12818 |
you might say |
| 12819 |
|
| 12820 |
export SCHEME_LOAD_PATH=/home/jimb/my-scheme:/home/jimb/guile-1.0b3 |
| 12821 |
|
| 12822 |
|
| 12823 |
** Guile's read-eval-print loop no longer prints #<unspecified> |
| 12824 |
results. If the user wants to see this, she can evaluate the |
| 12825 |
expression (assert-repl-print-unspecified #t), perhaps in her startup |
| 12826 |
file. |
| 12827 |
|
| 12828 |
** Guile no longer shows backtraces by default when an error occurs; |
| 12829 |
however, it does display a message saying how to get one, and how to |
| 12830 |
request that they be displayed by default. After an error, evaluate |
| 12831 |
(backtrace) |
| 12832 |
to see a backtrace, and |
| 12833 |
(debug-enable 'backtrace) |
| 12834 |
to see them by default. |
| 12835 |
|
| 12836 |
|
| 12837 |
|
| 12838 |
* Changes to Guile Scheme: |
| 12839 |
|
| 12840 |
** Guile now distinguishes between #f and the empty list. |
| 12841 |
|
| 12842 |
This is for compatibility with the IEEE standard, the (possibly) |
| 12843 |
upcoming Revised^5 Report on Scheme, and many extant Scheme |
| 12844 |
implementations. |
| 12845 |
|
| 12846 |
Guile used to have #f and '() denote the same object, to make Scheme's |
| 12847 |
type system more compatible with Emacs Lisp's. However, the change |
| 12848 |
caused too much trouble for Scheme programmers, and we found another |
| 12849 |
way to reconcile Emacs Lisp with Scheme that didn't require this. |
| 12850 |
|
| 12851 |
|
| 12852 |
** Guile's delq, delv, delete functions, and their destructive |
| 12853 |
counterparts, delq!, delv!, and delete!, now remove all matching |
| 12854 |
elements from the list, not just the first. This matches the behavior |
| 12855 |
of the corresponding Emacs Lisp functions, and (I believe) the Maclisp |
| 12856 |
functions which inspired them. |
| 12857 |
|
| 12858 |
I recognize that this change may break code in subtle ways, but it |
| 12859 |
seems best to make the change before the FSF's first Guile release, |
| 12860 |
rather than after. |
| 12861 |
|
| 12862 |
|
| 12863 |
** The compiled-library-path function has been deleted from libguile. |
| 12864 |
|
| 12865 |
** The facilities for loading Scheme source files have changed. |
| 12866 |
|
| 12867 |
*** The variable %load-path now tells Guile which directories to search |
| 12868 |
for Scheme code. Its value is a list of strings, each of which names |
| 12869 |
a directory. |
| 12870 |
|
| 12871 |
*** The variable %load-extensions now tells Guile which extensions to |
| 12872 |
try appending to a filename when searching the load path. Its value |
| 12873 |
is a list of strings. Its default value is ("" ".scm"). |
| 12874 |
|
| 12875 |
*** (%search-load-path FILENAME) searches the directories listed in the |
| 12876 |
value of the %load-path variable for a Scheme file named FILENAME, |
| 12877 |
with all the extensions listed in %load-extensions. If it finds a |
| 12878 |
match, then it returns its full filename. If FILENAME is absolute, it |
| 12879 |
returns it unchanged. Otherwise, it returns #f. |
| 12880 |
|
| 12881 |
%search-load-path will not return matches that refer to directories. |
| 12882 |
|
| 12883 |
*** (primitive-load FILENAME :optional CASE-INSENSITIVE-P SHARP) |
| 12884 |
uses %seach-load-path to find a file named FILENAME, and loads it if |
| 12885 |
it finds it. If it can't read FILENAME for any reason, it throws an |
| 12886 |
error. |
| 12887 |
|
| 12888 |
The arguments CASE-INSENSITIVE-P and SHARP are interpreted as by the |
| 12889 |
`read' function. |
| 12890 |
|
| 12891 |
*** load uses the same searching semantics as primitive-load. |
| 12892 |
|
| 12893 |
*** The functions %try-load, try-load-with-path, %load, load-with-path, |
| 12894 |
basic-try-load-with-path, basic-load-with-path, try-load-module-with- |
| 12895 |
path, and load-module-with-path have been deleted. The functions |
| 12896 |
above should serve their purposes. |
| 12897 |
|
| 12898 |
*** If the value of the variable %load-hook is a procedure, |
| 12899 |
`primitive-load' applies its value to the name of the file being |
| 12900 |
loaded (without the load path directory name prepended). If its value |
| 12901 |
is #f, it is ignored. Otherwise, an error occurs. |
| 12902 |
|
| 12903 |
This is mostly useful for printing load notification messages. |
| 12904 |
|
| 12905 |
|
| 12906 |
** The function `eval!' is no longer accessible from the scheme level. |
| 12907 |
We can't allow operations which introduce glocs into the scheme level, |
| 12908 |
because Guile's type system can't handle these as data. Use `eval' or |
| 12909 |
`read-and-eval!' (see below) as replacement. |
| 12910 |
|
| 12911 |
** The new function read-and-eval! reads an expression from PORT, |
| 12912 |
evaluates it, and returns the result. This is more efficient than |
| 12913 |
simply calling `read' and `eval', since it is not necessary to make a |
| 12914 |
copy of the expression for the evaluator to munge. |
| 12915 |
|
| 12916 |
Its optional arguments CASE_INSENSITIVE_P and SHARP are interpreted as |
| 12917 |
for the `read' function. |
| 12918 |
|
| 12919 |
|
| 12920 |
** The function `int?' has been removed; its definition was identical |
| 12921 |
to that of `integer?'. |
| 12922 |
|
| 12923 |
** The functions `<?', `<?', `<=?', `=?', `>?', and `>=?'. Code should |
| 12924 |
use the R4RS names for these functions. |
| 12925 |
|
| 12926 |
** The function object-properties no longer returns the hash handle; |
| 12927 |
it simply returns the object's property list. |
| 12928 |
|
| 12929 |
** Many functions have been changed to throw errors, instead of |
| 12930 |
returning #f on failure. The point of providing exception handling in |
| 12931 |
the language is to simplify the logic of user code, but this is less |
| 12932 |
useful if Guile's primitives don't throw exceptions. |
| 12933 |
|
| 12934 |
** The function `fileno' has been renamed from `%fileno'. |
| 12935 |
|
| 12936 |
** The function primitive-mode->fdes returns #t or #f now, not 1 or 0. |
| 12937 |
|
| 12938 |
|
| 12939 |
* Changes to Guile's C interface: |
| 12940 |
|
| 12941 |
** The library's initialization procedure has been simplified. |
| 12942 |
scm_boot_guile now has the prototype: |
| 12943 |
|
| 12944 |
void scm_boot_guile (int ARGC, |
| 12945 |
char **ARGV, |
| 12946 |
void (*main_func) (), |
| 12947 |
void *closure); |
| 12948 |
|
| 12949 |
scm_boot_guile calls MAIN_FUNC, passing it CLOSURE, ARGC, and ARGV. |
| 12950 |
MAIN_FUNC should do all the work of the program (initializing other |
| 12951 |
packages, reading user input, etc.) before returning. When MAIN_FUNC |
| 12952 |
returns, call exit (0); this function never returns. If you want some |
| 12953 |
other exit value, MAIN_FUNC may call exit itself. |
| 12954 |
|
| 12955 |
scm_boot_guile arranges for program-arguments to return the strings |
| 12956 |
given by ARGC and ARGV. If MAIN_FUNC modifies ARGC/ARGV, should call |
| 12957 |
scm_set_program_arguments with the final list, so Scheme code will |
| 12958 |
know which arguments have been processed. |
| 12959 |
|
| 12960 |
scm_boot_guile establishes a catch-all catch handler which prints an |
| 12961 |
error message and exits the process. This means that Guile exits in a |
| 12962 |
coherent way when system errors occur and the user isn't prepared to |
| 12963 |
handle it. If the user doesn't like this behavior, they can establish |
| 12964 |
their own universal catcher in MAIN_FUNC to shadow this one. |
| 12965 |
|
| 12966 |
Why must the caller do all the real work from MAIN_FUNC? The garbage |
| 12967 |
collector assumes that all local variables of type SCM will be above |
| 12968 |
scm_boot_guile's stack frame on the stack. If you try to manipulate |
| 12969 |
SCM values after this function returns, it's the luck of the draw |
| 12970 |
whether the GC will be able to find the objects you allocate. So, |
| 12971 |
scm_boot_guile function exits, rather than returning, to discourage |
| 12972 |
people from making that mistake. |
| 12973 |
|
| 12974 |
The IN, OUT, and ERR arguments were removed; there are other |
| 12975 |
convenient ways to override these when desired. |
| 12976 |
|
| 12977 |
The RESULT argument was deleted; this function should never return. |
| 12978 |
|
| 12979 |
The BOOT_CMD argument was deleted; the MAIN_FUNC argument is more |
| 12980 |
general. |
| 12981 |
|
| 12982 |
|
| 12983 |
** Guile's header files should no longer conflict with your system's |
| 12984 |
header files. |
| 12985 |
|
| 12986 |
In order to compile code which #included <libguile.h>, previous |
| 12987 |
versions of Guile required you to add a directory containing all the |
| 12988 |
Guile header files to your #include path. This was a problem, since |
| 12989 |
Guile's header files have names which conflict with many systems' |
| 12990 |
header files. |
| 12991 |
|
| 12992 |
Now only <libguile.h> need appear in your #include path; you must |
| 12993 |
refer to all Guile's other header files as <libguile/mumble.h>. |
| 12994 |
Guile's installation procedure puts libguile.h in $(includedir), and |
| 12995 |
the rest in $(includedir)/libguile. |
| 12996 |
|
| 12997 |
|
| 12998 |
** Two new C functions, scm_protect_object and scm_unprotect_object, |
| 12999 |
have been added to the Guile library. |
| 13000 |
|
| 13001 |
scm_protect_object (OBJ) protects OBJ from the garbage collector. |
| 13002 |
OBJ will not be freed, even if all other references are dropped, |
| 13003 |
until someone does scm_unprotect_object (OBJ). Both functions |
| 13004 |
return OBJ. |
| 13005 |
|
| 13006 |
Note that calls to scm_protect_object do not nest. You can call |
| 13007 |
scm_protect_object any number of times on a given object, and the |
| 13008 |
next call to scm_unprotect_object will unprotect it completely. |
| 13009 |
|
| 13010 |
Basically, scm_protect_object and scm_unprotect_object just |
| 13011 |
maintain a list of references to things. Since the GC knows about |
| 13012 |
this list, all objects it mentions stay alive. scm_protect_object |
| 13013 |
adds its argument to the list; scm_unprotect_object remove its |
| 13014 |
argument from the list. |
| 13015 |
|
| 13016 |
|
| 13017 |
** scm_eval_0str now returns the value of the last expression |
| 13018 |
evaluated. |
| 13019 |
|
| 13020 |
** The new function scm_read_0str reads an s-expression from a |
| 13021 |
null-terminated string, and returns it. |
| 13022 |
|
| 13023 |
** The new function `scm_stdio_to_port' converts a STDIO file pointer |
| 13024 |
to a Scheme port object. |
| 13025 |
|
| 13026 |
** The new function `scm_set_program_arguments' allows C code to set |
| 13027 |
the value returned by the Scheme `program-arguments' function. |
| 13028 |
|
| 13029 |
|
| 13030 |
Older changes: |
| 13031 |
|
| 13032 |
* Guile no longer includes sophisticated Tcl/Tk support. |
| 13033 |
|
| 13034 |
The old Tcl/Tk support was unsatisfying to us, because it required the |
| 13035 |
user to link against the Tcl library, as well as Tk and Guile. The |
| 13036 |
interface was also un-lispy, in that it preserved Tcl/Tk's practice of |
| 13037 |
referring to widgets by names, rather than exporting widgets to Scheme |
| 13038 |
code as a special datatype. |
| 13039 |
|
| 13040 |
In the Usenix Tk Developer's Workshop held in July 1996, the Tcl/Tk |
| 13041 |
maintainers described some very interesting changes in progress to the |
| 13042 |
Tcl/Tk internals, which would facilitate clean interfaces between lone |
| 13043 |
Tk and other interpreters --- even for garbage-collected languages |
| 13044 |
like Scheme. They expected the new Tk to be publicly available in the |
| 13045 |
fall of 1996. |
| 13046 |
|
| 13047 |
Since it seems that Guile might soon have a new, cleaner interface to |
| 13048 |
lone Tk, and that the old Guile/Tk glue code would probably need to be |
| 13049 |
completely rewritten, we (Jim Blandy and Richard Stallman) have |
| 13050 |
decided not to support the old code. We'll spend the time instead on |
| 13051 |
a good interface to the newer Tk, as soon as it is available. |
| 13052 |
|
| 13053 |
Until then, gtcltk-lib provides trivial, low-maintenance functionality. |
| 13054 |
|
| 13055 |
|
| 13056 |
Copyright information: |
| 13057 |
|
| 13058 |
Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2006 Free Software Foundation, Inc. |
| 13059 |
|
| 13060 |
Permission is granted to anyone to make or distribute verbatim copies |
| 13061 |
of this document as received, in any medium, provided that the |
| 13062 |
copyright notice and this permission notice are preserved, |
| 13063 |
thus giving the recipient permission to redistribute in turn. |
| 13064 |
|
| 13065 |
Permission is granted to distribute modified versions |
| 13066 |
of this document, or of portions of it, |
| 13067 |
under the above conditions, provided also that they |
| 13068 |
carry prominent notices stating who last changed them. |
| 13069 |
|
| 13070 |
|
| 13071 |
Local variables: |
| 13072 |
mode: outline |
| 13073 |
paragraph-separate: "[ ]*$" |
| 13074 |
end: |