| 1 |
.\" -*- mode: troff; coding: utf-8 -*- |
| 2 |
.\" Automatically generated by Pod::Man 5.0102 (Pod::Simple 3.45) |
| 3 |
.\" |
| 4 |
.\" Standard preamble: |
| 5 |
.\" ======================================================================== |
| 6 |
.de Sp \" Vertical space (when we can't use .PP) |
| 7 |
.if t .sp .5v |
| 8 |
.if n .sp |
| 9 |
.. |
| 10 |
.de Vb \" Begin verbatim text |
| 11 |
.ft CW |
| 12 |
.nf |
| 13 |
.ne \\$1 |
| 14 |
.. |
| 15 |
.de Ve \" End verbatim text |
| 16 |
.ft R |
| 17 |
.fi |
| 18 |
.. |
| 19 |
.\" \*(C` and \*(C' are quotes in nroff, nothing in troff, for use with C<>. |
| 20 |
.ie n \{\ |
| 21 |
. ds C` "" |
| 22 |
. ds C' "" |
| 23 |
'br\} |
| 24 |
.el\{\ |
| 25 |
. ds C` |
| 26 |
. ds C' |
| 27 |
'br\} |
| 28 |
.\" |
| 29 |
.\" Escape single quotes in literal strings from groff's Unicode transform. |
| 30 |
.ie \n(.g .ds Aq \(aq |
| 31 |
.el .ds Aq ' |
| 32 |
.\" |
| 33 |
.\" If the F register is >0, we'll generate index entries on stderr for |
| 34 |
.\" titles (.TH), headers (.SH), subsections (.SS), items (.Ip), and index |
| 35 |
.\" entries marked with X<> in POD. Of course, you'll have to process the |
| 36 |
.\" output yourself in some meaningful fashion. |
| 37 |
.\" |
| 38 |
.\" Avoid warning from groff about undefined register 'F'. |
| 39 |
.de IX |
| 40 |
.. |
| 41 |
.nr rF 0 |
| 42 |
.if \n(.g .if rF .nr rF 1 |
| 43 |
.if (\n(rF:(\n(.g==0)) \{\ |
| 44 |
. if \nF \{\ |
| 45 |
. de IX |
| 46 |
. tm Index:\\$1\t\\n%\t"\\$2" |
| 47 |
.. |
| 48 |
. if !\nF==2 \{\ |
| 49 |
. nr % 0 |
| 50 |
. nr F 2 |
| 51 |
. \} |
| 52 |
. \} |
| 53 |
.\} |
| 54 |
.rr rF |
| 55 |
.\" ======================================================================== |
| 56 |
.\" |
| 57 |
.IX Title "PERL5180DELTA 1" |
| 58 |
.TH PERL5180DELTA 1 2025-11-16 "perl v5.40.1" "Perl Programmers Reference Guide" |
| 59 |
.\" For nroff, turn off justification. Always turn off hyphenation; it makes |
| 60 |
.\" way too many mistakes in technical documents. |
| 61 |
.if n .ad l |
| 62 |
.nh |
| 63 |
.SH NAME |
| 64 |
perl5180delta \- what is new for perl v5.18.0 |
| 65 |
.SH DESCRIPTION |
| 66 |
.IX Header "DESCRIPTION" |
| 67 |
This document describes differences between the v5.16.0 release and the v5.18.0 |
| 68 |
release. |
| 69 |
.PP |
| 70 |
If you are upgrading from an earlier release such as v5.14.0, first read |
| 71 |
perl5160delta, which describes differences between v5.14.0 and v5.16.0. |
| 72 |
.SH "Core Enhancements" |
| 73 |
.IX Header "Core Enhancements" |
| 74 |
.SS "New mechanism for experimental features" |
| 75 |
.IX Subsection "New mechanism for experimental features" |
| 76 |
Newly-added experimental features will now require this incantation: |
| 77 |
.PP |
| 78 |
.Vb 2 |
| 79 |
\& no warnings "experimental::feature_name"; |
| 80 |
\& use feature "feature_name"; # would warn without the prev line |
| 81 |
.Ve |
| 82 |
.PP |
| 83 |
There is a new warnings category, called "experimental", containing |
| 84 |
warnings that the feature pragma emits when enabling experimental |
| 85 |
features. |
| 86 |
.PP |
| 87 |
Newly-added experimental features will also be given special warning IDs, |
| 88 |
which consist of "experimental::" followed by the name of the feature. (The |
| 89 |
plan is to extend this mechanism eventually to all warnings, to allow them |
| 90 |
to be enabled or disabled individually, and not just by category.) |
| 91 |
.PP |
| 92 |
By saying |
| 93 |
.PP |
| 94 |
.Vb 1 |
| 95 |
\& no warnings "experimental::feature_name"; |
| 96 |
.Ve |
| 97 |
.PP |
| 98 |
you are taking responsibility for any breakage that future changes to, or |
| 99 |
removal of, the feature may cause. |
| 100 |
.PP |
| 101 |
Since some features (like \f(CW\*(C`~~\*(C'\fR or \f(CW\*(C`my $_\*(C'\fR) now emit experimental warnings, |
| 102 |
and you may want to disable them in code that is also run on perls that do not |
| 103 |
recognize these warning categories, consider using the \f(CW\*(C`if\*(C'\fR pragma like this: |
| 104 |
.PP |
| 105 |
.Vb 1 |
| 106 |
\& no if $] >= 5.018, warnings => "experimental::feature_name"; |
| 107 |
.Ve |
| 108 |
.PP |
| 109 |
Existing experimental features may begin emitting these warnings, too. Please |
| 110 |
consult perlexperiment for information on which features are considered |
| 111 |
experimental. |
| 112 |
.SS "Hash overhaul" |
| 113 |
.IX Subsection "Hash overhaul" |
| 114 |
Changes to the implementation of hashes in perl v5.18.0 will be one of the most |
| 115 |
visible changes to the behavior of existing code. |
| 116 |
.PP |
| 117 |
By default, two distinct hash variables with identical keys and values may now |
| 118 |
provide their contents in a different order where it was previously identical. |
| 119 |
.PP |
| 120 |
When encountering these changes, the key to cleaning up from them is to accept |
| 121 |
that \fBhashes are unordered collections\fR and to act accordingly. |
| 122 |
.PP |
| 123 |
\fIHash randomization\fR |
| 124 |
.IX Subsection "Hash randomization" |
| 125 |
.PP |
| 126 |
The seed used by Perl's hash function is now random. This means that the |
| 127 |
order which keys/values will be returned from functions like \f(CWkeys()\fR, |
| 128 |
\&\f(CWvalues()\fR, and \f(CWeach()\fR will differ from run to run. |
| 129 |
.PP |
| 130 |
This change was introduced to make Perl's hashes more robust to algorithmic |
| 131 |
complexity attacks, and also because we discovered that it exposes hash |
| 132 |
ordering dependency bugs and makes them easier to track down. |
| 133 |
.PP |
| 134 |
Toolchain maintainers might want to invest in additional infrastructure to |
| 135 |
test for things like this. Running tests several times in a row and then |
| 136 |
comparing results will make it easier to spot hash order dependencies in |
| 137 |
code. Authors are strongly encouraged not to expose the key order of |
| 138 |
Perl's hashes to insecure audiences. |
| 139 |
.PP |
| 140 |
Further, every hash has its own iteration order, which should make it much |
| 141 |
more difficult to determine what the current hash seed is. |
| 142 |
.PP |
| 143 |
\fINew hash functions\fR |
| 144 |
.IX Subsection "New hash functions" |
| 145 |
.PP |
| 146 |
Perl v5.18 includes support for multiple hash functions, and changed |
| 147 |
the default (to ONE_AT_A_TIME_HARD), you can choose a different |
| 148 |
algorithm by defining a symbol at compile time. For a current list, |
| 149 |
consult the \fIINSTALL\fR document. Note that as of Perl v5.18 we can |
| 150 |
only recommend use of the default or SIPHASH. All the others are |
| 151 |
known to have security issues and are for research purposes only. |
| 152 |
.PP |
| 153 |
\fIPERL_HASH_SEED environment variable now takes a hex value\fR |
| 154 |
.IX Subsection "PERL_HASH_SEED environment variable now takes a hex value" |
| 155 |
.PP |
| 156 |
\&\f(CW\*(C`PERL_HASH_SEED\*(C'\fR no longer accepts an integer as a parameter; |
| 157 |
instead the value is expected to be a binary value encoded in a hex |
| 158 |
string, such as "0xf5867c55039dc724". This is to make the |
| 159 |
infrastructure support hash seeds of arbitrary lengths, which might |
| 160 |
exceed that of an integer. (SipHash uses a 16 byte seed.) |
| 161 |
.PP |
| 162 |
\fIPERL_PERTURB_KEYS environment variable added\fR |
| 163 |
.IX Subsection "PERL_PERTURB_KEYS environment variable added" |
| 164 |
.PP |
| 165 |
The \f(CW\*(C`PERL_PERTURB_KEYS\*(C'\fR environment variable allows one to control the level of |
| 166 |
randomization applied to \f(CW\*(C`keys\*(C'\fR and friends. |
| 167 |
.PP |
| 168 |
When \f(CW\*(C`PERL_PERTURB_KEYS\*(C'\fR is 0, perl will not randomize the key order at all. The |
| 169 |
chance that \f(CW\*(C`keys\*(C'\fR changes due to an insert will be the same as in previous |
| 170 |
perls, basically only when the bucket size is changed. |
| 171 |
.PP |
| 172 |
When \f(CW\*(C`PERL_PERTURB_KEYS\*(C'\fR is 1, perl will randomize keys in a non-repeatable |
| 173 |
way. The chance that \f(CW\*(C`keys\*(C'\fR changes due to an insert will be very high. This |
| 174 |
is the most secure and default mode. |
| 175 |
.PP |
| 176 |
When \f(CW\*(C`PERL_PERTURB_KEYS\*(C'\fR is 2, perl will randomize keys in a repeatable way. |
| 177 |
Repeated runs of the same program should produce the same output every time. |
| 178 |
.PP |
| 179 |
\&\f(CW\*(C`PERL_HASH_SEED\*(C'\fR implies a non-default \f(CW\*(C`PERL_PERTURB_KEYS\*(C'\fR setting. Setting |
| 180 |
\&\f(CW\*(C`PERL_HASH_SEED=0\*(C'\fR (exactly one 0) implies \f(CW\*(C`PERL_PERTURB_KEYS=0\*(C'\fR (hash key |
| 181 |
randomization disabled); setting \f(CW\*(C`PERL_HASH_SEED\*(C'\fR to any other value implies |
| 182 |
\&\f(CW\*(C`PERL_PERTURB_KEYS=2\*(C'\fR (deterministic and repeatable hash key randomization). |
| 183 |
Specifying \f(CW\*(C`PERL_PERTURB_KEYS\*(C'\fR explicitly to a different level overrides this |
| 184 |
behavior. |
| 185 |
.PP |
| 186 |
\fR\f(BIHash::Util::hash_seed()\fR\fI now returns a string\fR |
| 187 |
.IX Subsection "Hash::Util::hash_seed() now returns a string" |
| 188 |
.PP |
| 189 |
\&\fBHash::Util::hash_seed()\fR now returns a string instead of an integer. This |
| 190 |
is to make the infrastructure support hash seeds of arbitrary lengths |
| 191 |
which might exceed that of an integer. (SipHash uses a 16 byte seed.) |
| 192 |
.PP |
| 193 |
\fIOutput of PERL_HASH_SEED_DEBUG has been changed\fR |
| 194 |
.IX Subsection "Output of PERL_HASH_SEED_DEBUG has been changed" |
| 195 |
.PP |
| 196 |
The environment variable PERL_HASH_SEED_DEBUG now makes perl show both the |
| 197 |
hash function perl was built with, \fIand\fR the seed, in hex, in use for that |
| 198 |
process. Code parsing this output, should it exist, must change to accommodate |
| 199 |
the new format. Example of the new format: |
| 200 |
.PP |
| 201 |
.Vb 2 |
| 202 |
\& $ PERL_HASH_SEED_DEBUG=1 ./perl \-e1 |
| 203 |
\& HASH_FUNCTION = MURMUR3 HASH_SEED = 0x1476bb9f |
| 204 |
.Ve |
| 205 |
.SS "Upgrade to Unicode 6.2" |
| 206 |
.IX Subsection "Upgrade to Unicode 6.2" |
| 207 |
Perl now supports Unicode 6.2. A list of changes from Unicode |
| 208 |
6.1 is at <http://www.unicode.org/versions/Unicode6.2.0>. |
| 209 |
.SS "Character name aliases may now include non\-Latin1\-range characters" |
| 210 |
.IX Subsection "Character name aliases may now include non-Latin1-range characters" |
| 211 |
It is possible to define your own names for characters for use in |
| 212 |
\&\f(CW\*(C`\eN{...}\*(C'\fR, \f(CWcharnames::vianame()\fR, etc. These names can now |
| 213 |
consist of characters from the whole Unicode range. This allows for |
| 214 |
names to be in your native language, and not just English. Certain |
| 215 |
restrictions apply to the characters that may be used (you can't define |
| 216 |
a name that has punctuation in it, for example). See "CUSTOM |
| 217 |
ALIASES" in charnames. |
| 218 |
.SS "New DTrace probes" |
| 219 |
.IX Subsection "New DTrace probes" |
| 220 |
The following new DTrace probes have been added: |
| 221 |
.IP \(bu 4 |
| 222 |
\&\f(CW\*(C`op\-entry\*(C'\fR |
| 223 |
.IP \(bu 4 |
| 224 |
\&\f(CW\*(C`loading\-file\*(C'\fR |
| 225 |
.IP \(bu 4 |
| 226 |
\&\f(CW\*(C`loaded\-file\*(C'\fR |
| 227 |
.ie n .SS """${^LAST_FH}""" |
| 228 |
.el .SS \f(CW${^LAST_FH}\fP |
| 229 |
.IX Subsection "${^LAST_FH}" |
| 230 |
This new variable provides access to the filehandle that was last read. |
| 231 |
This is the handle used by \f(CW$.\fR and by \f(CW\*(C`tell\*(C'\fR and \f(CW\*(C`eof\*(C'\fR without |
| 232 |
arguments. |
| 233 |
.SS "Regular Expression Set Operations" |
| 234 |
.IX Subsection "Regular Expression Set Operations" |
| 235 |
This is an \fBexperimental\fR feature to allow matching against the union, |
| 236 |
intersection, etc., of sets of code points, similar to |
| 237 |
Unicode::Regex::Set. It can also be used to extend \f(CW\*(C`/x\*(C'\fR processing |
| 238 |
to [bracketed] character classes, and as a replacement of user-defined |
| 239 |
properties, allowing more complex expressions than they do. See |
| 240 |
"Extended Bracketed Character Classes" in perlrecharclass. |
| 241 |
.SS "Lexical subroutines" |
| 242 |
.IX Subsection "Lexical subroutines" |
| 243 |
This new feature is still considered \fBexperimental\fR. To enable it: |
| 244 |
.PP |
| 245 |
.Vb 3 |
| 246 |
\& use 5.018; |
| 247 |
\& no warnings "experimental::lexical_subs"; |
| 248 |
\& use feature "lexical_subs"; |
| 249 |
.Ve |
| 250 |
.PP |
| 251 |
You can now declare subroutines with \f(CW\*(C`state sub foo\*(C'\fR, \f(CW\*(C`my sub foo\*(C'\fR, and |
| 252 |
\&\f(CW\*(C`our sub foo\*(C'\fR. (\f(CW\*(C`state sub\*(C'\fR requires that the "state" feature be |
| 253 |
enabled, unless you write it as \f(CW\*(C`CORE::state sub foo\*(C'\fR.) |
| 254 |
.PP |
| 255 |
\&\f(CW\*(C`state sub\*(C'\fR creates a subroutine visible within the lexical scope in which |
| 256 |
it is declared. The subroutine is shared between calls to the outer sub. |
| 257 |
.PP |
| 258 |
\&\f(CW\*(C`my sub\*(C'\fR declares a lexical subroutine that is created each time the |
| 259 |
enclosing block is entered. \f(CW\*(C`state sub\*(C'\fR is generally slightly faster than |
| 260 |
\&\f(CW\*(C`my sub\*(C'\fR. |
| 261 |
.PP |
| 262 |
\&\f(CW\*(C`our sub\*(C'\fR declares a lexical alias to the package subroutine of the same |
| 263 |
name. |
| 264 |
.PP |
| 265 |
For more information, see "Lexical Subroutines" in perlsub. |
| 266 |
.SS "Computed Labels" |
| 267 |
.IX Subsection "Computed Labels" |
| 268 |
The loop controls \f(CW\*(C`next\*(C'\fR, \f(CW\*(C`last\*(C'\fR and \f(CW\*(C`redo\*(C'\fR, and the special \f(CW\*(C`dump\*(C'\fR |
| 269 |
operator, now allow arbitrary expressions to be used to compute labels at run |
| 270 |
time. Previously, any argument that was not a constant was treated as the |
| 271 |
empty string. |
| 272 |
.SS "More CORE:: subs" |
| 273 |
.IX Subsection "More CORE:: subs" |
| 274 |
Several more built-in functions have been added as subroutines to the |
| 275 |
CORE:: namespace \- namely, those non-overridable keywords that can be |
| 276 |
implemented without custom parsers: \f(CW\*(C`defined\*(C'\fR, \f(CW\*(C`delete\*(C'\fR, \f(CW\*(C`exists\*(C'\fR, |
| 277 |
\&\f(CW\*(C`glob\*(C'\fR, \f(CW\*(C`pos\*(C'\fR, \f(CW\*(C`prototype\*(C'\fR, \f(CW\*(C`scalar\*(C'\fR, \f(CW\*(C`split\*(C'\fR, \f(CW\*(C`study\*(C'\fR, and \f(CW\*(C`undef\*(C'\fR. |
| 278 |
.PP |
| 279 |
As some of these have prototypes, \f(CWprototype(\*(AqCORE::...\*(Aq)\fR has been |
| 280 |
changed to not make a distinction between overridable and non-overridable |
| 281 |
keywords. This is to make \f(CWprototype(\*(AqCORE::pos\*(Aq)\fR consistent with |
| 282 |
\&\f(CWprototype(&CORE::pos)\fR. |
| 283 |
.ie n .SS """kill"" with negative signal names" |
| 284 |
.el .SS "\f(CWkill\fP with negative signal names" |
| 285 |
.IX Subsection "kill with negative signal names" |
| 286 |
\&\f(CW\*(C`kill\*(C'\fR has always allowed a negative signal number, which kills the |
| 287 |
process group instead of a single process. It has also allowed signal |
| 288 |
names. But it did not behave consistently, because negative signal names |
| 289 |
were treated as 0. Now negative signals names like \f(CW\*(C`\-INT\*(C'\fR are supported |
| 290 |
and treated the same way as \-2 [perl #112990]. |
| 291 |
.SH Security |
| 292 |
.IX Header "Security" |
| 293 |
.SS "See also: hash overhaul" |
| 294 |
.IX Subsection "See also: hash overhaul" |
| 295 |
Some of the changes in the hash overhaul were made to |
| 296 |
enhance security. Please read that section. |
| 297 |
.ie n .SS """Storable"" security warning in documentation" |
| 298 |
.el .SS "\f(CWStorable\fP security warning in documentation" |
| 299 |
.IX Subsection "Storable security warning in documentation" |
| 300 |
The documentation for \f(CW\*(C`Storable\*(C'\fR now includes a section which warns readers |
| 301 |
of the danger of accepting Storable documents from untrusted sources. The |
| 302 |
short version is that deserializing certain types of data can lead to loading |
| 303 |
modules and other code execution. This is documented behavior and wanted |
| 304 |
behavior, but this opens an attack vector for malicious entities. |
| 305 |
.ie n .SS """Locale::Maketext"" allowed code injection via a malicious template" |
| 306 |
.el .SS "\f(CWLocale::Maketext\fP allowed code injection via a malicious template" |
| 307 |
.IX Subsection "Locale::Maketext allowed code injection via a malicious template" |
| 308 |
If users could provide a translation string to Locale::Maketext, this could be |
| 309 |
used to invoke arbitrary Perl subroutines available in the current process. |
| 310 |
.PP |
| 311 |
This has been fixed, but it is still possible to invoke any method provided by |
| 312 |
\&\f(CW\*(C`Locale::Maketext\*(C'\fR itself or a subclass that you are using. One of these |
| 313 |
methods in turn will invoke the Perl core's \f(CW\*(C`sprintf\*(C'\fR subroutine. |
| 314 |
.PP |
| 315 |
In summary, allowing users to provide translation strings without auditing |
| 316 |
them is a bad idea. |
| 317 |
.PP |
| 318 |
This vulnerability is documented in CVE\-2012\-6329. |
| 319 |
.SS "Avoid calling memset with a negative count" |
| 320 |
.IX Subsection "Avoid calling memset with a negative count" |
| 321 |
Poorly written perl code that allows an attacker to specify the count to perl's |
| 322 |
\&\f(CW\*(C`x\*(C'\fR string repeat operator can already cause a memory exhaustion |
| 323 |
denial-of-service attack. A flaw in versions of perl before v5.15.5 can escalate |
| 324 |
that into a heap buffer overrun; coupled with versions of glibc before 2.16, it |
| 325 |
possibly allows the execution of arbitrary code. |
| 326 |
.PP |
| 327 |
The flaw addressed to this commit has been assigned identifier CVE\-2012\-5195 |
| 328 |
and was researched by Tim Brown. |
| 329 |
.SH "Incompatible Changes" |
| 330 |
.IX Header "Incompatible Changes" |
| 331 |
.SS "See also: hash overhaul" |
| 332 |
.IX Subsection "See also: hash overhaul" |
| 333 |
Some of the changes in the hash overhaul are not fully |
| 334 |
compatible with previous versions of perl. Please read that section. |
| 335 |
.ie n .SS "An unknown character name in ""\eN{...}"" is now a syntax error" |
| 336 |
.el .SS "An unknown character name in \f(CW\eN{...}\fP is now a syntax error" |
| 337 |
.IX Subsection "An unknown character name in N{...} is now a syntax error" |
| 338 |
Previously, it warned, and the Unicode REPLACEMENT CHARACTER was |
| 339 |
substituted. Unicode now recommends that this situation be a syntax |
| 340 |
error. Also, the previous behavior led to some confusing warnings and |
| 341 |
behaviors, and since the REPLACEMENT CHARACTER has no use other than as |
| 342 |
a stand-in for some unknown character, any code that has this problem is |
| 343 |
buggy. |
| 344 |
.ie n .SS "Formerly deprecated characters in ""\eN{}"" character name aliases are now errors" |
| 345 |
.el .SS "Formerly deprecated characters in \f(CW\eN{}\fP character name aliases are now errors" |
| 346 |
.IX Subsection "Formerly deprecated characters in N{} character name aliases are now errors" |
| 347 |
Since v5.12.0, it has been deprecated to use certain characters in |
| 348 |
user-defined \f(CW\*(C`\eN{...}\*(C'\fR character names. These now cause a syntax |
| 349 |
error. For example, it is now an error to begin a name with a digit, |
| 350 |
such as in |
| 351 |
.PP |
| 352 |
.Vb 1 |
| 353 |
\& my $undraftable = "\eN{4F}"; # Syntax error! |
| 354 |
.Ve |
| 355 |
.PP |
| 356 |
or to have commas anywhere in the name. See "CUSTOM ALIASES" in charnames. |
| 357 |
.ie n .SS """\eN{BELL}"" now refers to U+1F514 instead of U+0007" |
| 358 |
.el .SS "\f(CW\eN{BELL}\fP now refers to U+1F514 instead of U+0007" |
| 359 |
.IX Subsection "N{BELL} now refers to U+1F514 instead of U+0007" |
| 360 |
Unicode 6.0 reused the name "BELL" for a different code point than it |
| 361 |
traditionally had meant. Since Perl v5.14, use of this name still |
| 362 |
referred to U+0007, but would raise a deprecation warning. Now, "BELL" |
| 363 |
refers to U+1F514, and the name for U+0007 is "ALERT". All the |
| 364 |
functions in charnames have been correspondingly updated. |
| 365 |
.SS "New Restrictions in Multi-Character Case-Insensitive Matching in Regular Expression Bracketed Character Classes" |
| 366 |
.IX Subsection "New Restrictions in Multi-Character Case-Insensitive Matching in Regular Expression Bracketed Character Classes" |
| 367 |
Unicode has now withdrawn their previous recommendation for regular |
| 368 |
expressions to automatically handle cases where a single character can |
| 369 |
match multiple characters case-insensitively, for example, the letter |
| 370 |
LATIN SMALL LETTER SHARP S and the sequence \f(CW\*(C`ss\*(C'\fR. This is because |
| 371 |
it turns out to be impracticable to do this correctly in all |
| 372 |
circumstances. Because Perl has tried to do this as best it can, it |
| 373 |
will continue to do so. (We are considering an option to turn it off.) |
| 374 |
However, a new restriction is being added on such matches when they |
| 375 |
occur in [bracketed] character classes. People were specifying |
| 376 |
things such as \f(CW\*(C`/[\e0\-\exff]/i\*(C'\fR, and being surprised that it matches the |
| 377 |
two character sequence \f(CW\*(C`ss\*(C'\fR (since LATIN SMALL LETTER SHARP S occurs in |
| 378 |
this range). This behavior is also inconsistent with using a |
| 379 |
property instead of a range: \f(CW\*(C`\ep{Block=Latin1}\*(C'\fR also includes LATIN |
| 380 |
SMALL LETTER SHARP S, but \f(CW\*(C`/[\ep{Block=Latin1}]/i\*(C'\fR does not match \f(CW\*(C`ss\*(C'\fR. |
| 381 |
The new rule is that for there to be a multi-character case-insensitive |
| 382 |
match within a bracketed character class, the character must be |
| 383 |
explicitly listed, and not as an end point of a range. This more |
| 384 |
closely obeys the Principle of Least Astonishment. See |
| 385 |
"Bracketed Character Classes" in perlrecharclass. Note that a bug [perl |
| 386 |
#89774], now fixed as part of this change, prevented the previous |
| 387 |
behavior from working fully. |
| 388 |
.SS "Explicit rules for variable names and identifiers" |
| 389 |
.IX Subsection "Explicit rules for variable names and identifiers" |
| 390 |
Due to an oversight, single character variable names in v5.16 were |
| 391 |
completely unrestricted. This opened the door to several kinds of |
| 392 |
insanity. As of v5.18, these now follow the rules of other identifiers, |
| 393 |
in addition to accepting characters that match the \f(CW\*(C`\ep{POSIX_Punct}\*(C'\fR |
| 394 |
property. |
| 395 |
.PP |
| 396 |
There is no longer any difference in the parsing of identifiers |
| 397 |
specified by using braces versus without braces. For instance, perl |
| 398 |
used to allow \f(CW\*(C`${foo:bar}\*(C'\fR (with a single colon) but not \f(CW$foo:bar\fR. |
| 399 |
Now that both are handled by a single code path, they are both treated |
| 400 |
the same way: both are forbidden. Note that this change is about the |
| 401 |
range of permissible literal identifiers, not other expressions. |
| 402 |
.SS "Vertical tabs are now whitespace" |
| 403 |
.IX Subsection "Vertical tabs are now whitespace" |
| 404 |
No one could recall why \f(CW\*(C`\es\*(C'\fR didn't match \f(CW\*(C`\ecK\*(C'\fR, the vertical tab. |
| 405 |
Now it does. Given the extreme rarity of that character, very little |
| 406 |
breakage is expected. That said, here's what it means: |
| 407 |
.PP |
| 408 |
\&\f(CW\*(C`\es\*(C'\fR in a regex now matches a vertical tab in all circumstances. |
| 409 |
.PP |
| 410 |
Literal vertical tabs in a regex literal are ignored when the \f(CW\*(C`/x\*(C'\fR |
| 411 |
modifier is used. |
| 412 |
.PP |
| 413 |
Leading vertical tabs, alone or mixed with other whitespace, are now |
| 414 |
ignored when interpreting a string as a number. For example: |
| 415 |
.PP |
| 416 |
.Vb 2 |
| 417 |
\& $dec = " \ecK \et 123"; |
| 418 |
\& $hex = " \ecK \et 0xF"; |
| 419 |
\& |
| 420 |
\& say 0 + $dec; # was 0 with warning, now 123 |
| 421 |
\& say int $dec; # was 0, now 123 |
| 422 |
\& say oct $hex; # was 0, now 15 |
| 423 |
.Ve |
| 424 |
.ie n .SS """/(?{})/"" and ""/(??{})/"" have been heavily reworked" |
| 425 |
.el .SS "\f(CW/(?{})/\fP and \f(CW/(??{})/\fP have been heavily reworked" |
| 426 |
.IX Subsection "/(?{})/ and /(??{})/ have been heavily reworked" |
| 427 |
The implementation of this feature has been almost completely rewritten. |
| 428 |
Although its main intent is to fix bugs, some behaviors, especially |
| 429 |
related to the scope of lexical variables, will have changed. This is |
| 430 |
described more fully in the "Selected Bug Fixes" section. |
| 431 |
.SS "Stricter parsing of substitution replacement" |
| 432 |
.IX Subsection "Stricter parsing of substitution replacement" |
| 433 |
It is no longer possible to abuse the way the parser parses \f(CW\*(C`s///e\*(C'\fR like |
| 434 |
this: |
| 435 |
.PP |
| 436 |
.Vb 3 |
| 437 |
\& %_=(_,"Just another "); |
| 438 |
\& $_="Perl hacker,\en"; |
| 439 |
\& s//_}\->{_/e;print |
| 440 |
.Ve |
| 441 |
.ie n .SS """given"" now aliases the global $_" |
| 442 |
.el .SS "\f(CWgiven\fP now aliases the global \f(CW$_\fP" |
| 443 |
.IX Subsection "given now aliases the global $_" |
| 444 |
Instead of assigning to an implicit lexical \f(CW$_\fR, \f(CW\*(C`given\*(C'\fR now makes the |
| 445 |
global \f(CW$_\fR an alias for its argument, just like \f(CW\*(C`foreach\*(C'\fR. However, it |
| 446 |
still uses lexical \f(CW$_\fR if there is lexical \f(CW$_\fR in scope (again, just like |
| 447 |
\&\f(CW\*(C`foreach\*(C'\fR) [perl #114020]. |
| 448 |
.SS "The smartmatch family of features are now experimental" |
| 449 |
.IX Subsection "The smartmatch family of features are now experimental" |
| 450 |
Smart match, added in v5.10.0 and significantly revised in v5.10.1, has been |
| 451 |
a regular point of complaint. Although there are a number of ways in which |
| 452 |
it is useful, it has also proven problematic and confusing for both users and |
| 453 |
implementors of Perl. There have been a number of proposals on how to best |
| 454 |
address the problem. It is clear that smartmatch is almost certainly either |
| 455 |
going to change or go away in the future. Relying on its current behavior |
| 456 |
is not recommended. |
| 457 |
.PP |
| 458 |
Warnings will now be issued when the parser sees \f(CW\*(C`~~\*(C'\fR, \f(CW\*(C`given\*(C'\fR, or \f(CW\*(C`when\*(C'\fR. |
| 459 |
To disable these warnings, you can add this line to the appropriate scope: |
| 460 |
.PP |
| 461 |
.Vb 1 |
| 462 |
\& no if $] >= 5.018, warnings => "experimental::smartmatch"; |
| 463 |
.Ve |
| 464 |
.PP |
| 465 |
Consider, though, replacing the use of these features, as they may change |
| 466 |
behavior again before becoming stable. |
| 467 |
.ie n .SS "Lexical $_ is now experimental" |
| 468 |
.el .SS "Lexical \f(CW$_\fP is now experimental" |
| 469 |
.IX Subsection "Lexical $_ is now experimental" |
| 470 |
Since it was introduced in Perl v5.10, it has caused much confusion with no |
| 471 |
obvious solution: |
| 472 |
.IP \(bu 4 |
| 473 |
Various modules (e.g., List::Util) expect callback routines to use the |
| 474 |
global \f(CW$_\fR. \f(CW\*(C`use List::Util \*(Aqfirst\*(Aq; my $_; first { $_ == 1 } @list\*(C'\fR |
| 475 |
does not work as one would expect. |
| 476 |
.IP \(bu 4 |
| 477 |
A \f(CW\*(C`my $_\*(C'\fR declaration earlier in the same file can cause confusing closure |
| 478 |
warnings. |
| 479 |
.IP \(bu 4 |
| 480 |
The "_" subroutine prototype character allows called subroutines to access |
| 481 |
your lexical \f(CW$_\fR, so it is not really private after all. |
| 482 |
.IP \(bu 4 |
| 483 |
Nevertheless, subroutines with a "(@)" prototype and methods cannot access |
| 484 |
the caller's lexical \f(CW$_\fR, unless they are written in XS. |
| 485 |
.IP \(bu 4 |
| 486 |
But even XS routines cannot access a lexical \f(CW$_\fR declared, not in the |
| 487 |
calling subroutine, but in an outer scope, iff that subroutine happened not |
| 488 |
to mention \f(CW$_\fR or use any operators that default to \f(CW$_\fR. |
| 489 |
.PP |
| 490 |
It is our hope that lexical \f(CW$_\fR can be rehabilitated, but this may |
| 491 |
cause changes in its behavior. Please use it with caution until it |
| 492 |
becomes stable. |
| 493 |
.ie n .SS "\fBreadline()\fP with ""$/ = \eN"" now reads N characters, not N bytes" |
| 494 |
.el .SS "\fBreadline()\fP with \f(CW$/ = \eN\fP now reads N characters, not N bytes" |
| 495 |
.IX Subsection "readline() with $/ = N now reads N characters, not N bytes" |
| 496 |
Previously, when reading from a stream with I/O layers such as |
| 497 |
\&\f(CW\*(C`encoding\*(C'\fR, the \fBreadline()\fR function, otherwise known as the \f(CW\*(C`<>\*(C'\fR |
| 498 |
operator, would read \fIN\fR bytes from the top-most layer. [perl #79960] |
| 499 |
.PP |
| 500 |
Now, \fIN\fR characters are read instead. |
| 501 |
.PP |
| 502 |
There is no change in behaviour when reading from streams with no |
| 503 |
extra layers, since bytes map exactly to characters. |
| 504 |
.ie n .SS "Overridden ""glob"" is now passed one argument" |
| 505 |
.el .SS "Overridden \f(CWglob\fP is now passed one argument" |
| 506 |
.IX Subsection "Overridden glob is now passed one argument" |
| 507 |
\&\f(CW\*(C`glob\*(C'\fR overrides used to be passed a magical undocumented second argument |
| 508 |
that identified the caller. Nothing on CPAN was using this, and it got in |
| 509 |
the way of a bug fix, so it was removed. If you really need to identify |
| 510 |
the caller, see Devel::Callsite on CPAN. |
| 511 |
.SS "Here doc parsing" |
| 512 |
.IX Subsection "Here doc parsing" |
| 513 |
The body of a here document inside a quote-like operator now always begins |
| 514 |
on the line after the "<<foo" marker. Previously, it was documented to |
| 515 |
begin on the line following the containing quote-like operator, but that |
| 516 |
was only sometimes the case [perl #114040]. |
| 517 |
.SS "Alphanumeric operators must now be separated from the closing delimiter of regular expressions" |
| 518 |
.IX Subsection "Alphanumeric operators must now be separated from the closing delimiter of regular expressions" |
| 519 |
You may no longer write something like: |
| 520 |
.PP |
| 521 |
.Vb 1 |
| 522 |
\& m/a/and 1 |
| 523 |
.Ve |
| 524 |
.PP |
| 525 |
Instead you must write |
| 526 |
.PP |
| 527 |
.Vb 1 |
| 528 |
\& m/a/ and 1 |
| 529 |
.Ve |
| 530 |
.PP |
| 531 |
with whitespace separating the operator from the closing delimiter of |
| 532 |
the regular expression. Not having whitespace has resulted in a |
| 533 |
deprecation warning since Perl v5.14.0. |
| 534 |
.SS "qw(...) can no longer be used as parentheses" |
| 535 |
.IX Subsection "qw(...) can no longer be used as parentheses" |
| 536 |
\&\f(CW\*(C`qw\*(C'\fR lists used to fool the parser into thinking they were always |
| 537 |
surrounded by parentheses. This permitted some surprising constructions |
| 538 |
such as \f(CW\*(C`foreach $x qw(a b c) {...}\*(C'\fR, which should really be written |
| 539 |
\&\f(CW\*(C`foreach $x (qw(a b c)) {...}\*(C'\fR. These would sometimes get the lexer into |
| 540 |
the wrong state, so they didn't fully work, and the similar \f(CW\*(C`foreach qw(a |
| 541 |
b c) {...}\*(C'\fR that one might expect to be permitted never worked at all. |
| 542 |
.PP |
| 543 |
This side effect of \f(CW\*(C`qw\*(C'\fR has now been abolished. It has been deprecated |
| 544 |
since Perl v5.13.11. It is now necessary to use real parentheses |
| 545 |
everywhere that the grammar calls for them. |
| 546 |
.SS "Interaction of lexical and default warnings" |
| 547 |
.IX Subsection "Interaction of lexical and default warnings" |
| 548 |
Turning on any lexical warnings used first to disable all default warnings |
| 549 |
if lexical warnings were not already enabled: |
| 550 |
.PP |
| 551 |
.Vb 3 |
| 552 |
\& $*; # deprecation warning |
| 553 |
\& use warnings "void"; |
| 554 |
\& $#; # void warning; no deprecation warning |
| 555 |
.Ve |
| 556 |
.PP |
| 557 |
Now, the \f(CW\*(C`debugging\*(C'\fR, \f(CW\*(C`deprecated\*(C'\fR, \f(CW\*(C`glob\*(C'\fR, \f(CW\*(C`inplace\*(C'\fR and \f(CW\*(C`malloc\*(C'\fR warnings |
| 558 |
categories are left on when turning on lexical warnings (unless they are |
| 559 |
turned off by \f(CW\*(C`no warnings\*(C'\fR, of course). |
| 560 |
.PP |
| 561 |
This may cause deprecation warnings to occur in code that used to be free |
| 562 |
of warnings. |
| 563 |
.PP |
| 564 |
Those are the only categories consisting only of default warnings. Default |
| 565 |
warnings in other categories are still disabled by \f(CW\*(C`use warnings "category"\*(C'\fR, |
| 566 |
as we do not yet have the infrastructure for controlling |
| 567 |
individual warnings. |
| 568 |
.ie n .SS """state sub"" and ""our sub""" |
| 569 |
.el .SS "\f(CWstate sub\fP and \f(CWour sub\fP" |
| 570 |
.IX Subsection "state sub and our sub" |
| 571 |
Due to an accident of history, \f(CW\*(C`state sub\*(C'\fR and \f(CW\*(C`our sub\*(C'\fR were equivalent |
| 572 |
to a plain \f(CW\*(C`sub\*(C'\fR, so one could even create an anonymous sub with |
| 573 |
\&\f(CW\*(C`our sub { ... }\*(C'\fR. These are now disallowed outside of the "lexical_subs" |
| 574 |
feature. Under the "lexical_subs" feature they have new meanings described |
| 575 |
in "Lexical Subroutines" in perlsub. |
| 576 |
.SS "Defined values stored in environment are forced to byte strings" |
| 577 |
.IX Subsection "Defined values stored in environment are forced to byte strings" |
| 578 |
A value stored in an environment variable has always been stringified when |
| 579 |
inherited by child processes. |
| 580 |
.PP |
| 581 |
In this release, when assigning to \f(CW%ENV\fR, values are immediately stringified, |
| 582 |
and converted to be only a byte string. |
| 583 |
.PP |
| 584 |
First, it is forced to be only a string. Then if the string is utf8 and the |
| 585 |
equivalent of \f(CWutf8::downgrade()\fR works, that result is used; otherwise, the |
| 586 |
equivalent of \f(CWutf8::encode()\fR is used, and a warning is issued about wide |
| 587 |
characters ("Diagnostics"). |
| 588 |
.ie n .SS """require"" dies for unreadable files" |
| 589 |
.el .SS "\f(CWrequire\fP dies for unreadable files" |
| 590 |
.IX Subsection "require dies for unreadable files" |
| 591 |
When \f(CW\*(C`require\*(C'\fR encounters an unreadable file, it now dies. It used to |
| 592 |
ignore the file and continue searching the directories in \f(CW@INC\fR |
| 593 |
[perl #113422]. |
| 594 |
.ie n .SS """gv_fetchmeth_*"" and SUPER" |
| 595 |
.el .SS "\f(CWgv_fetchmeth_*\fP and SUPER" |
| 596 |
.IX Subsection "gv_fetchmeth_* and SUPER" |
| 597 |
The various \f(CW\*(C`gv_fetchmeth_*\*(C'\fR XS functions used to treat a package whose |
| 598 |
named ended with \f(CW\*(C`::SUPER\*(C'\fR specially. A method lookup on the \f(CW\*(C`Foo::SUPER\*(C'\fR |
| 599 |
package would be treated as a \f(CW\*(C`SUPER\*(C'\fR method lookup on the \f(CW\*(C`Foo\*(C'\fR package. This |
| 600 |
is no longer the case. To do a \f(CW\*(C`SUPER\*(C'\fR lookup, pass the \f(CW\*(C`Foo\*(C'\fR stash and the |
| 601 |
\&\f(CW\*(C`GV_SUPER\*(C'\fR flag. |
| 602 |
.ie n .SS """split""'s first argument is more consistently interpreted" |
| 603 |
.el .SS "\f(CWsplit\fP's first argument is more consistently interpreted" |
| 604 |
.IX Subsection "split's first argument is more consistently interpreted" |
| 605 |
After some changes earlier in v5.17, \f(CW\*(C`split\*(C'\fR's behavior has been |
| 606 |
simplified: if the PATTERN argument evaluates to a string |
| 607 |
containing one space, it is treated the way that a \fIliteral\fR string |
| 608 |
containing one space once was. |
| 609 |
.SH Deprecations |
| 610 |
.IX Header "Deprecations" |
| 611 |
.SS "Module removals" |
| 612 |
.IX Subsection "Module removals" |
| 613 |
The following modules will be removed from the core distribution in a future |
| 614 |
release, and will at that time need to be installed from CPAN. Distributions |
| 615 |
on CPAN which require these modules will need to list them as prerequisites. |
| 616 |
.PP |
| 617 |
The core versions of these modules will now issue \f(CW"deprecated"\fR\-category |
| 618 |
warnings to alert you to this fact. To silence these deprecation warnings, |
| 619 |
install the modules in question from CPAN. |
| 620 |
.PP |
| 621 |
Note that these are (with rare exceptions) fine modules that you are encouraged |
| 622 |
to continue to use. Their disinclusion from core primarily hinges on their |
| 623 |
necessity to bootstrapping a fully functional, CPAN-capable Perl installation, |
| 624 |
not usually on concerns over their design. |
| 625 |
.IP encoding 4 |
| 626 |
.IX Item "encoding" |
| 627 |
The use of this pragma is now strongly discouraged. It conflates the encoding |
| 628 |
of source text with the encoding of I/O data, reinterprets escape sequences in |
| 629 |
source text (a questionable choice), and introduces the UTF\-8 bug to all runtime |
| 630 |
handling of character strings. It is broken as designed and beyond repair. |
| 631 |
.Sp |
| 632 |
For using non-ASCII literal characters in source text, please refer to utf8. |
| 633 |
For dealing with textual I/O data, please refer to Encode and open. |
| 634 |
.IP Archive::Extract 4 |
| 635 |
.IX Item "Archive::Extract" |
| 636 |
.PD 0 |
| 637 |
.IP B::Lint 4 |
| 638 |
.IX Item "B::Lint" |
| 639 |
.IP B::Lint::Debug 4 |
| 640 |
.IX Item "B::Lint::Debug" |
| 641 |
.ie n .IP "CPANPLUS and all included ""CPANPLUS::*"" modules" 4 |
| 642 |
.el .IP "CPANPLUS and all included \f(CWCPANPLUS::*\fR modules" 4 |
| 643 |
.IX Item "CPANPLUS and all included CPANPLUS::* modules" |
| 644 |
.IP Devel::InnerPackage 4 |
| 645 |
.IX Item "Devel::InnerPackage" |
| 646 |
.IP Log::Message 4 |
| 647 |
.IX Item "Log::Message" |
| 648 |
.IP Log::Message::Config 4 |
| 649 |
.IX Item "Log::Message::Config" |
| 650 |
.IP Log::Message::Handlers 4 |
| 651 |
.IX Item "Log::Message::Handlers" |
| 652 |
.IP Log::Message::Item 4 |
| 653 |
.IX Item "Log::Message::Item" |
| 654 |
.IP Log::Message::Simple 4 |
| 655 |
.IX Item "Log::Message::Simple" |
| 656 |
.IP Module::Pluggable 4 |
| 657 |
.IX Item "Module::Pluggable" |
| 658 |
.IP Module::Pluggable::Object 4 |
| 659 |
.IX Item "Module::Pluggable::Object" |
| 660 |
.IP Object::Accessor 4 |
| 661 |
.IX Item "Object::Accessor" |
| 662 |
.IP Pod::LaTeX 4 |
| 663 |
.IX Item "Pod::LaTeX" |
| 664 |
.IP Term::UI 4 |
| 665 |
.IX Item "Term::UI" |
| 666 |
.IP Term::UI::History 4 |
| 667 |
.IX Item "Term::UI::History" |
| 668 |
.PD |
| 669 |
.SS "Deprecated Utilities" |
| 670 |
.IX Subsection "Deprecated Utilities" |
| 671 |
The following utilities will be removed from the core distribution in a |
| 672 |
future release as their associated modules have been deprecated. They |
| 673 |
will remain available with the applicable CPAN distribution. |
| 674 |
.IP cpanp 4 |
| 675 |
.IX Item "cpanp" |
| 676 |
.PD 0 |
| 677 |
.ie n .IP """cpanp\-run\-perl""" 4 |
| 678 |
.el .IP \f(CWcpanp\-run\-perl\fR 4 |
| 679 |
.IX Item "cpanp-run-perl" |
| 680 |
.IP cpan2dist 4 |
| 681 |
.IX Item "cpan2dist" |
| 682 |
.PD |
| 683 |
These items are part of the \f(CW\*(C`CPANPLUS\*(C'\fR distribution. |
| 684 |
.IP pod2latex 4 |
| 685 |
.IX Item "pod2latex" |
| 686 |
This item is part of the \f(CW\*(C`Pod::LaTeX\*(C'\fR distribution. |
| 687 |
.SS PL_sv_objcount |
| 688 |
.IX Subsection "PL_sv_objcount" |
| 689 |
This interpreter-global variable used to track the total number of |
| 690 |
Perl objects in the interpreter. It is no longer maintained and will |
| 691 |
be removed altogether in Perl v5.20. |
| 692 |
.ie n .SS "Five additional characters should be escaped in patterns with ""/x""" |
| 693 |
.el .SS "Five additional characters should be escaped in patterns with \f(CW/x\fP" |
| 694 |
.IX Subsection "Five additional characters should be escaped in patterns with /x" |
| 695 |
When a regular expression pattern is compiled with \f(CW\*(C`/x\*(C'\fR, Perl treats 6 |
| 696 |
characters as white space to ignore, such as SPACE and TAB. However, |
| 697 |
Unicode recommends 11 characters be treated thusly. We will conform |
| 698 |
with this in a future Perl version. In the meantime, use of any of the |
| 699 |
missing characters will raise a deprecation warning, unless turned off. |
| 700 |
The five characters are: |
| 701 |
.PP |
| 702 |
.Vb 5 |
| 703 |
\& U+0085 NEXT LINE |
| 704 |
\& U+200E LEFT\-TO\-RIGHT MARK |
| 705 |
\& U+200F RIGHT\-TO\-LEFT MARK |
| 706 |
\& U+2028 LINE SEPARATOR |
| 707 |
\& U+2029 PARAGRAPH SEPARATOR |
| 708 |
.Ve |
| 709 |
.SS "User-defined charnames with surprising whitespace" |
| 710 |
.IX Subsection "User-defined charnames with surprising whitespace" |
| 711 |
A user-defined character name with trailing or multiple spaces in a row is |
| 712 |
likely a typo. This now generates a warning when defined, on the assumption |
| 713 |
that uses of it will be unlikely to include the excess whitespace. |
| 714 |
.SS "Various XS-callable functions are now deprecated" |
| 715 |
.IX Subsection "Various XS-callable functions are now deprecated" |
| 716 |
All the functions used to classify characters will be removed from a |
| 717 |
future version of Perl, and should not be used. With participating C |
| 718 |
compilers (e.g., gcc), compiling any file that uses any of these will |
| 719 |
generate a warning. These were not intended for public use; there are |
| 720 |
equivalent, faster, macros for most of them. |
| 721 |
.PP |
| 722 |
See "Character classes" in perlapi. The complete list is: |
| 723 |
.PP |
| 724 |
\&\f(CW\*(C`is_uni_alnum\*(C'\fR, \f(CW\*(C`is_uni_alnumc\*(C'\fR, \f(CW\*(C`is_uni_alnumc_lc\*(C'\fR, |
| 725 |
\&\f(CW\*(C`is_uni_alnum_lc\*(C'\fR, \f(CW\*(C`is_uni_alpha\*(C'\fR, \f(CW\*(C`is_uni_alpha_lc\*(C'\fR, |
| 726 |
\&\f(CW\*(C`is_uni_ascii\*(C'\fR, \f(CW\*(C`is_uni_ascii_lc\*(C'\fR, \f(CW\*(C`is_uni_blank\*(C'\fR, |
| 727 |
\&\f(CW\*(C`is_uni_blank_lc\*(C'\fR, \f(CW\*(C`is_uni_cntrl\*(C'\fR, \f(CW\*(C`is_uni_cntrl_lc\*(C'\fR, |
| 728 |
\&\f(CW\*(C`is_uni_digit\*(C'\fR, \f(CW\*(C`is_uni_digit_lc\*(C'\fR, \f(CW\*(C`is_uni_graph\*(C'\fR, |
| 729 |
\&\f(CW\*(C`is_uni_graph_lc\*(C'\fR, \f(CW\*(C`is_uni_idfirst\*(C'\fR, \f(CW\*(C`is_uni_idfirst_lc\*(C'\fR, |
| 730 |
\&\f(CW\*(C`is_uni_lower\*(C'\fR, \f(CW\*(C`is_uni_lower_lc\*(C'\fR, \f(CW\*(C`is_uni_print\*(C'\fR, |
| 731 |
\&\f(CW\*(C`is_uni_print_lc\*(C'\fR, \f(CW\*(C`is_uni_punct\*(C'\fR, \f(CW\*(C`is_uni_punct_lc\*(C'\fR, |
| 732 |
\&\f(CW\*(C`is_uni_space\*(C'\fR, \f(CW\*(C`is_uni_space_lc\*(C'\fR, \f(CW\*(C`is_uni_upper\*(C'\fR, |
| 733 |
\&\f(CW\*(C`is_uni_upper_lc\*(C'\fR, \f(CW\*(C`is_uni_xdigit\*(C'\fR, \f(CW\*(C`is_uni_xdigit_lc\*(C'\fR, |
| 734 |
\&\f(CW\*(C`is_utf8_alnum\*(C'\fR, \f(CW\*(C`is_utf8_alnumc\*(C'\fR, \f(CW\*(C`is_utf8_alpha\*(C'\fR, |
| 735 |
\&\f(CW\*(C`is_utf8_ascii\*(C'\fR, \f(CW\*(C`is_utf8_blank\*(C'\fR, \f(CW\*(C`is_utf8_char\*(C'\fR, |
| 736 |
\&\f(CW\*(C`is_utf8_cntrl\*(C'\fR, \f(CW\*(C`is_utf8_digit\*(C'\fR, \f(CW\*(C`is_utf8_graph\*(C'\fR, |
| 737 |
\&\f(CW\*(C`is_utf8_idcont\*(C'\fR, \f(CW\*(C`is_utf8_idfirst\*(C'\fR, \f(CW\*(C`is_utf8_lower\*(C'\fR, |
| 738 |
\&\f(CW\*(C`is_utf8_mark\*(C'\fR, \f(CW\*(C`is_utf8_perl_space\*(C'\fR, \f(CW\*(C`is_utf8_perl_word\*(C'\fR, |
| 739 |
\&\f(CW\*(C`is_utf8_posix_digit\*(C'\fR, \f(CW\*(C`is_utf8_print\*(C'\fR, \f(CW\*(C`is_utf8_punct\*(C'\fR, |
| 740 |
\&\f(CW\*(C`is_utf8_space\*(C'\fR, \f(CW\*(C`is_utf8_upper\*(C'\fR, \f(CW\*(C`is_utf8_xdigit\*(C'\fR, |
| 741 |
\&\f(CW\*(C`is_utf8_xidcont\*(C'\fR, \f(CW\*(C`is_utf8_xidfirst\*(C'\fR. |
| 742 |
.PP |
| 743 |
In addition these three functions that have never worked properly are |
| 744 |
deprecated: |
| 745 |
\&\f(CW\*(C`to_uni_lower_lc\*(C'\fR, \f(CW\*(C`to_uni_title_lc\*(C'\fR, and \f(CW\*(C`to_uni_upper_lc\*(C'\fR. |
| 746 |
.SS "Certain rare uses of backslashes within regexes are now deprecated" |
| 747 |
.IX Subsection "Certain rare uses of backslashes within regexes are now deprecated" |
| 748 |
There are three pairs of characters that Perl recognizes as |
| 749 |
metacharacters in regular expression patterns: \f(CW\*(C`{}\*(C'\fR, \f(CW\*(C`[]\*(C'\fR, and \f(CW\*(C`()\*(C'\fR. |
| 750 |
These can be used as well to delimit patterns, as in: |
| 751 |
.PP |
| 752 |
.Vb 2 |
| 753 |
\& m{foo} |
| 754 |
\& s(foo)(bar) |
| 755 |
.Ve |
| 756 |
.PP |
| 757 |
Since they are metacharacters, they have special meaning to regular |
| 758 |
expression patterns, and it turns out that you can't turn off that |
| 759 |
special meaning by the normal means of preceding them with a backslash, |
| 760 |
if you use them, paired, within a pattern delimited by them. For |
| 761 |
example, in |
| 762 |
.PP |
| 763 |
.Vb 1 |
| 764 |
\& m{foo\e{1,3\e}} |
| 765 |
.Ve |
| 766 |
.PP |
| 767 |
the backslashes do not change the behavior, and this matches |
| 768 |
\&\f(CW"f\ o"\fR followed by one to three more occurrences of \f(CW"o"\fR. |
| 769 |
.PP |
| 770 |
Usages like this, where they are interpreted as metacharacters, are |
| 771 |
exceedingly rare; we think there are none, for example, in all of CPAN. |
| 772 |
Hence, this deprecation should affect very little code. It does give |
| 773 |
notice, however, that any such code needs to change, which will in turn |
| 774 |
allow us to change the behavior in future Perl versions so that the |
| 775 |
backslashes do have an effect, and without fear that we are silently |
| 776 |
breaking any existing code. |
| 777 |
.ie n .SS "Splitting the tokens ""(?"" and ""(*"" in regular expressions" |
| 778 |
.el .SS "Splitting the tokens \f(CW(?\fP and \f(CW(*\fP in regular expressions" |
| 779 |
.IX Subsection "Splitting the tokens (? and (* in regular expressions" |
| 780 |
A deprecation warning is now raised if the \f(CW\*(C`(\*(C'\fR and \f(CW\*(C`?\*(C'\fR are separated |
| 781 |
by white space or comments in \f(CW\*(C`(?...)\*(C'\fR regular expression constructs. |
| 782 |
Similarly, if the \f(CW\*(C`(\*(C'\fR and \f(CW\*(C`*\*(C'\fR are separated in \f(CW\*(C`(*VERB...)\*(C'\fR |
| 783 |
constructs. |
| 784 |
.SS "Pre-PerlIO IO implementations" |
| 785 |
.IX Subsection "Pre-PerlIO IO implementations" |
| 786 |
In theory, you can currently build perl without PerlIO. Instead, you'd use a |
| 787 |
wrapper around stdio or sfio. In practice, this isn't very useful. It's not |
| 788 |
well tested, and without any support for IO layers or (thus) Unicode, it's not |
| 789 |
much of a perl. Building without PerlIO will most likely be removed in the |
| 790 |
next version of perl. |
| 791 |
.PP |
| 792 |
PerlIO supports a \f(CW\*(C`stdio\*(C'\fR layer if stdio use is desired. Similarly a |
| 793 |
sfio layer could be produced in the future, if needed. |
| 794 |
.SH "Future Deprecations" |
| 795 |
.IX Header "Future Deprecations" |
| 796 |
.IP \(bu 4 |
| 797 |
Platforms without support infrastructure |
| 798 |
.Sp |
| 799 |
Both Windows CE and z/OS have been historically under-maintained, and are |
| 800 |
currently neither successfully building nor regularly being smoke tested. |
| 801 |
Efforts are underway to change this situation, but it should not be taken for |
| 802 |
granted that the platforms are safe and supported. If they do not become |
| 803 |
buildable and regularly smoked, support for them may be actively removed in |
| 804 |
future releases. If you have an interest in these platforms and you can lend |
| 805 |
your time, expertise, or hardware to help support these platforms, please let |
| 806 |
the perl development effort know by emailing \f(CW\*(C`perl5\[email protected]\*(C'\fR. |
| 807 |
.Sp |
| 808 |
Some platforms that appear otherwise entirely dead are also on the short list |
| 809 |
for removal between now and v5.20.0: |
| 810 |
.RS 4 |
| 811 |
.IP DG/UX 4 |
| 812 |
.IX Item "DG/UX" |
| 813 |
.PD 0 |
| 814 |
.IP NeXT 4 |
| 815 |
.IX Item "NeXT" |
| 816 |
.RE |
| 817 |
.RS 4 |
| 818 |
.PD |
| 819 |
.Sp |
| 820 |
We also think it likely that current versions of Perl will no longer |
| 821 |
build AmigaOS, DJGPP, NetWare (natively), OS/2 and Plan 9. If you |
| 822 |
are using Perl on such a platform and have an interest in ensuring |
| 823 |
Perl's future on them, please contact us. |
| 824 |
.Sp |
| 825 |
We believe that Perl has long been unable to build on mixed endian |
| 826 |
architectures (such as PDP\-11s), and intend to remove any remaining |
| 827 |
support code. Similarly, code supporting the long unmaintained GNU |
| 828 |
dld will be removed soon if no-one makes themselves known as an |
| 829 |
active user. |
| 830 |
.RE |
| 831 |
.IP \(bu 4 |
| 832 |
Swapping of $< and $> |
| 833 |
.Sp |
| 834 |
Perl has supported the idiom of swapping $< and $> (and likewise $( and |
| 835 |
$)) to temporarily drop permissions since 5.0, like this: |
| 836 |
.Sp |
| 837 |
.Vb 1 |
| 838 |
\& ($<, $>) = ($>, $<); |
| 839 |
.Ve |
| 840 |
.Sp |
| 841 |
However, this idiom modifies the real user/group id, which can have |
| 842 |
undesirable side-effects, is no longer useful on any platform perl |
| 843 |
supports and complicates the implementation of these variables and list |
| 844 |
assignment in general. |
| 845 |
.Sp |
| 846 |
As an alternative, assignment only to \f(CW$>\fR is recommended: |
| 847 |
.Sp |
| 848 |
.Vb 1 |
| 849 |
\& local $> = $<; |
| 850 |
.Ve |
| 851 |
.Sp |
| 852 |
See also: Setuid Demystified <http://www.cs.berkeley.edu/~daw/papers/setuid-usenix02.pdf>. |
| 853 |
.IP \(bu 4 |
| 854 |
\&\f(CW\*(C`microperl\*(C'\fR, long broken and of unclear present purpose, will be removed. |
| 855 |
.IP \(bu 4 |
| 856 |
Revamping \f(CW"\eQ"\fR semantics in double-quotish strings when combined with |
| 857 |
other escapes. |
| 858 |
.Sp |
| 859 |
There are several bugs and inconsistencies involving combinations |
| 860 |
of \f(CW\*(C`\eQ\*(C'\fR and escapes like \f(CW\*(C`\ex\*(C'\fR, \f(CW\*(C`\eL\*(C'\fR, etc., within a \f(CW\*(C`\eQ...\eE\*(C'\fR pair. |
| 861 |
These need to be fixed, and doing so will necessarily change current |
| 862 |
behavior. The changes have not yet been settled. |
| 863 |
.IP \(bu 4 |
| 864 |
Use of \f(CW$x\fR, where \f(CW\*(C`x\*(C'\fR stands for any actual (non-printing) C0 control |
| 865 |
character will be disallowed in a future Perl version. Use \f(CW\*(C`${x}\*(C'\fR |
| 866 |
instead (where again \f(CW\*(C`x\*(C'\fR stands for a control character), |
| 867 |
or better, \f(CW$^A\fR , where \f(CW\*(C`^\*(C'\fR is a caret (CIRCUMFLEX ACCENT), |
| 868 |
and \f(CW\*(C`A\*(C'\fR stands for any of the characters listed at the end of |
| 869 |
"OPERATOR DIFFERENCES" in perlebcdic. |
| 870 |
.SH "Performance Enhancements" |
| 871 |
.IX Header "Performance Enhancements" |
| 872 |
.IP \(bu 4 |
| 873 |
Lists of lexical variable declarations (\f(CW\*(C`my($x, $y)\*(C'\fR) are now optimised |
| 874 |
down to a single op and are hence faster than before. |
| 875 |
.IP \(bu 4 |
| 876 |
A new C preprocessor define \f(CW\*(C`NO_TAINT_SUPPORT\*(C'\fR was added that, if set, |
| 877 |
disables Perl's taint support altogether. Using the \-T or \-t command |
| 878 |
line flags will cause a fatal error. Beware that both core tests as |
| 879 |
well as many a CPAN distribution's tests will fail with this change. On |
| 880 |
the upside, it provides a small performance benefit due to reduced |
| 881 |
branching. |
| 882 |
.Sp |
| 883 |
\&\fBDo not enable this unless you know exactly what you are getting yourself |
| 884 |
into.\fR |
| 885 |
.IP \(bu 4 |
| 886 |
\&\f(CW\*(C`pack\*(C'\fR with constant arguments is now constant folded in most cases |
| 887 |
[perl #113470]. |
| 888 |
.IP \(bu 4 |
| 889 |
Speed up in regular expression matching against Unicode properties. The |
| 890 |
largest gain is for \f(CW\*(C`\eX\*(C'\fR, the Unicode "extended grapheme cluster." The |
| 891 |
gain for it is about 35% \- 40%. Bracketed character classes, e.g., |
| 892 |
\&\f(CW\*(C`[0\-9\ex{100}]\*(C'\fR containing code points above 255 are also now faster. |
| 893 |
.IP \(bu 4 |
| 894 |
On platforms supporting it, several former macros are now implemented as static |
| 895 |
inline functions. This should speed things up slightly on non-GCC platforms. |
| 896 |
.IP \(bu 4 |
| 897 |
The optimisation of hashes in boolean context has been extended to |
| 898 |
affect \f(CWscalar(%hash)\fR, \f(CW\*(C`%hash ? ... : ...\*(C'\fR, and \f(CW\*(C`sub { %hash || ... }\*(C'\fR. |
| 899 |
.IP \(bu 4 |
| 900 |
Filetest operators manage the stack in a fractionally more efficient manner. |
| 901 |
.IP \(bu 4 |
| 902 |
Globs used in a numeric context are now numified directly in most cases, |
| 903 |
rather than being numified via stringification. |
| 904 |
.IP \(bu 4 |
| 905 |
The \f(CW\*(C`x\*(C'\fR repetition operator is now folded to a single constant at compile |
| 906 |
time if called in scalar context with constant operands and no parentheses |
| 907 |
around the left operand. |
| 908 |
.SH "Modules and Pragmata" |
| 909 |
.IX Header "Modules and Pragmata" |
| 910 |
.SS "New Modules and Pragmata" |
| 911 |
.IX Subsection "New Modules and Pragmata" |
| 912 |
.IP \(bu 4 |
| 913 |
Config::Perl::V version 0.16 has been added as a dual-lifed module. |
| 914 |
It provides structured data retrieval of \f(CW\*(C`perl \-V\*(C'\fR output including |
| 915 |
information only known to the \f(CW\*(C`perl\*(C'\fR binary and not available via Config. |
| 916 |
.SS "Updated Modules and Pragmata" |
| 917 |
.IX Subsection "Updated Modules and Pragmata" |
| 918 |
For a complete list of updates, run: |
| 919 |
.PP |
| 920 |
.Vb 1 |
| 921 |
\& $ corelist \-\-diff 5.16.0 5.18.0 |
| 922 |
.Ve |
| 923 |
.PP |
| 924 |
You can substitute your favorite version in place of \f(CW5.16.0\fR, too. |
| 925 |
.IP \(bu 4 |
| 926 |
Archive::Extract has been upgraded to 0.68. |
| 927 |
.Sp |
| 928 |
Work around an edge case on Linux with Busybox's unzip. |
| 929 |
.IP \(bu 4 |
| 930 |
Archive::Tar has been upgraded to 1.90. |
| 931 |
.Sp |
| 932 |
ptar now supports the \-T option as well as dashless options |
| 933 |
[rt.cpan.org #75473], [rt.cpan.org #75475]. |
| 934 |
.Sp |
| 935 |
Auto-encode filenames marked as UTF\-8 [rt.cpan.org #75474]. |
| 936 |
.Sp |
| 937 |
Don't use \f(CW\*(C`tell\*(C'\fR on IO::Zlib handles [rt.cpan.org #64339]. |
| 938 |
.Sp |
| 939 |
Don't try to \f(CW\*(C`chown\*(C'\fR on symlinks. |
| 940 |
.IP \(bu 4 |
| 941 |
autodie has been upgraded to 2.13. |
| 942 |
.Sp |
| 943 |
\&\f(CW\*(C`autodie\*(C'\fR now plays nicely with the 'open' pragma. |
| 944 |
.IP \(bu 4 |
| 945 |
B has been upgraded to 1.42. |
| 946 |
.Sp |
| 947 |
The \f(CW\*(C`stashoff\*(C'\fR method of COPs has been added. This provides access to an |
| 948 |
internal field added in perl 5.16 under threaded builds [perl #113034]. |
| 949 |
.Sp |
| 950 |
\&\f(CW\*(C`B::COP::stashpv\*(C'\fR now supports UTF\-8 package names and embedded NULs. |
| 951 |
.Sp |
| 952 |
All \f(CW\*(C`CVf_*\*(C'\fR and \f(CW\*(C`GVf_*\*(C'\fR |
| 953 |
and more SV-related flag values are now provided as constants in the \f(CW\*(C`B::\*(C'\fR |
| 954 |
namespace and available for export. The default export list has not changed. |
| 955 |
.Sp |
| 956 |
This makes the module work with the new pad API. |
| 957 |
.IP \(bu 4 |
| 958 |
B::Concise has been upgraded to 0.95. |
| 959 |
.Sp |
| 960 |
The \f(CW\*(C`\-nobanner\*(C'\fR option has been fixed, and \f(CW\*(C`format\*(C'\fRs can now be dumped. |
| 961 |
When passed a sub name to dump, it will check also to see whether it |
| 962 |
is the name of a format. If a sub and a format share the same name, |
| 963 |
it will dump both. |
| 964 |
.Sp |
| 965 |
This adds support for the new \f(CW\*(C`OpMAYBE_TRUEBOOL\*(C'\fR and \f(CW\*(C`OPpTRUEBOOL\*(C'\fR flags. |
| 966 |
.IP \(bu 4 |
| 967 |
B::Debug has been upgraded to 1.18. |
| 968 |
.Sp |
| 969 |
This adds support (experimentally) for \f(CW\*(C`B::PADLIST\*(C'\fR, which was |
| 970 |
added in Perl 5.17.4. |
| 971 |
.IP \(bu 4 |
| 972 |
B::Deparse has been upgraded to 1.20. |
| 973 |
.Sp |
| 974 |
Avoid warning when run under \f(CW\*(C`perl \-w\*(C'\fR. |
| 975 |
.Sp |
| 976 |
It now deparses |
| 977 |
loop controls with the correct precedence, and multiple statements in a |
| 978 |
\&\f(CW\*(C`format\*(C'\fR line are also now deparsed correctly. |
| 979 |
.Sp |
| 980 |
This release suppresses trailing semicolons in formats. |
| 981 |
.Sp |
| 982 |
This release adds stub deparsing for lexical subroutines. |
| 983 |
.Sp |
| 984 |
It no longer dies when deparsing \f(CW\*(C`sort\*(C'\fR without arguments. It now |
| 985 |
correctly omits the comma for \f(CW\*(C`system $prog @args\*(C'\fR and \f(CWexec $prog |
| 986 |
@args\fR. |
| 987 |
.IP \(bu 4 |
| 988 |
bignum, bigint and bigrat have been upgraded to 0.33. |
| 989 |
.Sp |
| 990 |
The overrides for \f(CW\*(C`hex\*(C'\fR and \f(CW\*(C`oct\*(C'\fR have been rewritten, eliminating |
| 991 |
several problems, and making one incompatible change: |
| 992 |
.RS 4 |
| 993 |
.IP \(bu 4 |
| 994 |
Formerly, whichever of \f(CW\*(C`use bigint\*(C'\fR or \f(CW\*(C`use bigrat\*(C'\fR was compiled later |
| 995 |
would take precedence over the other, causing \f(CW\*(C`hex\*(C'\fR and \f(CW\*(C`oct\*(C'\fR not to |
| 996 |
respect the other pragma when in scope. |
| 997 |
.IP \(bu 4 |
| 998 |
Using any of these three pragmata would cause \f(CW\*(C`hex\*(C'\fR and \f(CW\*(C`oct\*(C'\fR anywhere |
| 999 |
else in the program to evaluate their arguments in list context and prevent |
| 1000 |
them from inferring \f(CW$_\fR when called without arguments. |
| 1001 |
.IP \(bu 4 |
| 1002 |
Using any of these three pragmata would make \f(CWoct("1234")\fR return 1234 |
| 1003 |
(for any number not beginning with 0) anywhere in the program. Now "1234" |
| 1004 |
is translated from octal to decimal, whether within the pragma's scope or |
| 1005 |
not. |
| 1006 |
.IP \(bu 4 |
| 1007 |
The global overrides that facilitate lexical use of \f(CW\*(C`hex\*(C'\fR and \f(CW\*(C`oct\*(C'\fR now |
| 1008 |
respect any existing overrides that were in place before the new overrides |
| 1009 |
were installed, falling back to them outside of the scope of \f(CW\*(C`use bignum\*(C'\fR. |
| 1010 |
.IP \(bu 4 |
| 1011 |
\&\f(CW\*(C`use bignum "hex"\*(C'\fR, \f(CW\*(C`use bignum "oct"\*(C'\fR and similar invocations for bigint |
| 1012 |
and bigrat now export a \f(CW\*(C`hex\*(C'\fR or \f(CW\*(C`oct\*(C'\fR function, instead of providing a |
| 1013 |
global override. |
| 1014 |
.RE |
| 1015 |
.RS 4 |
| 1016 |
.RE |
| 1017 |
.IP \(bu 4 |
| 1018 |
Carp has been upgraded to 1.29. |
| 1019 |
.Sp |
| 1020 |
Carp is no longer confused when \f(CW\*(C`caller\*(C'\fR returns undef for a package that |
| 1021 |
has been deleted. |
| 1022 |
.Sp |
| 1023 |
The \f(CWlongmess()\fR and \f(CWshortmess()\fR functions are now documented. |
| 1024 |
.IP \(bu 4 |
| 1025 |
CGI has been upgraded to 3.63. |
| 1026 |
.Sp |
| 1027 |
Unrecognized HTML escape sequences are now handled better, problematic |
| 1028 |
trailing newlines are no longer inserted after <form> tags |
| 1029 |
by \f(CWstartform()\fR or \f(CWstart_form()\fR, and bogus "Insecure Dependency" |
| 1030 |
warnings appearing with some versions of perl are now worked around. |
| 1031 |
.IP \(bu 4 |
| 1032 |
Class::Struct has been upgraded to 0.64. |
| 1033 |
.Sp |
| 1034 |
The constructor now respects overridden accessor methods [perl #29230]. |
| 1035 |
.IP \(bu 4 |
| 1036 |
Compress::Raw::Bzip2 has been upgraded to 2.060. |
| 1037 |
.Sp |
| 1038 |
The misuse of Perl's "magic" API has been fixed. |
| 1039 |
.IP \(bu 4 |
| 1040 |
Compress::Raw::Zlib has been upgraded to 2.060. |
| 1041 |
.Sp |
| 1042 |
Upgrade bundled zlib to version 1.2.7. |
| 1043 |
.Sp |
| 1044 |
Fix build failures on Irix, Solaris, and Win32, and also when building as C++ |
| 1045 |
[rt.cpan.org #69985], [rt.cpan.org #77030], [rt.cpan.org #75222]. |
| 1046 |
.Sp |
| 1047 |
The misuse of Perl's "magic" API has been fixed. |
| 1048 |
.Sp |
| 1049 |
\&\f(CWcompress()\fR, \f(CWuncompress()\fR, \f(CWmemGzip()\fR and \f(CWmemGunzip()\fR have |
| 1050 |
been speeded up by making parameter validation more efficient. |
| 1051 |
.IP \(bu 4 |
| 1052 |
CPAN::Meta::Requirements has been upgraded to 2.122. |
| 1053 |
.Sp |
| 1054 |
Treat undef requirements to \f(CW\*(C`from_string_hash\*(C'\fR as 0 (with a warning). |
| 1055 |
.Sp |
| 1056 |
Added \f(CW\*(C`requirements_for_module\*(C'\fR method. |
| 1057 |
.IP \(bu 4 |
| 1058 |
CPANPLUS has been upgraded to 0.9135. |
| 1059 |
.Sp |
| 1060 |
Allow adding \fIblib/script\fR to PATH. |
| 1061 |
.Sp |
| 1062 |
Save the history between invocations of the shell. |
| 1063 |
.Sp |
| 1064 |
Handle multiple \f(CW\*(C`makemakerargs\*(C'\fR and \f(CW\*(C`makeflags\*(C'\fR arguments better. |
| 1065 |
.Sp |
| 1066 |
This resolves issues with the SQLite source engine. |
| 1067 |
.IP \(bu 4 |
| 1068 |
Data::Dumper has been upgraded to 2.145. |
| 1069 |
.Sp |
| 1070 |
It has been optimized to only build a seen-scalar hash as necessary, |
| 1071 |
thereby speeding up serialization drastically. |
| 1072 |
.Sp |
| 1073 |
Additional tests were added in order to improve statement, branch, condition |
| 1074 |
and subroutine coverage. On the basis of the coverage analysis, some of the |
| 1075 |
internals of Dumper.pm were refactored. Almost all methods are now |
| 1076 |
documented. |
| 1077 |
.IP \(bu 4 |
| 1078 |
DB_File has been upgraded to 1.827. |
| 1079 |
.Sp |
| 1080 |
The main Perl module no longer uses the \f(CW"@_"\fR construct. |
| 1081 |
.IP \(bu 4 |
| 1082 |
Devel::Peek has been upgraded to 1.11. |
| 1083 |
.Sp |
| 1084 |
This fixes compilation with C++ compilers and makes the module work with |
| 1085 |
the new pad API. |
| 1086 |
.IP \(bu 4 |
| 1087 |
Digest::MD5 has been upgraded to 2.52. |
| 1088 |
.Sp |
| 1089 |
Fix \f(CW\*(C`Digest::Perl::MD5\*(C'\fR OO fallback [rt.cpan.org #66634]. |
| 1090 |
.IP \(bu 4 |
| 1091 |
Digest::SHA has been upgraded to 5.84. |
| 1092 |
.Sp |
| 1093 |
This fixes a double-free bug, which might have caused vulnerabilities |
| 1094 |
in some cases. |
| 1095 |
.IP \(bu 4 |
| 1096 |
DynaLoader has been upgraded to 1.18. |
| 1097 |
.Sp |
| 1098 |
This is due to a minor code change in the XS for the VMS implementation. |
| 1099 |
.Sp |
| 1100 |
This fixes warnings about using \f(CW\*(C`CODE\*(C'\fR sections without an \f(CW\*(C`OUTPUT\*(C'\fR |
| 1101 |
section. |
| 1102 |
.IP \(bu 4 |
| 1103 |
Encode has been upgraded to 2.49. |
| 1104 |
.Sp |
| 1105 |
The Mac alias x\-mac-ce has been added, and various bugs have been fixed |
| 1106 |
in Encode::Unicode, Encode::UTF7 and Encode::GSM0338. |
| 1107 |
.IP \(bu 4 |
| 1108 |
Env has been upgraded to 1.04. |
| 1109 |
.Sp |
| 1110 |
Its SPLICE implementation no longer misbehaves in list context. |
| 1111 |
.IP \(bu 4 |
| 1112 |
ExtUtils::CBuilder has been upgraded to 0.280210. |
| 1113 |
.Sp |
| 1114 |
Manifest files are now correctly embedded for those versions of VC++ which |
| 1115 |
make use of them. [perl #111782, #111798]. |
| 1116 |
.Sp |
| 1117 |
A list of symbols to export can now be passed to \f(CWlink()\fR when on |
| 1118 |
Windows, as on other OSes [perl #115100]. |
| 1119 |
.IP \(bu 4 |
| 1120 |
ExtUtils::ParseXS has been upgraded to 3.18. |
| 1121 |
.Sp |
| 1122 |
The generated C code now avoids unnecessarily incrementing |
| 1123 |
\&\f(CW\*(C`PL_amagic_generation\*(C'\fR on Perl versions where it's done automatically |
| 1124 |
(or on current Perl where the variable no longer exists). |
| 1125 |
.Sp |
| 1126 |
This avoids a bogus warning for initialised XSUB non-parameters [perl |
| 1127 |
#112776]. |
| 1128 |
.IP \(bu 4 |
| 1129 |
File::Copy has been upgraded to 2.26. |
| 1130 |
.Sp |
| 1131 |
\&\f(CWcopy()\fR no longer zeros files when copying into the same directory, |
| 1132 |
and also now fails (as it has long been documented to do) when attempting |
| 1133 |
to copy a file over itself. |
| 1134 |
.IP \(bu 4 |
| 1135 |
File::DosGlob has been upgraded to 1.10. |
| 1136 |
.Sp |
| 1137 |
The internal cache of file names that it keeps for each caller is now |
| 1138 |
freed when that caller is freed. This means |
| 1139 |
\&\f(CW\*(C`use File::DosGlob \*(Aqglob\*(Aq; eval \*(Aqscalar <*>\*(Aq\*(C'\fR no longer leaks memory. |
| 1140 |
.IP \(bu 4 |
| 1141 |
File::Fetch has been upgraded to 0.38. |
| 1142 |
.Sp |
| 1143 |
Added the 'file_default' option for URLs that do not have a file |
| 1144 |
component. |
| 1145 |
.Sp |
| 1146 |
Use \f(CW\*(C`File::HomeDir\*(C'\fR when available, and provide \f(CW\*(C`PERL5_CPANPLUS_HOME\*(C'\fR to |
| 1147 |
override the autodetection. |
| 1148 |
.Sp |
| 1149 |
Always re-fetch \fICHECKSUMS\fR if \f(CW\*(C`fetchdir\*(C'\fR is set. |
| 1150 |
.IP \(bu 4 |
| 1151 |
File::Find has been upgraded to 1.23. |
| 1152 |
.Sp |
| 1153 |
This fixes inconsistent unixy path handling on VMS. |
| 1154 |
.Sp |
| 1155 |
Individual files may now appear in list of directories to be searched |
| 1156 |
[perl #59750]. |
| 1157 |
.IP \(bu 4 |
| 1158 |
File::Glob has been upgraded to 1.20. |
| 1159 |
.Sp |
| 1160 |
File::Glob has had exactly the same fix as File::DosGlob. Since it is |
| 1161 |
what Perl's own \f(CW\*(C`glob\*(C'\fR operator itself uses (except on VMS), this means |
| 1162 |
\&\f(CW\*(C`eval \*(Aqscalar <*>\*(Aq\*(C'\fR no longer leaks. |
| 1163 |
.Sp |
| 1164 |
A space-separated list of patterns return long lists of results no longer |
| 1165 |
results in memory corruption or crashes. This bug was introduced in |
| 1166 |
Perl 5.16.0. [perl #114984] |
| 1167 |
.IP \(bu 4 |
| 1168 |
File::Spec::Unix has been upgraded to 3.40. |
| 1169 |
.Sp |
| 1170 |
\&\f(CW\*(C`abs2rel\*(C'\fR could produce incorrect results when given two relative paths or |
| 1171 |
the root directory twice [perl #111510]. |
| 1172 |
.IP \(bu 4 |
| 1173 |
File::stat has been upgraded to 1.07. |
| 1174 |
.Sp |
| 1175 |
\&\f(CW\*(C`File::stat\*(C'\fR ignores the filetest pragma, and warns when used in |
| 1176 |
combination therewith. But it was not warning for \f(CW\*(C`\-r\*(C'\fR. This has been |
| 1177 |
fixed [perl #111640]. |
| 1178 |
.Sp |
| 1179 |
\&\f(CW\*(C`\-p\*(C'\fR now works, and does not return false for pipes [perl #111638]. |
| 1180 |
.Sp |
| 1181 |
Previously \f(CW\*(C`File::stat\*(C'\fR's overloaded \f(CW\*(C`\-x\*(C'\fR and \f(CW\*(C`\-X\*(C'\fR operators did not give |
| 1182 |
the correct results for directories or executable files when running as |
| 1183 |
root. They had been treating executable permissions for root just like for |
| 1184 |
any other user, performing group membership tests \fIetc\fR for files not owned |
| 1185 |
by root. They now follow the correct Unix behaviour \- for a directory they |
| 1186 |
are always true, and for a file if any of the three execute permission bits |
| 1187 |
are set then they report that root can execute the file. Perl's builtin |
| 1188 |
\&\f(CW\*(C`\-x\*(C'\fR and \f(CW\*(C`\-X\*(C'\fR operators have always been correct. |
| 1189 |
.IP \(bu 4 |
| 1190 |
File::Temp has been upgraded to 0.23 |
| 1191 |
.Sp |
| 1192 |
Fixes various bugs involving directory removal. Defers unlinking tempfiles if |
| 1193 |
the initial unlink fails, which fixes problems on NFS. |
| 1194 |
.IP \(bu 4 |
| 1195 |
GDBM_File has been upgraded to 1.15. |
| 1196 |
.Sp |
| 1197 |
The undocumented optional fifth parameter to \f(CW\*(C`TIEHASH\*(C'\fR has been |
| 1198 |
removed. This was intended to provide control of the callback used by |
| 1199 |
\&\f(CW\*(C`gdbm*\*(C'\fR functions in case of fatal errors (such as filesystem problems), |
| 1200 |
but did not work (and could never have worked). No code on CPAN even |
| 1201 |
attempted to use it. The callback is now always the previous default, |
| 1202 |
\&\f(CW\*(C`croak\*(C'\fR. Problems on some platforms with how the \f(CW\*(C`C\*(C'\fR \f(CW\*(C`croak\*(C'\fR function |
| 1203 |
is called have also been resolved. |
| 1204 |
.IP \(bu 4 |
| 1205 |
Hash::Util has been upgraded to 0.15. |
| 1206 |
.Sp |
| 1207 |
\&\f(CW\*(C`hash_unlocked\*(C'\fR and \f(CW\*(C`hashref_unlocked\*(C'\fR now returns true if the hash is |
| 1208 |
unlocked, instead of always returning false [perl #112126]. |
| 1209 |
.Sp |
| 1210 |
\&\f(CW\*(C`hash_unlocked\*(C'\fR, \f(CW\*(C`hashref_unlocked\*(C'\fR, \f(CW\*(C`lock_hash_recurse\*(C'\fR and |
| 1211 |
\&\f(CW\*(C`unlock_hash_recurse\*(C'\fR are now exportable [perl #112126]. |
| 1212 |
.Sp |
| 1213 |
Two new functions, \f(CW\*(C`hash_locked\*(C'\fR and \f(CW\*(C`hashref_locked\*(C'\fR, have been added. |
| 1214 |
Oddly enough, these two functions were already exported, even though they |
| 1215 |
did not exist [perl #112126]. |
| 1216 |
.IP \(bu 4 |
| 1217 |
HTTP::Tiny has been upgraded to 0.025. |
| 1218 |
.Sp |
| 1219 |
Add SSL verification features [github #6], [github #9]. |
| 1220 |
.Sp |
| 1221 |
Include the final URL in the response hashref. |
| 1222 |
.Sp |
| 1223 |
Add \f(CW\*(C`local_address\*(C'\fR option. |
| 1224 |
.Sp |
| 1225 |
This improves SSL support. |
| 1226 |
.IP \(bu 4 |
| 1227 |
IO has been upgraded to 1.28. |
| 1228 |
.Sp |
| 1229 |
\&\f(CWsync()\fR can now be called on read-only file handles [perl #64772]. |
| 1230 |
.Sp |
| 1231 |
IO::Socket tries harder to cache or otherwise fetch socket |
| 1232 |
information. |
| 1233 |
.IP \(bu 4 |
| 1234 |
IPC::Cmd has been upgraded to 0.80. |
| 1235 |
.Sp |
| 1236 |
Use \f(CW\*(C`POSIX::_exit\*(C'\fR instead of \f(CW\*(C`exit\*(C'\fR in \f(CW\*(C`run_forked\*(C'\fR [rt.cpan.org #76901]. |
| 1237 |
.IP \(bu 4 |
| 1238 |
IPC::Open3 has been upgraded to 1.13. |
| 1239 |
.Sp |
| 1240 |
The \f(CWopen3()\fR function no longer uses \f(CWPOSIX::close()\fR to close file |
| 1241 |
descriptors since that breaks the ref-counting of file descriptors done by |
| 1242 |
PerlIO in cases where the file descriptors are shared by PerlIO streams, |
| 1243 |
leading to attempts to close the file descriptors a second time when |
| 1244 |
any such PerlIO streams are closed later on. |
| 1245 |
.IP \(bu 4 |
| 1246 |
Locale::Codes has been upgraded to 3.25. |
| 1247 |
.Sp |
| 1248 |
It includes some new codes. |
| 1249 |
.IP \(bu 4 |
| 1250 |
Memoize has been upgraded to 1.03. |
| 1251 |
.Sp |
| 1252 |
Fix the \f(CW\*(C`MERGE\*(C'\fR cache option. |
| 1253 |
.IP \(bu 4 |
| 1254 |
Module::Build has been upgraded to 0.4003. |
| 1255 |
.Sp |
| 1256 |
Fixed bug where modules without \f(CW$VERSION\fR might have a version of '0' listed |
| 1257 |
in 'provides' metadata, which will be rejected by PAUSE. |
| 1258 |
.Sp |
| 1259 |
Fixed bug in PodParser to allow numerals in module names. |
| 1260 |
.Sp |
| 1261 |
Fixed bug where giving arguments twice led to them becoming arrays, resulting |
| 1262 |
in install paths like \fR\f(BIARRAY\fR\fI\|(0xdeadbeef)/lib/Foo.pm\fR. |
| 1263 |
.Sp |
| 1264 |
A minor bug fix allows markup to be used around the leading "Name" in |
| 1265 |
a POD "abstract" line, and some documentation improvements have been made. |
| 1266 |
.IP \(bu 4 |
| 1267 |
Module::CoreList has been upgraded to 2.90 |
| 1268 |
.Sp |
| 1269 |
Version information is now stored as a delta, which greatly reduces the |
| 1270 |
size of the \fICoreList.pm\fR file. |
| 1271 |
.Sp |
| 1272 |
This restores compatibility with older versions of perl and cleans up |
| 1273 |
the corelist data for various modules. |
| 1274 |
.IP \(bu 4 |
| 1275 |
Module::Load::Conditional has been upgraded to 0.54. |
| 1276 |
.Sp |
| 1277 |
Fix use of \f(CW\*(C`requires\*(C'\fR on perls installed to a path with spaces. |
| 1278 |
.Sp |
| 1279 |
Various enhancements include the new use of Module::Metadata. |
| 1280 |
.IP \(bu 4 |
| 1281 |
Module::Metadata has been upgraded to 1.000011. |
| 1282 |
.Sp |
| 1283 |
The creation of a Module::Metadata object for a typical module file has |
| 1284 |
been sped up by about 40%, and some spurious warnings about \f(CW$VERSION\fRs |
| 1285 |
have been suppressed. |
| 1286 |
.IP \(bu 4 |
| 1287 |
Module::Pluggable has been upgraded to 4.7. |
| 1288 |
.Sp |
| 1289 |
Amongst other changes, triggers are now allowed on events, which gives |
| 1290 |
a powerful way to modify behaviour. |
| 1291 |
.IP \(bu 4 |
| 1292 |
Net::Ping has been upgraded to 2.41. |
| 1293 |
.Sp |
| 1294 |
This fixes some test failures on Windows. |
| 1295 |
.IP \(bu 4 |
| 1296 |
Opcode has been upgraded to 1.25. |
| 1297 |
.Sp |
| 1298 |
Reflect the removal of the boolkeys opcode and the addition of the |
| 1299 |
clonecv, introcv and padcv opcodes. |
| 1300 |
.IP \(bu 4 |
| 1301 |
overload has been upgraded to 1.22. |
| 1302 |
.Sp |
| 1303 |
\&\f(CW\*(C`no overload\*(C'\fR now warns for invalid arguments, just like \f(CW\*(C`use overload\*(C'\fR. |
| 1304 |
.IP \(bu 4 |
| 1305 |
PerlIO::encoding has been upgraded to 0.16. |
| 1306 |
.Sp |
| 1307 |
This is the module implementing the ":encoding(...)" I/O layer. It no |
| 1308 |
longer corrupts memory or crashes when the encoding back-end reallocates |
| 1309 |
the buffer or gives it a typeglob or shared hash key scalar. |
| 1310 |
.IP \(bu 4 |
| 1311 |
PerlIO::scalar has been upgraded to 0.16. |
| 1312 |
.Sp |
| 1313 |
The buffer scalar supplied may now only contain code points 0xFF or |
| 1314 |
lower. [perl #109828] |
| 1315 |
.IP \(bu 4 |
| 1316 |
Perl::OSType has been upgraded to 1.003. |
| 1317 |
.Sp |
| 1318 |
This fixes a bug detecting the VOS operating system. |
| 1319 |
.IP \(bu 4 |
| 1320 |
Pod::Html has been upgraded to 1.18. |
| 1321 |
.Sp |
| 1322 |
The option \f(CW\*(C`\-\-libpods\*(C'\fR has been reinstated. It is deprecated, and its use |
| 1323 |
does nothing other than issue a warning that it is no longer supported. |
| 1324 |
.Sp |
| 1325 |
Since the HTML files generated by pod2html claim to have a UTF\-8 charset, |
| 1326 |
actually write the files out using UTF\-8 [perl #111446]. |
| 1327 |
.IP \(bu 4 |
| 1328 |
Pod::Simple has been upgraded to 3.28. |
| 1329 |
.Sp |
| 1330 |
Numerous improvements have been made, mostly to Pod::Simple::XHTML, |
| 1331 |
which also has a compatibility change: the \f(CW\*(C`codes_in_verbatim\*(C'\fR option |
| 1332 |
is now disabled by default. See \fIcpan/Pod\-Simple/ChangeLog\fR for the |
| 1333 |
full details. |
| 1334 |
.IP \(bu 4 |
| 1335 |
re has been upgraded to 0.23 |
| 1336 |
.Sp |
| 1337 |
Single character [class]es like \f(CW\*(C`/[s]/\*(C'\fR or \f(CW\*(C`/[s]/i\*(C'\fR are now optimized |
| 1338 |
as if they did not have the brackets, i.e. \f(CW\*(C`/s/\*(C'\fR or \f(CW\*(C`/s/i\*(C'\fR. |
| 1339 |
.Sp |
| 1340 |
See note about \f(CW\*(C`op_comp\*(C'\fR in the "Internal Changes" section below. |
| 1341 |
.IP \(bu 4 |
| 1342 |
Safe has been upgraded to 2.35. |
| 1343 |
.Sp |
| 1344 |
Fix interactions with \f(CW\*(C`Devel::Cover\*(C'\fR. |
| 1345 |
.Sp |
| 1346 |
Don't eval code under \f(CW\*(C`no strict\*(C'\fR. |
| 1347 |
.IP \(bu 4 |
| 1348 |
Scalar::Util has been upgraded to version 1.27. |
| 1349 |
.Sp |
| 1350 |
Fix an overloading issue with \f(CW\*(C`sum\*(C'\fR. |
| 1351 |
.Sp |
| 1352 |
\&\f(CW\*(C`first\*(C'\fR and \f(CW\*(C`reduce\*(C'\fR now check the callback first (so \f(CW&first(1)\fR is |
| 1353 |
disallowed). |
| 1354 |
.Sp |
| 1355 |
Fix \f(CW\*(C`tainted\*(C'\fR on magical values [rt.cpan.org #55763]. |
| 1356 |
.Sp |
| 1357 |
Fix \f(CW\*(C`sum\*(C'\fR on previously magical values [rt.cpan.org #61118]. |
| 1358 |
.Sp |
| 1359 |
Fix reading past the end of a fixed buffer [rt.cpan.org #72700]. |
| 1360 |
.IP \(bu 4 |
| 1361 |
Search::Dict has been upgraded to 1.07. |
| 1362 |
.Sp |
| 1363 |
No longer require \f(CW\*(C`stat\*(C'\fR on filehandles. |
| 1364 |
.Sp |
| 1365 |
Use \f(CW\*(C`fc\*(C'\fR for casefolding. |
| 1366 |
.IP \(bu 4 |
| 1367 |
Socket has been upgraded to 2.009. |
| 1368 |
.Sp |
| 1369 |
Constants and functions required for IP multicast source group membership |
| 1370 |
have been added. |
| 1371 |
.Sp |
| 1372 |
\&\f(CWunpack_sockaddr_in()\fR and \f(CWunpack_sockaddr_in6()\fR now return just the IP |
| 1373 |
address in scalar context, and \f(CWinet_ntop()\fR now guards against incorrect |
| 1374 |
length scalars being passed in. |
| 1375 |
.Sp |
| 1376 |
This fixes an uninitialized memory read. |
| 1377 |
.IP \(bu 4 |
| 1378 |
Storable has been upgraded to 2.41. |
| 1379 |
.Sp |
| 1380 |
Modifying \f(CW$_[0]\fR within \f(CW\*(C`STORABLE_freeze\*(C'\fR no longer results in crashes |
| 1381 |
[perl #112358]. |
| 1382 |
.Sp |
| 1383 |
An object whose class implements \f(CW\*(C`STORABLE_attach\*(C'\fR is now thawed only once |
| 1384 |
when there are multiple references to it in the structure being thawed |
| 1385 |
[perl #111918]. |
| 1386 |
.Sp |
| 1387 |
Restricted hashes were not always thawed correctly [perl #73972]. |
| 1388 |
.Sp |
| 1389 |
Storable would croak when freezing a blessed REF object with a |
| 1390 |
\&\f(CWSTORABLE_freeze()\fR method [perl #113880]. |
| 1391 |
.Sp |
| 1392 |
It can now freeze and thaw vstrings correctly. This causes a slight |
| 1393 |
incompatible change in the storage format, so the format version has |
| 1394 |
increased to 2.9. |
| 1395 |
.Sp |
| 1396 |
This contains various bugfixes, including compatibility fixes for older |
| 1397 |
versions of Perl and vstring handling. |
| 1398 |
.IP \(bu 4 |
| 1399 |
Sys::Syslog has been upgraded to 0.32. |
| 1400 |
.Sp |
| 1401 |
This contains several bug fixes relating to \f(CWgetservbyname()\fR, |
| 1402 |
\&\f(CWsetlogsock()\fRand log levels in \f(CWsyslog()\fR, together with fixes for |
| 1403 |
Windows, Haiku-OS and GNU/kFreeBSD. See \fIcpan/Sys\-Syslog/Changes\fR |
| 1404 |
for the full details. |
| 1405 |
.IP \(bu 4 |
| 1406 |
Term::ANSIColor has been upgraded to 4.02. |
| 1407 |
.Sp |
| 1408 |
Add support for italics. |
| 1409 |
.Sp |
| 1410 |
Improve error handling. |
| 1411 |
.IP \(bu 4 |
| 1412 |
Term::ReadLine has been upgraded to 1.10. This fixes the |
| 1413 |
use of the \fBcpan\fR and \fBcpanp\fR shells on Windows in the event that the current |
| 1414 |
drive happens to contain a \fI\edev\etty\fR file. |
| 1415 |
.IP \(bu 4 |
| 1416 |
Test::Harness has been upgraded to 3.26. |
| 1417 |
.Sp |
| 1418 |
Fix glob semantics on Win32 [rt.cpan.org #49732]. |
| 1419 |
.Sp |
| 1420 |
Don't use \f(CW\*(C`Win32::GetShortPathName\*(C'\fR when calling perl [rt.cpan.org #47890]. |
| 1421 |
.Sp |
| 1422 |
Ignore \-T when reading shebang [rt.cpan.org #64404]. |
| 1423 |
.Sp |
| 1424 |
Handle the case where we don't know the wait status of the test more |
| 1425 |
gracefully. |
| 1426 |
.Sp |
| 1427 |
Make the test summary 'ok' line overridable so that it can be changed to a |
| 1428 |
plugin to make the output of prove idempotent. |
| 1429 |
.Sp |
| 1430 |
Don't run world-writable files. |
| 1431 |
.IP \(bu 4 |
| 1432 |
Text::Tabs and Text::Wrap have been upgraded to |
| 1433 |
2012.0818. Support for Unicode combining characters has been added to them |
| 1434 |
both. |
| 1435 |
.IP \(bu 4 |
| 1436 |
threads::shared has been upgraded to 1.31. |
| 1437 |
.Sp |
| 1438 |
This adds the option to warn about or ignore attempts to clone structures |
| 1439 |
that can't be cloned, as opposed to just unconditionally dying in |
| 1440 |
that case. |
| 1441 |
.Sp |
| 1442 |
This adds support for dual-valued values as created by |
| 1443 |
Scalar::Util::dualvar. |
| 1444 |
.IP \(bu 4 |
| 1445 |
Tie::StdHandle has been upgraded to 4.3. |
| 1446 |
.Sp |
| 1447 |
\&\f(CW\*(C`READ\*(C'\fR now respects the offset argument to \f(CW\*(C`read\*(C'\fR [perl #112826]. |
| 1448 |
.IP \(bu 4 |
| 1449 |
Time::Local has been upgraded to 1.2300. |
| 1450 |
.Sp |
| 1451 |
Seconds values greater than 59 but less than 60 no longer cause |
| 1452 |
\&\f(CWtimegm()\fR and \f(CWtimelocal()\fR to croak. |
| 1453 |
.IP \(bu 4 |
| 1454 |
Unicode::UCD has been upgraded to 0.53. |
| 1455 |
.Sp |
| 1456 |
This adds a function \fBall_casefolds()\fR |
| 1457 |
that returns all the casefolds. |
| 1458 |
.IP \(bu 4 |
| 1459 |
Win32 has been upgraded to 0.47. |
| 1460 |
.Sp |
| 1461 |
New APIs have been added for getting and setting the current code page. |
| 1462 |
.SS "Removed Modules and Pragmata" |
| 1463 |
.IX Subsection "Removed Modules and Pragmata" |
| 1464 |
.IP \(bu 4 |
| 1465 |
Version::Requirements has been removed from the core distribution. It is |
| 1466 |
available under a different name: CPAN::Meta::Requirements. |
| 1467 |
.SH Documentation |
| 1468 |
.IX Header "Documentation" |
| 1469 |
.SS "Changes to Existing Documentation" |
| 1470 |
.IX Subsection "Changes to Existing Documentation" |
| 1471 |
\fIperlcheat\fR |
| 1472 |
.IX Subsection "perlcheat" |
| 1473 |
.IP \(bu 4 |
| 1474 |
perlcheat has been reorganized, and a few new sections were added. |
| 1475 |
.PP |
| 1476 |
\fIperldata\fR |
| 1477 |
.IX Subsection "perldata" |
| 1478 |
.IP \(bu 4 |
| 1479 |
Now explicitly documents the behaviour of hash initializer lists that |
| 1480 |
contain duplicate keys. |
| 1481 |
.PP |
| 1482 |
\fIperldiag\fR |
| 1483 |
.IX Subsection "perldiag" |
| 1484 |
.IP \(bu 4 |
| 1485 |
The explanation of symbolic references being prevented by "strict refs" |
| 1486 |
now doesn't assume that the reader knows what symbolic references are. |
| 1487 |
.PP |
| 1488 |
\fIperlfaq\fR |
| 1489 |
.IX Subsection "perlfaq" |
| 1490 |
.IP \(bu 4 |
| 1491 |
perlfaq has been synchronized with version 5.0150040 from CPAN. |
| 1492 |
.PP |
| 1493 |
\fIperlfunc\fR |
| 1494 |
.IX Subsection "perlfunc" |
| 1495 |
.IP \(bu 4 |
| 1496 |
The return value of \f(CW\*(C`pipe\*(C'\fR is now documented. |
| 1497 |
.IP \(bu 4 |
| 1498 |
Clarified documentation of \f(CW\*(C`our\*(C'\fR. |
| 1499 |
.PP |
| 1500 |
\fIperlop\fR |
| 1501 |
.IX Subsection "perlop" |
| 1502 |
.IP \(bu 4 |
| 1503 |
Loop control verbs (\f(CW\*(C`dump\*(C'\fR, \f(CW\*(C`goto\*(C'\fR, \f(CW\*(C`next\*(C'\fR, \f(CW\*(C`last\*(C'\fR and \f(CW\*(C`redo\*(C'\fR) have always |
| 1504 |
had the same precedence as assignment operators, but this was not documented |
| 1505 |
until now. |
| 1506 |
.PP |
| 1507 |
\fIDiagnostics\fR |
| 1508 |
.IX Subsection "Diagnostics" |
| 1509 |
.PP |
| 1510 |
The following additions or changes have been made to diagnostic output, |
| 1511 |
including warnings and fatal error messages. For the complete list of |
| 1512 |
diagnostic messages, see perldiag. |
| 1513 |
.SS "New Diagnostics" |
| 1514 |
.IX Subsection "New Diagnostics" |
| 1515 |
\fINew Errors\fR |
| 1516 |
.IX Subsection "New Errors" |
| 1517 |
.IP \(bu 4 |
| 1518 |
Unterminated delimiter for here document |
| 1519 |
.Sp |
| 1520 |
This message now occurs when a here document label has an initial quotation |
| 1521 |
mark but the final quotation mark is missing. |
| 1522 |
.Sp |
| 1523 |
This replaces a bogus and misleading error message about not finding the label |
| 1524 |
itself [perl #114104]. |
| 1525 |
.IP \(bu 4 |
| 1526 |
panic: child pseudo-process was never scheduled |
| 1527 |
.Sp |
| 1528 |
This error is thrown when a child pseudo-process in the ithreads implementation |
| 1529 |
on Windows was not scheduled within the time period allowed and therefore was |
| 1530 |
not able to initialize properly [perl #88840]. |
| 1531 |
.IP \(bu 4 |
| 1532 |
Group name must start with a non-digit word character in regex; marked by <\-\- HERE in m/%s/ |
| 1533 |
.Sp |
| 1534 |
This error has been added for \f(CW\*(C`(?&0)\*(C'\fR, which is invalid. It used to |
| 1535 |
produce an incomprehensible error message [perl #101666]. |
| 1536 |
.IP \(bu 4 |
| 1537 |
Can't use an undefined value as a subroutine reference |
| 1538 |
.Sp |
| 1539 |
Calling an undefined value as a subroutine now produces this error message. |
| 1540 |
It used to, but was accidentally disabled, first in Perl 5.004 for |
| 1541 |
non-magical variables, and then in Perl v5.14 for magical (e.g., tied) |
| 1542 |
variables. It has now been restored. In the mean time, undef was treated |
| 1543 |
as an empty string [perl #113576]. |
| 1544 |
.IP \(bu 4 |
| 1545 |
Experimental "%s" subs not enabled |
| 1546 |
.Sp |
| 1547 |
To use lexical subs, you must first enable them: |
| 1548 |
.Sp |
| 1549 |
.Vb 3 |
| 1550 |
\& no warnings \*(Aqexperimental::lexical_subs\*(Aq; |
| 1551 |
\& use feature \*(Aqlexical_subs\*(Aq; |
| 1552 |
\& my sub foo { ... } |
| 1553 |
.Ve |
| 1554 |
.PP |
| 1555 |
\fINew Warnings\fR |
| 1556 |
.IX Subsection "New Warnings" |
| 1557 |
.IP \(bu 4 |
| 1558 |
\&'Strings with code points over 0xFF may not be mapped into in-memory file handles' |
| 1559 |
.IP \(bu 4 |
| 1560 |
\&'%s' resolved to '\eo{%s}%d' |
| 1561 |
.IP \(bu 4 |
| 1562 |
\&'Trailing white-space in a charnames alias definition is deprecated' |
| 1563 |
.IP \(bu 4 |
| 1564 |
\&'A sequence of multiple spaces in a charnames alias definition is deprecated' |
| 1565 |
.IP \(bu 4 |
| 1566 |
\&'Passing malformed UTF\-8 to "%s" is deprecated' |
| 1567 |
.IP \(bu 4 |
| 1568 |
Subroutine "&%s" is not available |
| 1569 |
.Sp |
| 1570 |
(W closure) During compilation, an inner named subroutine or eval is |
| 1571 |
attempting to capture an outer lexical subroutine that is not currently |
| 1572 |
available. This can happen for one of two reasons. First, the lexical |
| 1573 |
subroutine may be declared in an outer anonymous subroutine that has not |
| 1574 |
yet been created. (Remember that named subs are created at compile time, |
| 1575 |
while anonymous subs are created at run-time.) For example, |
| 1576 |
.Sp |
| 1577 |
.Vb 1 |
| 1578 |
\& sub { my sub a {...} sub f { \e&a } } |
| 1579 |
.Ve |
| 1580 |
.Sp |
| 1581 |
At the time that f is created, it can't capture the current the "a" sub, |
| 1582 |
since the anonymous subroutine hasn't been created yet. Conversely, the |
| 1583 |
following won't give a warning since the anonymous subroutine has by now |
| 1584 |
been created and is live: |
| 1585 |
.Sp |
| 1586 |
.Vb 1 |
| 1587 |
\& sub { my sub a {...} eval \*(Aqsub f { \e&a }\*(Aq }\->(); |
| 1588 |
.Ve |
| 1589 |
.Sp |
| 1590 |
The second situation is caused by an eval accessing a variable that has |
| 1591 |
gone out of scope, for example, |
| 1592 |
.Sp |
| 1593 |
.Vb 5 |
| 1594 |
\& sub f { |
| 1595 |
\& my sub a {...} |
| 1596 |
\& sub { eval \*(Aq\e&a\*(Aq } |
| 1597 |
\& } |
| 1598 |
\& f()\->(); |
| 1599 |
.Ve |
| 1600 |
.Sp |
| 1601 |
Here, when the '\e&a' in the eval is being compiled, f() is not currently |
| 1602 |
being executed, so its &a is not available for capture. |
| 1603 |
.IP \(bu 4 |
| 1604 |
"%s" subroutine &%s masks earlier declaration in same \f(CW%s\fR |
| 1605 |
.Sp |
| 1606 |
(W misc) A "my" or "state" subroutine has been redeclared in the |
| 1607 |
current scope or statement, effectively eliminating all access to |
| 1608 |
the previous instance. This is almost always a typographical error. |
| 1609 |
Note that the earlier subroutine will still exist until the end of |
| 1610 |
the scope or until all closure references to it are destroyed. |
| 1611 |
.IP \(bu 4 |
| 1612 |
The \f(CW%s\fR feature is experimental |
| 1613 |
.Sp |
| 1614 |
(S experimental) This warning is emitted if you enable an experimental |
| 1615 |
feature via \f(CW\*(C`use feature\*(C'\fR. Simply suppress the warning if you want |
| 1616 |
to use the feature, but know that in doing so you are taking the risk |
| 1617 |
of using an experimental feature which may change or be removed in a |
| 1618 |
future Perl version: |
| 1619 |
.Sp |
| 1620 |
.Vb 2 |
| 1621 |
\& no warnings "experimental::lexical_subs"; |
| 1622 |
\& use feature "lexical_subs"; |
| 1623 |
.Ve |
| 1624 |
.IP \(bu 4 |
| 1625 |
sleep(%u) too large |
| 1626 |
.Sp |
| 1627 |
(W overflow) You called \f(CW\*(C`sleep\*(C'\fR with a number that was larger than it can |
| 1628 |
reliably handle and \f(CW\*(C`sleep\*(C'\fR probably slept for less time than requested. |
| 1629 |
.IP \(bu 4 |
| 1630 |
Wide character in setenv |
| 1631 |
.Sp |
| 1632 |
Attempts to put wide characters into environment variables via \f(CW%ENV\fR now |
| 1633 |
provoke this warning. |
| 1634 |
.IP \(bu 4 |
| 1635 |
"Invalid negative number (%s) in chr" |
| 1636 |
.Sp |
| 1637 |
\&\f(CWchr()\fR now warns when passed a negative value [perl #83048]. |
| 1638 |
.IP \(bu 4 |
| 1639 |
"Integer overflow in srand" |
| 1640 |
.Sp |
| 1641 |
\&\f(CWsrand()\fR now warns when passed a value that doesn't fit in a \f(CW\*(C`UV\*(C'\fR (since the |
| 1642 |
value will be truncated rather than overflowing) [perl #40605]. |
| 1643 |
.IP \(bu 4 |
| 1644 |
"\-i used with no filenames on the command line, reading from STDIN" |
| 1645 |
.Sp |
| 1646 |
Running perl with the \f(CW\*(C`\-i\*(C'\fR flag now warns if no input files are provided on |
| 1647 |
the command line [perl #113410]. |
| 1648 |
.SS "Changes to Existing Diagnostics" |
| 1649 |
.IX Subsection "Changes to Existing Diagnostics" |
| 1650 |
.IP \(bu 4 |
| 1651 |
$* is no longer supported |
| 1652 |
.Sp |
| 1653 |
The warning that use of \f(CW$*\fR and \f(CW$#\fR is no longer supported is now |
| 1654 |
generated for every location that references them. Previously it would fail |
| 1655 |
to be generated if another variable using the same typeglob was seen first |
| 1656 |
(e.g. \f(CW\*(C`@*\*(C'\fR before \f(CW$*\fR), and would not be generated for the second and |
| 1657 |
subsequent uses. (It's hard to fix the failure to generate warnings at all |
| 1658 |
without also generating them every time, and warning every time is |
| 1659 |
consistent with the warnings that \f(CW$[\fR used to generate.) |
| 1660 |
.IP \(bu 4 |
| 1661 |
The warnings for \f(CW\*(C`\eb{\*(C'\fR and \f(CW\*(C`\eB{\*(C'\fR were added. They are a deprecation |
| 1662 |
warning which should be turned off by that category. One should not |
| 1663 |
have to turn off regular regexp warnings as well to get rid of these. |
| 1664 |
.IP \(bu 4 |
| 1665 |
Constant(%s): Call to &{$^H{%s}} did not return a defined value |
| 1666 |
.Sp |
| 1667 |
Constant overloading that returns \f(CW\*(C`undef\*(C'\fR results in this error message. |
| 1668 |
For numeric constants, it used to say "Constant(undef)". "undef" has been |
| 1669 |
replaced with the number itself. |
| 1670 |
.IP \(bu 4 |
| 1671 |
The error produced when a module cannot be loaded now includes a hint that |
| 1672 |
the module may need to be installed: "Can't locate hopping.pm in \f(CW@INC\fR (you |
| 1673 |
may need to install the hopping module) (@INC contains: ...)" |
| 1674 |
.IP \(bu 4 |
| 1675 |
vector argument not supported with alpha versions |
| 1676 |
.Sp |
| 1677 |
This warning was not suppressible, even with \f(CW\*(C`no warnings\*(C'\fR. Now it is |
| 1678 |
suppressible, and has been moved from the "internal" category to the |
| 1679 |
"printf" category. |
| 1680 |
.IP \(bu 4 |
| 1681 |
\&\f(CW\*(C`Can\*(Aqt do {n,m} with n > m in regex; marked by <\-\- HERE in m/%s/\*(C'\fR |
| 1682 |
.Sp |
| 1683 |
This fatal error has been turned into a warning that reads: |
| 1684 |
.Sp |
| 1685 |
Quantifier {n,m} with n > m can't match in regex |
| 1686 |
.Sp |
| 1687 |
(W regexp) Minima should be less than or equal to maxima. If you really want |
| 1688 |
your regexp to match something 0 times, just put {0}. |
| 1689 |
.IP \(bu 4 |
| 1690 |
The "Runaway prototype" warning that occurs in bizarre cases has been |
| 1691 |
removed as being unhelpful and inconsistent. |
| 1692 |
.IP \(bu 4 |
| 1693 |
The "Not a format reference" error has been removed, as the only case in |
| 1694 |
which it could be triggered was a bug. |
| 1695 |
.IP \(bu 4 |
| 1696 |
The "Unable to create sub named \f(CW%s\fR" error has been removed for the same |
| 1697 |
reason. |
| 1698 |
.IP \(bu 4 |
| 1699 |
The 'Can't use "my \f(CW%s\fR" in sort comparison' error has been downgraded to a |
| 1700 |
warning, '"my \f(CW%s\fR" used in sort comparison' (with 'state' instead of 'my' |
| 1701 |
for state variables). In addition, the heuristics for guessing whether |
| 1702 |
lexical \f(CW$a\fR or \f(CW$b\fR has been misused have been improved to generate fewer |
| 1703 |
false positives. Lexical \f(CW$a\fR and \f(CW$b\fR are no longer disallowed if they are |
| 1704 |
outside the sort block. Also, a named unary or list operator inside the |
| 1705 |
sort block no longer causes the \f(CW$a\fR or \f(CW$b\fR to be ignored [perl #86136]. |
| 1706 |
.SH "Utility Changes" |
| 1707 |
.IX Header "Utility Changes" |
| 1708 |
\fIh2xs\fR |
| 1709 |
.IX Subsection "h2xs" |
| 1710 |
.IP \(bu 4 |
| 1711 |
\&\fIh2xs\fR no longer produces invalid code for empty defines. [perl #20636] |
| 1712 |
.SH "Configuration and Compilation" |
| 1713 |
.IX Header "Configuration and Compilation" |
| 1714 |
.IP \(bu 4 |
| 1715 |
Added \f(CW\*(C`useversionedarchname\*(C'\fR option to Configure |
| 1716 |
.Sp |
| 1717 |
When set, it includes 'api_versionstring' in 'archname'. E.g. |
| 1718 |
x86_64\-linux\-5.13.6\-thread\-multi. It is unset by default. |
| 1719 |
.Sp |
| 1720 |
This feature was requested by Tim Bunce, who observed that |
| 1721 |
\&\f(CW\*(C`INSTALL_BASE\*(C'\fR creates a library structure that does not |
| 1722 |
differentiate by perl version. Instead, it places architecture |
| 1723 |
specific files in "$install_base/lib/perl5/$archname". This makes |
| 1724 |
it difficult to use a common \f(CW\*(C`INSTALL_BASE\*(C'\fR library path with |
| 1725 |
multiple versions of perl. |
| 1726 |
.Sp |
| 1727 |
By setting \f(CW\*(C`\-Duseversionedarchname\*(C'\fR, the \f(CW$archname\fR will be |
| 1728 |
distinct for architecture \fIand\fR API version, allowing mixed use of |
| 1729 |
\&\f(CW\*(C`INSTALL_BASE\*(C'\fR. |
| 1730 |
.IP \(bu 4 |
| 1731 |
Add a \f(CW\*(C`PERL_NO_INLINE_FUNCTIONS\*(C'\fR option |
| 1732 |
.Sp |
| 1733 |
If \f(CW\*(C`PERL_NO_INLINE_FUNCTIONS\*(C'\fR is defined, don't include "inline.h" |
| 1734 |
.Sp |
| 1735 |
This permits test code to include the perl headers for definitions without |
| 1736 |
creating a link dependency on the perl library (which may not exist yet). |
| 1737 |
.IP \(bu 4 |
| 1738 |
Configure will honour the external \f(CW\*(C`MAILDOMAIN\*(C'\fR environment variable, if set. |
| 1739 |
.IP \(bu 4 |
| 1740 |
\&\f(CW\*(C`installman\*(C'\fR no longer ignores the silent option |
| 1741 |
.IP \(bu 4 |
| 1742 |
Both \f(CW\*(C`META.yml\*(C'\fR and \f(CW\*(C`META.json\*(C'\fR files are now included in the distribution. |
| 1743 |
.IP \(bu 4 |
| 1744 |
\&\fIConfigure\fR will now correctly detect \f(CWisblank()\fR when compiling with a C++ |
| 1745 |
compiler. |
| 1746 |
.IP \(bu 4 |
| 1747 |
The pager detection in \fIConfigure\fR has been improved to allow responses which |
| 1748 |
specify options after the program name, e.g. \fB/usr/bin/less \-R\fR, if the user |
| 1749 |
accepts the default value. This helps \fBperldoc\fR when handling ANSI escapes |
| 1750 |
[perl #72156]. |
| 1751 |
.SH Testing |
| 1752 |
.IX Header "Testing" |
| 1753 |
.IP \(bu 4 |
| 1754 |
The test suite now has a section for tests that require very large amounts |
| 1755 |
of memory. These tests won't run by default; they can be enabled by |
| 1756 |
setting the \f(CW\*(C`PERL_TEST_MEMORY\*(C'\fR environment variable to the number of |
| 1757 |
gibibytes of memory that may be safely used. |
| 1758 |
.SH "Platform Support" |
| 1759 |
.IX Header "Platform Support" |
| 1760 |
.SS "Discontinued Platforms" |
| 1761 |
.IX Subsection "Discontinued Platforms" |
| 1762 |
.IP BeOS 4 |
| 1763 |
.IX Item "BeOS" |
| 1764 |
BeOS was an operating system for personal computers developed by Be Inc, |
| 1765 |
initially for their BeBox hardware. The OS Haiku was written as an open |
| 1766 |
source replacement for/continuation of BeOS, and its perl port is current and |
| 1767 |
actively maintained. |
| 1768 |
.IP "UTS Global" 4 |
| 1769 |
.IX Item "UTS Global" |
| 1770 |
Support code relating to UTS global has been removed. UTS was a mainframe |
| 1771 |
version of System V created by Amdahl, subsequently sold to UTS Global. The |
| 1772 |
port has not been touched since before Perl v5.8.0, and UTS Global is now |
| 1773 |
defunct. |
| 1774 |
.IP VM/ESA 4 |
| 1775 |
.IX Item "VM/ESA" |
| 1776 |
Support for VM/ESA has been removed. The port was tested on 2.3.0, which |
| 1777 |
IBM ended service on in March 2002. 2.4.0 ended service in June 2003, and |
| 1778 |
was superseded by Z/VM. The current version of Z/VM is V6.2.0, and scheduled |
| 1779 |
for end of service on 2015/04/30. |
| 1780 |
.IP MPE/IX 4 |
| 1781 |
.IX Item "MPE/IX" |
| 1782 |
Support for MPE/IX has been removed. |
| 1783 |
.IP EPOC 4 |
| 1784 |
.IX Item "EPOC" |
| 1785 |
Support code relating to EPOC has been removed. EPOC was a family of |
| 1786 |
operating systems developed by Psion for mobile devices. It was the |
| 1787 |
predecessor of Symbian. The port was last updated in April 2002. |
| 1788 |
.IP Rhapsody 4 |
| 1789 |
.IX Item "Rhapsody" |
| 1790 |
Support for Rhapsody has been removed. |
| 1791 |
.SS "Platform-Specific Notes" |
| 1792 |
.IX Subsection "Platform-Specific Notes" |
| 1793 |
\fIAIX\fR |
| 1794 |
.IX Subsection "AIX" |
| 1795 |
.PP |
| 1796 |
Configure now always adds \f(CW\*(C`\-qlanglvl=extc99\*(C'\fR to the CC flags on AIX when |
| 1797 |
using xlC. This will make it easier to compile a number of XS-based modules |
| 1798 |
that assume C99 [perl #113778]. |
| 1799 |
.PP |
| 1800 |
\fIclang++\fR |
| 1801 |
.IX Subsection "clang++" |
| 1802 |
.PP |
| 1803 |
There is now a workaround for a compiler bug that prevented compiling |
| 1804 |
with clang++ since Perl v5.15.7 [perl #112786]. |
| 1805 |
.PP |
| 1806 |
\fIC++\fR |
| 1807 |
.IX Subsection "C++" |
| 1808 |
.PP |
| 1809 |
When compiling the Perl core as C++ (which is only semi-supported), the |
| 1810 |
mathom functions are now compiled as \f(CW\*(C`extern "C"\*(C'\fR, to ensure proper |
| 1811 |
binary compatibility. (However, binary compatibility isn't generally |
| 1812 |
guaranteed anyway in the situations where this would matter.) |
| 1813 |
.PP |
| 1814 |
\fIDarwin\fR |
| 1815 |
.IX Subsection "Darwin" |
| 1816 |
.PP |
| 1817 |
Stop hardcoding an alignment on 8 byte boundaries to fix builds using |
| 1818 |
\&\-Dusemorebits. |
| 1819 |
.PP |
| 1820 |
\fIHaiku\fR |
| 1821 |
.IX Subsection "Haiku" |
| 1822 |
.PP |
| 1823 |
Perl should now work out of the box on Haiku R1 Alpha 4. |
| 1824 |
.PP |
| 1825 |
\fIMidnightBSD\fR |
| 1826 |
.IX Subsection "MidnightBSD" |
| 1827 |
.PP |
| 1828 |
\&\f(CW\*(C`libc_r\*(C'\fR was removed from recent versions of MidnightBSD and older versions |
| 1829 |
work better with \f(CW\*(C`pthread\*(C'\fR. Threading is now enabled using \f(CW\*(C`pthread\*(C'\fR which |
| 1830 |
corrects build errors with threading enabled on 0.4\-CURRENT. |
| 1831 |
.PP |
| 1832 |
\fISolaris\fR |
| 1833 |
.IX Subsection "Solaris" |
| 1834 |
.PP |
| 1835 |
In Configure, avoid running sed commands with flags not supported on Solaris. |
| 1836 |
.PP |
| 1837 |
\fIVMS\fR |
| 1838 |
.IX Subsection "VMS" |
| 1839 |
.IP \(bu 4 |
| 1840 |
Where possible, the case of filenames and command-line arguments is now |
| 1841 |
preserved by enabling the CRTL features \f(CW\*(C`DECC$EFS_CASE_PRESERVE\*(C'\fR and |
| 1842 |
\&\f(CW\*(C`DECC$ARGV_PARSE_STYLE\*(C'\fR at start-up time. The latter only takes effect |
| 1843 |
when extended parse is enabled in the process from which Perl is run. |
| 1844 |
.IP \(bu 4 |
| 1845 |
The character set for Extended Filename Syntax (EFS) is now enabled by default |
| 1846 |
on VMS. Among other things, this provides better handling of dots in directory |
| 1847 |
names, multiple dots in filenames, and spaces in filenames. To obtain the old |
| 1848 |
behavior, set the logical name \f(CW\*(C`DECC$EFS_CHARSET\*(C'\fR to \f(CW\*(C`DISABLE\*(C'\fR. |
| 1849 |
.IP \(bu 4 |
| 1850 |
Fixed linking on builds configured with \f(CW\*(C`\-Dusemymalloc=y\*(C'\fR. |
| 1851 |
.IP \(bu 4 |
| 1852 |
Experimental support for building Perl with the HP C++ compiler is available |
| 1853 |
by configuring with \f(CW\*(C`\-Dusecxx\*(C'\fR. |
| 1854 |
.IP \(bu 4 |
| 1855 |
All C header files from the top-level directory of the distribution are now |
| 1856 |
installed on VMS, providing consistency with a long-standing practice on other |
| 1857 |
platforms. Previously only a subset were installed, which broke non-core |
| 1858 |
extension builds for extensions that depended on the missing include files. |
| 1859 |
.IP \(bu 4 |
| 1860 |
Quotes are now removed from the command verb (but not the parameters) for |
| 1861 |
commands spawned via \f(CW\*(C`system\*(C'\fR, backticks, or a piped \f(CW\*(C`open\*(C'\fR. Previously, |
| 1862 |
quotes on the verb were passed through to DCL, which would fail to recognize |
| 1863 |
the command. Also, if the verb is actually a path to an image or command |
| 1864 |
procedure on an ODS\-5 volume, quoting it now allows the path to contain spaces. |
| 1865 |
.IP \(bu 4 |
| 1866 |
The \fBa2p\fR build has been fixed for the HP C++ compiler on OpenVMS. |
| 1867 |
.PP |
| 1868 |
\fIWin32\fR |
| 1869 |
.IX Subsection "Win32" |
| 1870 |
.IP \(bu 4 |
| 1871 |
Perl can now be built using Microsoft's Visual C++ 2012 compiler by specifying |
| 1872 |
CCTYPE=MSVC110 (or MSVC110FREE if you are using the free Express edition for |
| 1873 |
Windows Desktop) in \fIwin32/Makefile\fR. |
| 1874 |
.IP \(bu 4 |
| 1875 |
The option to build without \f(CW\*(C`USE_SOCKETS_AS_HANDLES\*(C'\fR has been removed. |
| 1876 |
.IP \(bu 4 |
| 1877 |
Fixed a problem where perl could crash while cleaning up threads (including the |
| 1878 |
main thread) in threaded debugging builds on Win32 and possibly other platforms |
| 1879 |
[perl #114496]. |
| 1880 |
.IP \(bu 4 |
| 1881 |
A rare race condition that would lead to sleep taking more |
| 1882 |
time than requested, and possibly even hanging, has been fixed [perl #33096]. |
| 1883 |
.IP \(bu 4 |
| 1884 |
\&\f(CW\*(C`link\*(C'\fR on Win32 now attempts to set \f(CW$!\fR to more appropriate values |
| 1885 |
based on the Win32 API error code. [perl #112272] |
| 1886 |
.Sp |
| 1887 |
Perl no longer mangles the environment block, e.g. when launching a new |
| 1888 |
sub-process, when the environment contains non-ASCII characters. Known |
| 1889 |
problems still remain, however, when the environment contains characters |
| 1890 |
outside of the current ANSI codepage (e.g. see the item about Unicode in |
| 1891 |
\&\f(CW%ENV\fR in <http://perl5.git.perl.org/perl.git/blob/HEAD:/Porting/todo.pod>). |
| 1892 |
[perl #113536] |
| 1893 |
.IP \(bu 4 |
| 1894 |
Building perl with some Windows compilers used to fail due to a problem |
| 1895 |
with miniperl's \f(CW\*(C`glob\*(C'\fR operator (which uses the \f(CW\*(C`perlglob\*(C'\fR program) |
| 1896 |
deleting the PATH environment variable [perl #113798]. |
| 1897 |
.IP \(bu 4 |
| 1898 |
A new makefile option, \f(CW\*(C`USE_64_BIT_INT\*(C'\fR, has been added to the Windows |
| 1899 |
makefiles. Set this to "define" when building a 32\-bit perl if you want |
| 1900 |
it to use 64\-bit integers. |
| 1901 |
.Sp |
| 1902 |
Machine code size reductions, already made to the DLLs of XS modules in |
| 1903 |
Perl v5.17.2, have now been extended to the perl DLL itself. |
| 1904 |
.Sp |
| 1905 |
Building with VC++ 6.0 was inadvertently broken in Perl v5.17.2 but has |
| 1906 |
now been fixed again. |
| 1907 |
.PP |
| 1908 |
\fIWinCE\fR |
| 1909 |
.IX Subsection "WinCE" |
| 1910 |
.PP |
| 1911 |
Building on WinCE is now possible once again, although more work is required |
| 1912 |
to fully restore a clean build. |
| 1913 |
.SH "Internal Changes" |
| 1914 |
.IX Header "Internal Changes" |
| 1915 |
.IP \(bu 4 |
| 1916 |
Synonyms for the misleadingly named \f(CWav_len()\fR have been created: |
| 1917 |
\&\f(CWav_top_index()\fR and \f(CW\*(C`av_tindex\*(C'\fR. All three of these return the |
| 1918 |
number of the highest index in the array, not the number of elements it |
| 1919 |
contains. |
| 1920 |
.IP \(bu 4 |
| 1921 |
\&\fBSvUPGRADE()\fR is no longer an expression. Originally this macro (and its |
| 1922 |
underlying function, \fBsv_upgrade()\fR) were documented as boolean, although |
| 1923 |
in reality they always croaked on error and never returned false. In 2005 |
| 1924 |
the documentation was updated to specify a void return value, but |
| 1925 |
\&\fBSvUPGRADE()\fR was left always returning 1 for backwards compatibility. This |
| 1926 |
has now been removed, and \fBSvUPGRADE()\fR is now a statement with no return |
| 1927 |
value. |
| 1928 |
.Sp |
| 1929 |
So this is now a syntax error: |
| 1930 |
.Sp |
| 1931 |
.Vb 1 |
| 1932 |
\& if (!SvUPGRADE(sv)) { croak(...); } |
| 1933 |
.Ve |
| 1934 |
.Sp |
| 1935 |
If you have code like that, simply replace it with |
| 1936 |
.Sp |
| 1937 |
.Vb 1 |
| 1938 |
\& SvUPGRADE(sv); |
| 1939 |
.Ve |
| 1940 |
.Sp |
| 1941 |
or to avoid compiler warnings with older perls, possibly |
| 1942 |
.Sp |
| 1943 |
.Vb 1 |
| 1944 |
\& (void)SvUPGRADE(sv); |
| 1945 |
.Ve |
| 1946 |
.IP \(bu 4 |
| 1947 |
Perl has a new copy-on-write mechanism that allows any SvPOK scalar to be |
| 1948 |
upgraded to a copy-on-write scalar. A reference count on the string buffer |
| 1949 |
is stored in the string buffer itself. This feature is \fBnot enabled by |
| 1950 |
default\fR. |
| 1951 |
.Sp |
| 1952 |
It can be enabled in a perl build by running \fIConfigure\fR with |
| 1953 |
\&\fB\-Accflags=\-DPERL_NEW_COPY_ON_WRITE\fR, and we would encourage XS authors |
| 1954 |
to try their code with such an enabled perl, and provide feedback. |
| 1955 |
Unfortunately, there is not yet a good guide to updating XS code to cope |
| 1956 |
with COW. Until such a document is available, consult the perl5\-porters |
| 1957 |
mailing list. |
| 1958 |
.Sp |
| 1959 |
It breaks a few XS modules by allowing copy-on-write scalars to go |
| 1960 |
through code paths that never encountered them before. |
| 1961 |
.IP \(bu 4 |
| 1962 |
Copy-on-write no longer uses the SvFAKE and SvREADONLY flags. Hence, |
| 1963 |
SvREADONLY indicates a true read-only SV. |
| 1964 |
.Sp |
| 1965 |
Use the SvIsCOW macro (as before) to identify a copy-on-write scalar. |
| 1966 |
.IP \(bu 4 |
| 1967 |
\&\f(CW\*(C`PL_glob_index\*(C'\fR is gone. |
| 1968 |
.IP \(bu 4 |
| 1969 |
The private Perl_croak_no_modify has had its context parameter removed. It is |
| 1970 |
now has a void prototype. Users of the public API croak_no_modify remain |
| 1971 |
unaffected. |
| 1972 |
.IP \(bu 4 |
| 1973 |
Copy-on-write (shared hash key) scalars are no longer marked read-only. |
| 1974 |
\&\f(CW\*(C`SvREADONLY\*(C'\fR returns false on such an SV, but \f(CW\*(C`SvIsCOW\*(C'\fR still returns |
| 1975 |
true. |
| 1976 |
.IP \(bu 4 |
| 1977 |
A new op type, \f(CW\*(C`OP_PADRANGE\*(C'\fR has been introduced. The perl peephole |
| 1978 |
optimiser will, where possible, substitute a single padrange op for a |
| 1979 |
pushmark followed by one or more pad ops, and possibly also skipping list |
| 1980 |
and nextstate ops. In addition, the op can carry out the tasks associated |
| 1981 |
with the RHS of a \f(CW\*(C`my(...) = @_\*(C'\fR assignment, so those ops may be optimised |
| 1982 |
away too. |
| 1983 |
.IP \(bu 4 |
| 1984 |
Case-insensitive matching inside a [bracketed] character class with a |
| 1985 |
multi-character fold no longer excludes one of the possibilities in the |
| 1986 |
circumstances that it used to. [perl #89774]. |
| 1987 |
.IP \(bu 4 |
| 1988 |
\&\f(CW\*(C`PL_formfeed\*(C'\fR has been removed. |
| 1989 |
.IP \(bu 4 |
| 1990 |
The regular expression engine no longer reads one byte past the end of the |
| 1991 |
target string. While for all internally well-formed scalars this should |
| 1992 |
never have been a problem, this change facilitates clever tricks with |
| 1993 |
string buffers in CPAN modules. [perl #73542] |
| 1994 |
.IP \(bu 4 |
| 1995 |
Inside a BEGIN block, \f(CW\*(C`PL_compcv\*(C'\fR now points to the currently-compiling |
| 1996 |
subroutine, rather than the BEGIN block itself. |
| 1997 |
.IP \(bu 4 |
| 1998 |
\&\f(CW\*(C`mg_length\*(C'\fR has been deprecated. |
| 1999 |
.IP \(bu 4 |
| 2000 |
\&\f(CW\*(C`sv_len\*(C'\fR now always returns a byte count and \f(CW\*(C`sv_len_utf8\*(C'\fR a character |
| 2001 |
count. Previously, \f(CW\*(C`sv_len\*(C'\fR and \f(CW\*(C`sv_len_utf8\*(C'\fR were both buggy and would |
| 2002 |
sometimes returns bytes and sometimes characters. \f(CW\*(C`sv_len_utf8\*(C'\fR no longer |
| 2003 |
assumes that its argument is in UTF\-8. Neither of these creates UTF\-8 caches |
| 2004 |
for tied or overloaded values or for non-PVs any more. |
| 2005 |
.IP \(bu 4 |
| 2006 |
\&\f(CW\*(C`sv_mortalcopy\*(C'\fR now copies string buffers of shared hash key scalars when |
| 2007 |
called from XS modules [perl #79824]. |
| 2008 |
.IP \(bu 4 |
| 2009 |
The new \f(CW\*(C`RXf_MODIFIES_VARS\*(C'\fR flag can be set by custom regular expression |
| 2010 |
engines to indicate that the execution of the regular expression may cause |
| 2011 |
variables to be modified. This lets \f(CW\*(C`s///\*(C'\fR know to skip certain |
| 2012 |
optimisations. Perl's own regular expression engine sets this flag for the |
| 2013 |
special backtracking verbs that set \f(CW$REGMARK\fR and \f(CW$REGERROR\fR. |
| 2014 |
.IP \(bu 4 |
| 2015 |
The APIs for accessing lexical pads have changed considerably. |
| 2016 |
.Sp |
| 2017 |
\&\f(CW\*(C`PADLIST\*(C'\fRs are now longer \f(CW\*(C`AV\*(C'\fRs, but their own type instead. |
| 2018 |
\&\f(CW\*(C`PADLIST\*(C'\fRs now contain a \f(CW\*(C`PAD\*(C'\fR and a \f(CW\*(C`PADNAMELIST\*(C'\fR of \f(CW\*(C`PADNAME\*(C'\fRs, |
| 2019 |
rather than \f(CW\*(C`AV\*(C'\fRs for the pad and the list of pad names. \f(CW\*(C`PAD\*(C'\fRs, |
| 2020 |
\&\f(CW\*(C`PADNAMELIST\*(C'\fRs, and \f(CW\*(C`PADNAME\*(C'\fRs are to be accessed as such through the |
| 2021 |
newly added pad API instead of the plain \f(CW\*(C`AV\*(C'\fR and \f(CW\*(C`SV\*(C'\fR APIs. See |
| 2022 |
perlapi for details. |
| 2023 |
.IP \(bu 4 |
| 2024 |
In the regex API, the numbered capture callbacks are passed an index |
| 2025 |
indicating what match variable is being accessed. There are special |
| 2026 |
index values for the \f(CW\*(C`$\`, $&, $&\*(C'\fR variables. Previously the same three |
| 2027 |
values were used to retrieve \f(CW\*(C`${^PREMATCH}, ${^MATCH}, ${^POSTMATCH}\*(C'\fR |
| 2028 |
too, but these have now been assigned three separate values. See |
| 2029 |
"Numbered capture callbacks" in perlreapi. |
| 2030 |
.IP \(bu 4 |
| 2031 |
\&\f(CW\*(C`PL_sawampersand\*(C'\fR was previously a boolean indicating that any of |
| 2032 |
\&\f(CW\*(C`$\`, $&, $&\*(C'\fR had been seen; it now contains three one-bit flags |
| 2033 |
indicating the presence of each of the variables individually. |
| 2034 |
.IP \(bu 4 |
| 2035 |
The \f(CW\*(C`CV *\*(C'\fR typemap entry now supports \f(CW\*(C`&{}\*(C'\fR overloading and typeglobs, |
| 2036 |
just like \f(CW\*(C`&{...}\*(C'\fR [perl #96872]. |
| 2037 |
.IP \(bu 4 |
| 2038 |
The \f(CW\*(C`SVf_AMAGIC\*(C'\fR flag to indicate overloading is now on the stash, not the |
| 2039 |
object. It is now set automatically whenever a method or \f(CW@ISA\fR changes, so |
| 2040 |
its meaning has changed, too. It now means "potentially overloaded". When |
| 2041 |
the overload table is calculated, the flag is automatically turned off if |
| 2042 |
there is no overloading, so there should be no noticeable slowdown. |
| 2043 |
.Sp |
| 2044 |
The staleness of the overload tables is now checked when overload methods |
| 2045 |
are invoked, rather than during \f(CW\*(C`bless\*(C'\fR. |
| 2046 |
.Sp |
| 2047 |
"A" magic is gone. The changes to the handling of the \f(CW\*(C`SVf_AMAGIC\*(C'\fR flag |
| 2048 |
eliminate the need for it. |
| 2049 |
.Sp |
| 2050 |
\&\f(CW\*(C`PL_amagic_generation\*(C'\fR has been removed as no longer necessary. For XS |
| 2051 |
modules, it is now a macro alias to \f(CW\*(C`PL_na\*(C'\fR. |
| 2052 |
.Sp |
| 2053 |
The fallback overload setting is now stored in a stash entry separate from |
| 2054 |
overloadedness itself. |
| 2055 |
.IP \(bu 4 |
| 2056 |
The character-processing code has been cleaned up in places. The changes |
| 2057 |
should be operationally invisible. |
| 2058 |
.IP \(bu 4 |
| 2059 |
The \f(CW\*(C`study\*(C'\fR function was made a no-op in v5.16. It was simply disabled via |
| 2060 |
a \f(CW\*(C`return\*(C'\fR statement; the code was left in place. Now the code supporting |
| 2061 |
what \f(CW\*(C`study\*(C'\fR used to do has been removed. |
| 2062 |
.IP \(bu 4 |
| 2063 |
Under threaded perls, there is no longer a separate PV allocated for every |
| 2064 |
COP to store its package name (\f(CW\*(C`cop\->stashpv\*(C'\fR). Instead, there is an |
| 2065 |
offset (\f(CW\*(C`cop\->stashoff\*(C'\fR) into the new \f(CW\*(C`PL_stashpad\*(C'\fR array, which |
| 2066 |
holds stash pointers. |
| 2067 |
.IP \(bu 4 |
| 2068 |
In the pluggable regex API, the \f(CW\*(C`regexp_engine\*(C'\fR struct has acquired a new |
| 2069 |
field \f(CW\*(C`op_comp\*(C'\fR, which is currently just for perl's internal use, and |
| 2070 |
should be initialized to NULL by other regex plugin modules. |
| 2071 |
.IP \(bu 4 |
| 2072 |
A new function \f(CW\*(C`alloccopstash\*(C'\fR has been added to the API, but is considered |
| 2073 |
experimental. See perlapi. |
| 2074 |
.IP \(bu 4 |
| 2075 |
Perl used to implement get magic in a way that would sometimes hide bugs in |
| 2076 |
code that could call \fBmg_get()\fR too many times on magical values. This hiding of |
| 2077 |
errors no longer occurs, so long-standing bugs may become visible now. If |
| 2078 |
you see magic-related errors in XS code, check to make sure it, together |
| 2079 |
with the Perl API functions it uses, calls \fBmg_get()\fR only once on \fBSvGMAGICAL()\fR |
| 2080 |
values. |
| 2081 |
.IP \(bu 4 |
| 2082 |
OP allocation for CVs now uses a slab allocator. This simplifies |
| 2083 |
memory management for OPs allocated to a CV, so cleaning up after a |
| 2084 |
compilation error is simpler and safer [perl #111462][perl #112312]. |
| 2085 |
.IP \(bu 4 |
| 2086 |
\&\f(CW\*(C`PERL_DEBUG_READONLY_OPS\*(C'\fR has been rewritten to work with the new slab |
| 2087 |
allocator, allowing it to catch more violations than before. |
| 2088 |
.IP \(bu 4 |
| 2089 |
The old slab allocator for ops, which was only enabled for \f(CW\*(C`PERL_IMPLICIT_SYS\*(C'\fR |
| 2090 |
and \f(CW\*(C`PERL_DEBUG_READONLY_OPS\*(C'\fR, has been retired. |
| 2091 |
.SH "Selected Bug Fixes" |
| 2092 |
.IX Header "Selected Bug Fixes" |
| 2093 |
.IP \(bu 4 |
| 2094 |
Here document terminators no longer require a terminating newline character when |
| 2095 |
they occur at the end of a file. This was already the case at the end of a |
| 2096 |
string eval [perl #65838]. |
| 2097 |
.IP \(bu 4 |
| 2098 |
\&\f(CW\*(C`\-DPERL_GLOBAL_STRUCT\*(C'\fR builds now free the global struct \fBafter\fR |
| 2099 |
they've finished using it. |
| 2100 |
.IP \(bu 4 |
| 2101 |
A trailing '/' on a path in \f(CW@INC\fR will no longer have an additional '/' |
| 2102 |
appended. |
| 2103 |
.IP \(bu 4 |
| 2104 |
The \f(CW\*(C`:crlf\*(C'\fR layer now works when unread data doesn't fit into its own |
| 2105 |
buffer. [perl #112244]. |
| 2106 |
.IP \(bu 4 |
| 2107 |
\&\f(CWungetc()\fR now handles UTF\-8 encoded data. [perl #116322]. |
| 2108 |
.IP \(bu 4 |
| 2109 |
A bug in the core typemap caused any C types that map to the T_BOOL core |
| 2110 |
typemap entry to not be set, updated, or modified when the T_BOOL variable was |
| 2111 |
used in an OUTPUT: section with an exception for RETVAL. T_BOOL in an INPUT: |
| 2112 |
section was not affected. Using a T_BOOL return type for an XSUB (RETVAL) |
| 2113 |
was not affected. A side effect of fixing this bug is, if a T_BOOL is specified |
| 2114 |
in the OUTPUT: section (which previous did nothing to the SV), and a read only |
| 2115 |
SV (literal) is passed to the XSUB, croaks like "Modification of a read-only |
| 2116 |
value attempted" will happen. [perl #115796] |
| 2117 |
.IP \(bu 4 |
| 2118 |
On many platforms, providing a directory name as the script name caused perl |
| 2119 |
to do nothing and report success. It should now universally report an error |
| 2120 |
and exit nonzero. [perl #61362] |
| 2121 |
.IP \(bu 4 |
| 2122 |
\&\f(CW\*(C`sort {undef} ...\*(C'\fR under fatal warnings no longer crashes. It had |
| 2123 |
begun crashing in Perl v5.16. |
| 2124 |
.IP \(bu 4 |
| 2125 |
Stashes blessed into each other |
| 2126 |
(\f(CW\*(C`bless \e%Foo::, \*(AqBar\*(Aq; bless \e%Bar::, \*(AqFoo\*(Aq\*(C'\fR) no longer result in double |
| 2127 |
frees. This bug started happening in Perl v5.16. |
| 2128 |
.IP \(bu 4 |
| 2129 |
Numerous memory leaks have been fixed, mostly involving fatal warnings and |
| 2130 |
syntax errors. |
| 2131 |
.IP \(bu 4 |
| 2132 |
Some failed regular expression matches such as \f(CW\*(C`\*(Aqf\*(Aq =~ /../g\*(C'\fR were not |
| 2133 |
resetting \f(CW\*(C`pos\*(C'\fR. Also, "match-once" patterns (\f(CW\*(C`m?...?g\*(C'\fR) failed to reset |
| 2134 |
it, too, when invoked a second time [perl #23180]. |
| 2135 |
.IP \(bu 4 |
| 2136 |
Several bugs involving \f(CW\*(C`local *ISA\*(C'\fR and \f(CW\*(C`local *Foo::\*(C'\fR causing stale |
| 2137 |
MRO caches have been fixed. |
| 2138 |
.IP \(bu 4 |
| 2139 |
Defining a subroutine when its typeglob has been aliased no longer results |
| 2140 |
in stale method caches. This bug was introduced in Perl v5.10. |
| 2141 |
.IP \(bu 4 |
| 2142 |
Localising a typeglob containing a subroutine when the typeglob's package |
| 2143 |
has been deleted from its parent stash no longer produces an error. This |
| 2144 |
bug was introduced in Perl v5.14. |
| 2145 |
.IP \(bu 4 |
| 2146 |
Under some circumstances, \f(CW\*(C`local *method=...\*(C'\fR would fail to reset method |
| 2147 |
caches upon scope exit. |
| 2148 |
.IP \(bu 4 |
| 2149 |
\&\f(CW\*(C`/[.foo.]/\*(C'\fR is no longer an error, but produces a warning (as before) and |
| 2150 |
is treated as \f(CW\*(C`/[.fo]/\*(C'\fR [perl #115818]. |
| 2151 |
.IP \(bu 4 |
| 2152 |
\&\f(CW\*(C`goto $tied_var\*(C'\fR now calls FETCH before deciding what type of goto |
| 2153 |
(subroutine or label) this is. |
| 2154 |
.IP \(bu 4 |
| 2155 |
Renaming packages through glob assignment |
| 2156 |
(\f(CW\*(C`*Foo:: = *Bar::; *Bar:: = *Baz::\*(C'\fR) in combination with \f(CW\*(C`m?...?\*(C'\fR and |
| 2157 |
\&\f(CW\*(C`reset\*(C'\fR no longer makes threaded builds crash. |
| 2158 |
.IP \(bu 4 |
| 2159 |
A number of bugs related to assigning a list to hash have been fixed. Many of |
| 2160 |
these involve lists with repeated keys like \f(CW\*(C`(1, 1, 1, 1)\*(C'\fR. |
| 2161 |
.RS 4 |
| 2162 |
.IP \(bu 4 |
| 2163 |
The expression \f(CW\*(C`scalar(%h = (1, 1, 1, 1))\*(C'\fR now returns \f(CW4\fR, not \f(CW2\fR. |
| 2164 |
.IP \(bu 4 |
| 2165 |
The return value of \f(CW\*(C`%h = (1, 1, 1)\*(C'\fR in list context was wrong. Previously |
| 2166 |
this would return \f(CW\*(C`(1, undef, 1)\*(C'\fR, now it returns \f(CW\*(C`(1, undef)\*(C'\fR. |
| 2167 |
.IP \(bu 4 |
| 2168 |
Perl now issues the same warning on \f(CW\*(C`($s, %h) = (1, {})\*(C'\fR as it does for |
| 2169 |
\&\f(CW\*(C`(%h) = ({})\*(C'\fR, "Reference found where even-sized list expected". |
| 2170 |
.IP \(bu 4 |
| 2171 |
A number of additional edge cases in list assignment to hashes were |
| 2172 |
corrected. For more details see commit 23b7025ebc. |
| 2173 |
.RE |
| 2174 |
.RS 4 |
| 2175 |
.RE |
| 2176 |
.IP \(bu 4 |
| 2177 |
Attributes applied to lexical variables no longer leak memory. |
| 2178 |
[perl #114764] |
| 2179 |
.IP \(bu 4 |
| 2180 |
\&\f(CW\*(C`dump\*(C'\fR, \f(CW\*(C`goto\*(C'\fR, \f(CW\*(C`last\*(C'\fR, \f(CW\*(C`next\*(C'\fR, \f(CW\*(C`redo\*(C'\fR or \f(CW\*(C`require\*(C'\fR followed by a |
| 2181 |
bareword (or version) and then an infix operator is no longer a syntax |
| 2182 |
error. It used to be for those infix operators (like \f(CW\*(C`+\*(C'\fR) that have a |
| 2183 |
different meaning where a term is expected. [perl #105924] |
| 2184 |
.IP \(bu 4 |
| 2185 |
\&\f(CW\*(C`require a::b . 1\*(C'\fR and \f(CW\*(C`require a::b + 1\*(C'\fR no longer produce erroneous |
| 2186 |
ambiguity warnings. [perl #107002] |
| 2187 |
.IP \(bu 4 |
| 2188 |
Class method calls are now allowed on any string, and not just strings |
| 2189 |
beginning with an alphanumeric character. [perl #105922] |
| 2190 |
.IP \(bu 4 |
| 2191 |
An empty pattern created with \f(CW\*(C`qr//\*(C'\fR used in \f(CW\*(C`m///\*(C'\fR no longer triggers |
| 2192 |
the "empty pattern reuses last pattern" behaviour. [perl #96230] |
| 2193 |
.IP \(bu 4 |
| 2194 |
Tying a hash during iteration no longer results in a memory leak. |
| 2195 |
.IP \(bu 4 |
| 2196 |
Freeing a tied hash during iteration no longer results in a memory leak. |
| 2197 |
.IP \(bu 4 |
| 2198 |
List assignment to a tied array or hash that dies on STORE no longer |
| 2199 |
results in a memory leak. |
| 2200 |
.IP \(bu 4 |
| 2201 |
If the hint hash (\f(CW\*(C`%^H\*(C'\fR) is tied, compile-time scope entry (which copies |
| 2202 |
the hint hash) no longer leaks memory if FETCH dies. [perl #107000] |
| 2203 |
.IP \(bu 4 |
| 2204 |
Constant folding no longer inappropriately triggers the special |
| 2205 |
\&\f(CW\*(C`split " "\*(C'\fR behaviour. [perl #94490] |
| 2206 |
.IP \(bu 4 |
| 2207 |
\&\f(CW\*(C`defined scalar(@array)\*(C'\fR, \f(CW\*(C`defined do { &foo }\*(C'\fR, and similar constructs |
| 2208 |
now treat the argument to \f(CW\*(C`defined\*(C'\fR as a simple scalar. [perl #97466] |
| 2209 |
.IP \(bu 4 |
| 2210 |
Running a custom debugging that defines no \f(CW*DB::DB\fR glob or provides a |
| 2211 |
subroutine stub for \f(CW&DB::DB\fR no longer results in a crash, but an error |
| 2212 |
instead. [perl #114990] |
| 2213 |
.IP \(bu 4 |
| 2214 |
\&\f(CW\*(C`reset ""\*(C'\fR now matches its documentation. \f(CW\*(C`reset\*(C'\fR only resets \f(CW\*(C`m?...?\*(C'\fR |
| 2215 |
patterns when called with no argument. An empty string for an argument now |
| 2216 |
does nothing. (It used to be treated as no argument.) [perl #97958] |
| 2217 |
.IP \(bu 4 |
| 2218 |
\&\f(CW\*(C`printf\*(C'\fR with an argument returning an empty list no longer reads past the |
| 2219 |
end of the stack, resulting in erratic behaviour. [perl #77094] |
| 2220 |
.IP \(bu 4 |
| 2221 |
\&\f(CW\*(C`\-\-subname\*(C'\fR no longer produces erroneous ambiguity warnings. |
| 2222 |
[perl #77240] |
| 2223 |
.IP \(bu 4 |
| 2224 |
\&\f(CW\*(C`v10\*(C'\fR is now allowed as a label or package name. This was inadvertently |
| 2225 |
broken when v\-strings were added in Perl v5.6. [perl #56880] |
| 2226 |
.IP \(bu 4 |
| 2227 |
\&\f(CW\*(C`length\*(C'\fR, \f(CW\*(C`pos\*(C'\fR, \f(CW\*(C`substr\*(C'\fR and \f(CW\*(C`sprintf\*(C'\fR could be confused by ties, |
| 2228 |
overloading, references and typeglobs if the stringification of such |
| 2229 |
changed the internal representation to or from UTF\-8. [perl #114410] |
| 2230 |
.IP \(bu 4 |
| 2231 |
utf8::encode now calls FETCH and STORE on tied variables. utf8::decode now |
| 2232 |
calls STORE (it was already calling FETCH). |
| 2233 |
.IP \(bu 4 |
| 2234 |
\&\f(CW\*(C`$tied =~ s/$non_utf8/$utf8/\*(C'\fR no longer loops infinitely if the tied |
| 2235 |
variable returns a Latin\-1 string, shared hash key scalar, or reference or |
| 2236 |
typeglob that stringifies as ASCII or Latin\-1. This was a regression from |
| 2237 |
v5.12. |
| 2238 |
.IP \(bu 4 |
| 2239 |
\&\f(CW\*(C`s///\*(C'\fR without /e is now better at detecting when it needs to forego |
| 2240 |
certain optimisations, fixing some buggy cases: |
| 2241 |
.RS 4 |
| 2242 |
.IP \(bu 4 |
| 2243 |
Match variables in certain constructs (\f(CW\*(C`&&\*(C'\fR, \f(CW\*(C`||\*(C'\fR, \f(CW\*(C`..\*(C'\fR and others) in |
| 2244 |
the replacement part; e.g., \f(CW\*(C`s/(.)/$l{$a||$1}/g\*(C'\fR. [perl #26986] |
| 2245 |
.IP \(bu 4 |
| 2246 |
Aliases to match variables in the replacement. |
| 2247 |
.IP \(bu 4 |
| 2248 |
\&\f(CW$REGERROR\fR or \f(CW$REGMARK\fR in the replacement. [perl #49190] |
| 2249 |
.IP \(bu 4 |
| 2250 |
An empty pattern (\f(CW\*(C`s//$foo/\*(C'\fR) that causes the last-successful pattern to |
| 2251 |
be used, when that pattern contains code blocks that modify the variables |
| 2252 |
in the replacement. |
| 2253 |
.RE |
| 2254 |
.RS 4 |
| 2255 |
.RE |
| 2256 |
.IP \(bu 4 |
| 2257 |
The taintedness of the replacement string no longer affects the taintedness |
| 2258 |
of the return value of \f(CW\*(C`s///e\*(C'\fR. |
| 2259 |
.IP \(bu 4 |
| 2260 |
The \f(CW$|\fR autoflush variable is created on-the-fly when needed. If this |
| 2261 |
happened (e.g., if it was mentioned in a module or eval) when the |
| 2262 |
currently-selected filehandle was a typeglob with an empty IO slot, it used |
| 2263 |
to crash. [perl #115206] |
| 2264 |
.IP \(bu 4 |
| 2265 |
Line numbers at the end of a string eval are no longer off by one. |
| 2266 |
[perl #114658] |
| 2267 |
.IP \(bu 4 |
| 2268 |
\&\f(CW@INC\fR filters (subroutines returned by subroutines in \f(CW@INC\fR) that set \f(CW$_\fR to a |
| 2269 |
copy-on-write scalar no longer cause the parser to modify that string |
| 2270 |
buffer in place. |
| 2271 |
.IP \(bu 4 |
| 2272 |
\&\f(CWlength($object)\fR no longer returns the undefined value if the object has |
| 2273 |
string overloading that returns undef. [perl #115260] |
| 2274 |
.IP \(bu 4 |
| 2275 |
The use of \f(CW\*(C`PL_stashcache\*(C'\fR, the stash name lookup cache for method calls, has |
| 2276 |
been restored, |
| 2277 |
.Sp |
| 2278 |
Commit da6b625f78f5f133 in August 2011 inadvertently broke the code that looks |
| 2279 |
up values in \f(CW\*(C`PL_stashcache\*(C'\fR. As it's only a cache, quite correctly everything |
| 2280 |
carried on working without it. |
| 2281 |
.IP \(bu 4 |
| 2282 |
The error "Can't localize through a reference" had disappeared in v5.16.0 |
| 2283 |
when \f(CW\*(C`local %$ref\*(C'\fR appeared on the last line of an lvalue subroutine. |
| 2284 |
This error disappeared for \f(CW\*(C`\elocal %$ref\*(C'\fR in perl v5.8.1. It has now |
| 2285 |
been restored. |
| 2286 |
.IP \(bu 4 |
| 2287 |
The parsing of here-docs has been improved significantly, fixing several |
| 2288 |
parsing bugs and crashes and one memory leak, and correcting wrong |
| 2289 |
subsequent line numbers under certain conditions. |
| 2290 |
.IP \(bu 4 |
| 2291 |
Inside an eval, the error message for an unterminated here-doc no longer |
| 2292 |
has a newline in the middle of it [perl #70836]. |
| 2293 |
.IP \(bu 4 |
| 2294 |
A substitution inside a substitution pattern (\f(CW\*(C`s/${s|||}//\*(C'\fR) no longer |
| 2295 |
confuses the parser. |
| 2296 |
.IP \(bu 4 |
| 2297 |
It may be an odd place to allow comments, but \f(CW\*(C`s//"" # hello/e\*(C'\fR has |
| 2298 |
always worked, \fIunless\fR there happens to be a null character before the |
| 2299 |
first #. Now it works even in the presence of nulls. |
| 2300 |
.IP \(bu 4 |
| 2301 |
An invalid range in \f(CW\*(C`tr///\*(C'\fR or \f(CW\*(C`y///\*(C'\fR no longer results in a memory leak. |
| 2302 |
.IP \(bu 4 |
| 2303 |
String eval no longer treats a semicolon-delimited quote-like operator at |
| 2304 |
the very end (\f(CW\*(C`eval \*(Aqq;;\*(Aq\*(C'\fR) as a syntax error. |
| 2305 |
.IP \(bu 4 |
| 2306 |
\&\f(CW\*(C`warn {$_ => 1} + 1\*(C'\fR is no longer a syntax error. The parser used to |
| 2307 |
get confused with certain list operators followed by an anonymous hash and |
| 2308 |
then an infix operator that shares its form with a unary operator. |
| 2309 |
.IP \(bu 4 |
| 2310 |
\&\f(CW\*(C`(caller $n)[6]\*(C'\fR (which gives the text of the eval) used to return the |
| 2311 |
actual parser buffer. Modifying it could result in crashes. Now it always |
| 2312 |
returns a copy. The string returned no longer has "\en;" tacked on to the |
| 2313 |
end. The returned text also includes here-doc bodies, which used to be |
| 2314 |
omitted. |
| 2315 |
.IP \(bu 4 |
| 2316 |
The UTF\-8 position cache is now reset when accessing magical variables, to |
| 2317 |
avoid the string buffer and the UTF\-8 position cache getting out of sync |
| 2318 |
[perl #114410]. |
| 2319 |
.IP \(bu 4 |
| 2320 |
Various cases of get magic being called twice for magical UTF\-8 |
| 2321 |
strings have been fixed. |
| 2322 |
.IP \(bu 4 |
| 2323 |
This code (when not in the presence of \f(CW$&\fR etc) |
| 2324 |
.Sp |
| 2325 |
.Vb 2 |
| 2326 |
\& $_ = \*(Aqx\*(Aq x 1_000_000; |
| 2327 |
\& 1 while /(.)/; |
| 2328 |
.Ve |
| 2329 |
.Sp |
| 2330 |
used to skip the buffer copy for performance reasons, but suffered from \f(CW$1\fR |
| 2331 |
etc changing if the original string changed. That's now been fixed. |
| 2332 |
.IP \(bu 4 |
| 2333 |
Perl doesn't use PerlIO anymore to report out of memory messages, as PerlIO |
| 2334 |
might attempt to allocate more memory. |
| 2335 |
.IP \(bu 4 |
| 2336 |
In a regular expression, if something is quantified with \f(CW\*(C`{n,m}\*(C'\fR where |
| 2337 |
\&\f(CW\*(C`n\ >\ m\*(C'\fR, it can't possibly match. Previously this was a fatal |
| 2338 |
error, but now is merely a warning (and that something won't match). |
| 2339 |
[perl #82954]. |
| 2340 |
.IP \(bu 4 |
| 2341 |
It used to be possible for formats defined in subroutines that have |
| 2342 |
subsequently been undefined and redefined to close over variables in the |
| 2343 |
wrong pad (the newly-defined enclosing sub), resulting in crashes or |
| 2344 |
"Bizarre copy" errors. |
| 2345 |
.IP \(bu 4 |
| 2346 |
Redefinition of XSUBs at run time could produce warnings with the wrong |
| 2347 |
line number. |
| 2348 |
.IP \(bu 4 |
| 2349 |
The \f(CW%vd\fR sprintf format does not support version objects for alpha versions. |
| 2350 |
It used to output the format itself (%vd) when passed an alpha version, and |
| 2351 |
also emit an "Invalid conversion in printf" warning. It no longer does, |
| 2352 |
but produces the empty string in the output. It also no longer leaks |
| 2353 |
memory in this case. |
| 2354 |
.IP \(bu 4 |
| 2355 |
\&\f(CW\*(C`$obj\->SUPER::method\*(C'\fR calls in the main package could fail if the |
| 2356 |
SUPER package had already been accessed by other means. |
| 2357 |
.IP \(bu 4 |
| 2358 |
Stash aliasing (\f(CW\*(C`*foo:: = *bar::\*(C'\fR) no longer causes SUPER calls to ignore |
| 2359 |
changes to methods or \f(CW@ISA\fR or use the wrong package. |
| 2360 |
.IP \(bu 4 |
| 2361 |
Method calls on packages whose names end in ::SUPER are no longer treated |
| 2362 |
as SUPER method calls, resulting in failure to find the method. |
| 2363 |
Furthermore, defining subroutines in such packages no longer causes them to |
| 2364 |
be found by SUPER method calls on the containing package [perl #114924]. |
| 2365 |
.IP \(bu 4 |
| 2366 |
\&\f(CW\*(C`\ew\*(C'\fR now matches the code points U+200C (ZERO WIDTH NON-JOINER) and U+200D |
| 2367 |
(ZERO WIDTH JOINER). \f(CW\*(C`\eW\*(C'\fR no longer matches these. This change is because |
| 2368 |
Unicode corrected their definition of what \f(CW\*(C`\ew\*(C'\fR should match. |
| 2369 |
.IP \(bu 4 |
| 2370 |
\&\f(CW\*(C`dump LABEL\*(C'\fR no longer leaks its label. |
| 2371 |
.IP \(bu 4 |
| 2372 |
Constant folding no longer changes the behaviour of functions like \f(CWstat()\fR |
| 2373 |
and \f(CWtruncate()\fR that can take either filenames or handles. |
| 2374 |
\&\f(CW\*(C`stat 1 ? foo : bar\*(C'\fR nows treats its argument as a file name (since it is an |
| 2375 |
arbitrary expression), rather than the handle "foo". |
| 2376 |
.IP \(bu 4 |
| 2377 |
\&\f(CW\*(C`truncate FOO, $len\*(C'\fR no longer falls back to treating "FOO" as a file name if |
| 2378 |
the filehandle has been deleted. This was broken in Perl v5.16.0. |
| 2379 |
.IP \(bu 4 |
| 2380 |
Subroutine redefinitions after sub-to-glob and glob-to-glob assignments no |
| 2381 |
longer cause double frees or panic messages. |
| 2382 |
.IP \(bu 4 |
| 2383 |
\&\f(CW\*(C`s///\*(C'\fR now turns vstrings into plain strings when performing a substitution, |
| 2384 |
even if the resulting string is the same (\f(CW\*(C`s/a/a/\*(C'\fR). |
| 2385 |
.IP \(bu 4 |
| 2386 |
Prototype mismatch warnings no longer erroneously treat constant subs as having |
| 2387 |
no prototype when they actually have "". |
| 2388 |
.IP \(bu 4 |
| 2389 |
Constant subroutines and forward declarations no longer prevent prototype |
| 2390 |
mismatch warnings from omitting the sub name. |
| 2391 |
.IP \(bu 4 |
| 2392 |
\&\f(CW\*(C`undef\*(C'\fR on a subroutine now clears call checkers. |
| 2393 |
.IP \(bu 4 |
| 2394 |
The \f(CW\*(C`ref\*(C'\fR operator started leaking memory on blessed objects in Perl v5.16.0. |
| 2395 |
This has been fixed [perl #114340]. |
| 2396 |
.IP \(bu 4 |
| 2397 |
\&\f(CW\*(C`use\*(C'\fR no longer tries to parse its arguments as a statement, making |
| 2398 |
\&\f(CW\*(C`use constant { () };\*(C'\fR a syntax error [perl #114222]. |
| 2399 |
.IP \(bu 4 |
| 2400 |
On debugging builds, "uninitialized" warnings inside formats no longer cause |
| 2401 |
assertion failures. |
| 2402 |
.IP \(bu 4 |
| 2403 |
On debugging builds, subroutines nested inside formats no longer cause |
| 2404 |
assertion failures [perl #78550]. |
| 2405 |
.IP \(bu 4 |
| 2406 |
Formats and \f(CW\*(C`use\*(C'\fR statements are now permitted inside formats. |
| 2407 |
.IP \(bu 4 |
| 2408 |
\&\f(CW\*(C`print $x\*(C'\fR and \f(CW\*(C`sub { print $x }\->()\*(C'\fR now always produce the same output. |
| 2409 |
It was possible for the latter to refuse to close over \f(CW$x\fR if the variable was |
| 2410 |
not active; e.g., if it was defined outside a currently-running named |
| 2411 |
subroutine. |
| 2412 |
.IP \(bu 4 |
| 2413 |
Similarly, \f(CW\*(C`print $x\*(C'\fR and \f(CW\*(C`print eval \*(Aq$x\*(Aq\*(C'\fR now produce the same output. |
| 2414 |
This also allows "my \f(CW$x\fR if 0" variables to be seen in the debugger [perl |
| 2415 |
#114018]. |
| 2416 |
.IP \(bu 4 |
| 2417 |
Formats called recursively no longer stomp on their own lexical variables, but |
| 2418 |
each recursive call has its own set of lexicals. |
| 2419 |
.IP \(bu 4 |
| 2420 |
Attempting to free an active format or the handle associated with it no longer |
| 2421 |
results in a crash. |
| 2422 |
.IP \(bu 4 |
| 2423 |
Format parsing no longer gets confused by braces, semicolons and low-precedence |
| 2424 |
operators. It used to be possible to use braces as format delimiters (instead |
| 2425 |
of \f(CW\*(C`=\*(C'\fR and \f(CW\*(C`.\*(C'\fR), but only sometimes. Semicolons and low-precedence operators |
| 2426 |
in format argument lines no longer confuse the parser into ignoring the line's |
| 2427 |
return value. In format argument lines, braces can now be used for anonymous |
| 2428 |
hashes, instead of being treated always as \f(CW\*(C`do\*(C'\fR blocks. |
| 2429 |
.IP \(bu 4 |
| 2430 |
Formats can now be nested inside code blocks in regular expressions and other |
| 2431 |
quoted constructs (\f(CW\*(C`/(?{...})/\*(C'\fR and \f(CW\*(C`qq/${...}/\*(C'\fR) [perl #114040]. |
| 2432 |
.IP \(bu 4 |
| 2433 |
Formats are no longer created after compilation errors. |
| 2434 |
.IP \(bu 4 |
| 2435 |
Under debugging builds, the \fB\-DA\fR command line option started crashing in Perl |
| 2436 |
v5.16.0. It has been fixed [perl #114368]. |
| 2437 |
.IP \(bu 4 |
| 2438 |
A potential deadlock scenario involving the premature termination of a pseudo\- |
| 2439 |
forked child in a Windows build with ithreads enabled has been fixed. This |
| 2440 |
resolves the common problem of the \fIt/op/fork.t\fR test hanging on Windows [perl |
| 2441 |
#88840]. |
| 2442 |
.IP \(bu 4 |
| 2443 |
The code which generates errors from \f(CWrequire()\fR could potentially read one or |
| 2444 |
two bytes before the start of the filename for filenames less than three bytes |
| 2445 |
long and ending \f(CW\*(C`/\e.p?\ez/\*(C'\fR. This has now been fixed. Note that it could |
| 2446 |
never have happened with module names given to \f(CWuse()\fR or \f(CWrequire()\fR anyway. |
| 2447 |
.IP \(bu 4 |
| 2448 |
The handling of pathnames of modules given to \f(CWrequire()\fR has been made |
| 2449 |
thread-safe on VMS. |
| 2450 |
.IP \(bu 4 |
| 2451 |
Non-blocking sockets have been fixed on VMS. |
| 2452 |
.IP \(bu 4 |
| 2453 |
Pod can now be nested in code inside a quoted construct outside of a string |
| 2454 |
eval. This used to work only within string evals [perl #114040]. |
| 2455 |
.IP \(bu 4 |
| 2456 |
\&\f(CW\*(C`goto \*(Aq\*(Aq\*(C'\fR now looks for an empty label, producing the "goto must have |
| 2457 |
label" error message, instead of exiting the program [perl #111794]. |
| 2458 |
.IP \(bu 4 |
| 2459 |
\&\f(CW\*(C`goto "\e0"\*(C'\fR now dies with "Can't find label" instead of "goto must have |
| 2460 |
label". |
| 2461 |
.IP \(bu 4 |
| 2462 |
The C function \f(CW\*(C`hv_store\*(C'\fR used to result in crashes when used on \f(CW\*(C`%^H\*(C'\fR |
| 2463 |
[perl #111000]. |
| 2464 |
.IP \(bu 4 |
| 2465 |
A call checker attached to a closure prototype via \f(CW\*(C`cv_set_call_checker\*(C'\fR |
| 2466 |
is now copied to closures cloned from it. So \f(CW\*(C`cv_set_call_checker\*(C'\fR now |
| 2467 |
works inside an attribute handler for a closure. |
| 2468 |
.IP \(bu 4 |
| 2469 |
Writing to \f(CW$^N\fR used to have no effect. Now it croaks with "Modification |
| 2470 |
of a read-only value" by default, but that can be overridden by a custom |
| 2471 |
regular expression engine, as with \f(CW$1\fR [perl #112184]. |
| 2472 |
.IP \(bu 4 |
| 2473 |
\&\f(CW\*(C`undef\*(C'\fR on a control character glob (\f(CW\*(C`undef *^H\*(C'\fR) no longer emits an |
| 2474 |
erroneous warning about ambiguity [perl #112456]. |
| 2475 |
.IP \(bu 4 |
| 2476 |
For efficiency's sake, many operators and built-in functions return the |
| 2477 |
same scalar each time. Lvalue subroutines and subroutines in the CORE:: |
| 2478 |
namespace were allowing this implementation detail to leak through. |
| 2479 |
\&\f(CW\*(C`print &CORE::uc("a"), &CORE::uc("b")\*(C'\fR used to print "BB". The same thing |
| 2480 |
would happen with an lvalue subroutine returning the return value of \f(CW\*(C`uc\*(C'\fR. |
| 2481 |
Now the value is copied in such cases. |
| 2482 |
.IP \(bu 4 |
| 2483 |
\&\f(CW\*(C`method {}\*(C'\fR syntax with an empty block or a block returning an empty list |
| 2484 |
used to crash or use some random value left on the stack as its invocant. |
| 2485 |
Now it produces an error. |
| 2486 |
.IP \(bu 4 |
| 2487 |
\&\f(CW\*(C`vec\*(C'\fR now works with extremely large offsets (>2 GB) [perl #111730]. |
| 2488 |
.IP \(bu 4 |
| 2489 |
Changes to overload settings now take effect immediately, as do changes to |
| 2490 |
inheritance that affect overloading. They used to take effect only after |
| 2491 |
\&\f(CW\*(C`bless\*(C'\fR. |
| 2492 |
.Sp |
| 2493 |
Objects that were created before a class had any overloading used to remain |
| 2494 |
non-overloaded even if the class gained overloading through \f(CW\*(C`use overload\*(C'\fR |
| 2495 |
or \f(CW@ISA\fR changes, and even after \f(CW\*(C`bless\*(C'\fR. This has been fixed |
| 2496 |
[perl #112708]. |
| 2497 |
.IP \(bu 4 |
| 2498 |
Classes with overloading can now inherit fallback values. |
| 2499 |
.IP \(bu 4 |
| 2500 |
Overloading was not respecting a fallback value of 0 if there were |
| 2501 |
overloaded objects on both sides of an assignment operator like \f(CW\*(C`+=\*(C'\fR |
| 2502 |
[perl #111856]. |
| 2503 |
.IP \(bu 4 |
| 2504 |
\&\f(CW\*(C`pos\*(C'\fR now croaks with hash and array arguments, instead of producing |
| 2505 |
erroneous warnings. |
| 2506 |
.IP \(bu 4 |
| 2507 |
\&\f(CW\*(C`while(each %h)\*(C'\fR now implies \f(CW\*(C`while(defined($_ = each %h))\*(C'\fR, like |
| 2508 |
\&\f(CW\*(C`readline\*(C'\fR and \f(CW\*(C`readdir\*(C'\fR. |
| 2509 |
.IP \(bu 4 |
| 2510 |
Subs in the CORE:: namespace no longer crash after \f(CW\*(C`undef *_\*(C'\fR when called |
| 2511 |
with no argument list (\f(CW&CORE::time\fR with no parentheses). |
| 2512 |
.IP \(bu 4 |
| 2513 |
\&\f(CW\*(C`unpack\*(C'\fR no longer produces the "'/' must follow a numeric type in unpack" |
| 2514 |
error when it is the data that are at fault [perl #60204]. |
| 2515 |
.IP \(bu 4 |
| 2516 |
\&\f(CW\*(C`join\*(C'\fR and \f(CW"@array"\fR now call FETCH only once on a tied \f(CW$"\fR |
| 2517 |
[perl #8931]. |
| 2518 |
.IP \(bu 4 |
| 2519 |
Some subroutine calls generated by compiling core ops affected by a |
| 2520 |
\&\f(CW\*(C`CORE::GLOBAL\*(C'\fR override had op checking performed twice. The checking |
| 2521 |
is always idempotent for pure Perl code, but the double checking can |
| 2522 |
matter when custom call checkers are involved. |
| 2523 |
.IP \(bu 4 |
| 2524 |
A race condition used to exist around fork that could cause a signal sent to |
| 2525 |
the parent to be handled by both parent and child. Signals are now blocked |
| 2526 |
briefly around fork to prevent this from happening [perl #82580]. |
| 2527 |
.IP \(bu 4 |
| 2528 |
The implementation of code blocks in regular expressions, such as \f(CW\*(C`(?{})\*(C'\fR |
| 2529 |
and \f(CW\*(C`(??{})\*(C'\fR, has been heavily reworked to eliminate a whole slew of bugs. |
| 2530 |
The main user-visible changes are: |
| 2531 |
.RS 4 |
| 2532 |
.IP \(bu 4 |
| 2533 |
Code blocks within patterns are now parsed in the same pass as the |
| 2534 |
surrounding code; in particular it is no longer necessary to have balanced |
| 2535 |
braces: this now works: |
| 2536 |
.Sp |
| 2537 |
.Vb 1 |
| 2538 |
\& /(?{ $x=\*(Aq{\*(Aq })/ |
| 2539 |
.Ve |
| 2540 |
.Sp |
| 2541 |
This means that this error message is no longer generated: |
| 2542 |
.Sp |
| 2543 |
.Vb 1 |
| 2544 |
\& Sequence (?{...}) not terminated or not {}\-balanced in regex |
| 2545 |
.Ve |
| 2546 |
.Sp |
| 2547 |
but a new error may be seen: |
| 2548 |
.Sp |
| 2549 |
.Vb 1 |
| 2550 |
\& Sequence (?{...}) not terminated with \*(Aq)\*(Aq |
| 2551 |
.Ve |
| 2552 |
.Sp |
| 2553 |
In addition, literal code blocks within run-time patterns are only |
| 2554 |
compiled once, at perl compile-time: |
| 2555 |
.Sp |
| 2556 |
.Vb 5 |
| 2557 |
\& for my $p (...) { |
| 2558 |
\& # this \*(AqFOO\*(Aq block of code is compiled once, |
| 2559 |
\& # at the same time as the surrounding \*(Aqfor\*(Aq loop |
| 2560 |
\& /$p{(?{FOO;})/; |
| 2561 |
\& } |
| 2562 |
.Ve |
| 2563 |
.IP \(bu 4 |
| 2564 |
Lexical variables are now sane as regards scope, recursion and closure |
| 2565 |
behavior. In particular, \f(CW\*(C`/A(?{B})C/\*(C'\fR behaves (from a closure viewpoint) |
| 2566 |
exactly like \f(CW\*(C`/A/ && do { B } && /C/\*(C'\fR, while \f(CW\*(C`qr/A(?{B})C/\*(C'\fR is like |
| 2567 |
\&\f(CW\*(C`sub {/A/ && do { B } && /C/}\*(C'\fR. So this code now works how you might |
| 2568 |
expect, creating three regexes that match 0, 1, and 2: |
| 2569 |
.Sp |
| 2570 |
.Vb 4 |
| 2571 |
\& for my $i (0..2) { |
| 2572 |
\& push @r, qr/^(??{$i})$/; |
| 2573 |
\& } |
| 2574 |
\& "1" =~ $r[1]; # matches |
| 2575 |
.Ve |
| 2576 |
.IP \(bu 4 |
| 2577 |
The \f(CW\*(C`use re \*(Aqeval\*(Aq\*(C'\fR pragma is now only required for code blocks defined |
| 2578 |
at runtime; in particular in the following, the text of the \f(CW$r\fR pattern is |
| 2579 |
still interpolated into the new pattern and recompiled, but the individual |
| 2580 |
compiled code-blocks within \f(CW$r\fR are reused rather than being recompiled, |
| 2581 |
and \f(CW\*(C`use re \*(Aqeval\*(Aq\*(C'\fR isn't needed any more: |
| 2582 |
.Sp |
| 2583 |
.Vb 2 |
| 2584 |
\& my $r = qr/abc(?{....})def/; |
| 2585 |
\& /xyz$r/; |
| 2586 |
.Ve |
| 2587 |
.IP \(bu 4 |
| 2588 |
Flow control operators no longer crash. Each code block runs in a new |
| 2589 |
dynamic scope, so \f(CW\*(C`next\*(C'\fR etc. will not see |
| 2590 |
any enclosing loops. \f(CW\*(C`return\*(C'\fR returns a value |
| 2591 |
from the code block, not from any enclosing subroutine. |
| 2592 |
.IP \(bu 4 |
| 2593 |
Perl normally caches the compilation of run-time patterns, and doesn't |
| 2594 |
recompile if the pattern hasn't changed, but this is now disabled if |
| 2595 |
required for the correct behavior of closures. For example: |
| 2596 |
.Sp |
| 2597 |
.Vb 5 |
| 2598 |
\& my $code = \*(Aq(??{$x})\*(Aq; |
| 2599 |
\& for my $x (1..3) { |
| 2600 |
\& # recompile to see fresh value of $x each time |
| 2601 |
\& $x =~ /$code/; |
| 2602 |
\& } |
| 2603 |
.Ve |
| 2604 |
.IP \(bu 4 |
| 2605 |
The \f(CW\*(C`/msix\*(C'\fR and \f(CW\*(C`(?msix)\*(C'\fR etc. flags are now propagated into the return |
| 2606 |
value from \f(CW\*(C`(??{})\*(C'\fR; this now works: |
| 2607 |
.Sp |
| 2608 |
.Vb 1 |
| 2609 |
\& "AB" =~ /a(??{\*(Aqb\*(Aq})/i; |
| 2610 |
.Ve |
| 2611 |
.IP \(bu 4 |
| 2612 |
Warnings and errors will appear to come from the surrounding code (or for |
| 2613 |
run-time code blocks, from an eval) rather than from an \f(CW\*(C`re_eval\*(C'\fR: |
| 2614 |
.Sp |
| 2615 |
.Vb 2 |
| 2616 |
\& use re \*(Aqeval\*(Aq; $c = \*(Aq(?{ warn "foo" })\*(Aq; /$c/; |
| 2617 |
\& /(?{ warn "foo" })/; |
| 2618 |
.Ve |
| 2619 |
.Sp |
| 2620 |
formerly gave: |
| 2621 |
.Sp |
| 2622 |
.Vb 2 |
| 2623 |
\& foo at (re_eval 1) line 1. |
| 2624 |
\& foo at (re_eval 2) line 1. |
| 2625 |
.Ve |
| 2626 |
.Sp |
| 2627 |
and now gives: |
| 2628 |
.Sp |
| 2629 |
.Vb 2 |
| 2630 |
\& foo at (eval 1) line 1. |
| 2631 |
\& foo at /some/prog line 2. |
| 2632 |
.Ve |
| 2633 |
.RE |
| 2634 |
.RS 4 |
| 2635 |
.RE |
| 2636 |
.IP \(bu 4 |
| 2637 |
Perl now can be recompiled to use any Unicode version. In v5.16, it |
| 2638 |
worked on Unicodes 6.0 and 6.1, but there were various bugs if earlier |
| 2639 |
releases were used; the older the release the more problems. |
| 2640 |
.IP \(bu 4 |
| 2641 |
\&\f(CW\*(C`vec\*(C'\fR no longer produces "uninitialized" warnings in lvalue context |
| 2642 |
[perl #9423]. |
| 2643 |
.IP \(bu 4 |
| 2644 |
An optimization involving fixed strings in regular expressions could cause |
| 2645 |
a severe performance penalty in edge cases. This has been fixed |
| 2646 |
[perl #76546]. |
| 2647 |
.IP \(bu 4 |
| 2648 |
In certain cases, including empty subpatterns within a regular expression (such |
| 2649 |
as \f(CW\*(C`(?:)\*(C'\fR or \f(CW\*(C`(?:|)\*(C'\fR) could disable some optimizations. This has been fixed. |
| 2650 |
.IP \(bu 4 |
| 2651 |
The "Can't find an opnumber" message that \f(CW\*(C`prototype\*(C'\fR produces when passed |
| 2652 |
a string like "CORE::nonexistent_keyword" now passes UTF\-8 and embedded |
| 2653 |
NULs through unchanged [perl #97478]. |
| 2654 |
.IP \(bu 4 |
| 2655 |
\&\f(CW\*(C`prototype\*(C'\fR now treats magical variables like \f(CW$1\fR the same way as |
| 2656 |
non-magical variables when checking for the CORE:: prefix, instead of |
| 2657 |
treating them as subroutine names. |
| 2658 |
.IP \(bu 4 |
| 2659 |
Under threaded perls, a runtime code block in a regular expression could |
| 2660 |
corrupt the package name stored in the op tree, resulting in bad reads |
| 2661 |
in \f(CW\*(C`caller\*(C'\fR, and possibly crashes [perl #113060]. |
| 2662 |
.IP \(bu 4 |
| 2663 |
Referencing a closure prototype (\f(CW\*(C`\e&{$_[1]}\*(C'\fR in an attribute handler for a |
| 2664 |
closure) no longer results in a copy of the subroutine (or assertion |
| 2665 |
failures on debugging builds). |
| 2666 |
.IP \(bu 4 |
| 2667 |
\&\f(CW\*(C`eval \*(Aq_\|_PACKAGE_\|_\*(Aq\*(C'\fR now returns the right answer on threaded builds if |
| 2668 |
the current package has been assigned over (as in |
| 2669 |
\&\f(CW\*(C`*ThisPackage:: = *ThatPackage::\*(C'\fR) [perl #78742]. |
| 2670 |
.IP \(bu 4 |
| 2671 |
If a package is deleted by code that it calls, it is possible for \f(CW\*(C`caller\*(C'\fR |
| 2672 |
to see a stack frame belonging to that deleted package. \f(CW\*(C`caller\*(C'\fR could |
| 2673 |
crash if the stash's memory address was reused for a scalar and a |
| 2674 |
substitution was performed on the same scalar [perl #113486]. |
| 2675 |
.IP \(bu 4 |
| 2676 |
\&\f(CW\*(C`UNIVERSAL::can\*(C'\fR no longer treats its first argument differently |
| 2677 |
depending on whether it is a string or number internally. |
| 2678 |
.IP \(bu 4 |
| 2679 |
\&\f(CW\*(C`open\*(C'\fR with \f(CW\*(C`<&\*(C'\fR for the mode checks to see whether the third argument is |
| 2680 |
a number, in determining whether to treat it as a file descriptor or a handle |
| 2681 |
name. Magical variables like \f(CW$1\fR were always failing the numeric check and |
| 2682 |
being treated as handle names. |
| 2683 |
.IP \(bu 4 |
| 2684 |
\&\f(CW\*(C`warn\*(C'\fR's handling of magical variables (\f(CW$1\fR, ties) has undergone several |
| 2685 |
fixes. \f(CW\*(C`FETCH\*(C'\fR is only called once now on a tied argument or a tied \f(CW$@\fR |
| 2686 |
[perl #97480]. Tied variables returning objects that stringify as "" are |
| 2687 |
no longer ignored. A tied \f(CW$@\fR that happened to return a reference the |
| 2688 |
\&\fIprevious\fR time it was used is no longer ignored. |
| 2689 |
.IP \(bu 4 |
| 2690 |
\&\f(CW\*(C`warn ""\*(C'\fR now treats \f(CW$@\fR with a number in it the same way, regardless of |
| 2691 |
whether it happened via \f(CW\*(C`$@=3\*(C'\fR or \f(CW\*(C`$@="3"\*(C'\fR. It used to ignore the |
| 2692 |
former. Now it appends "\et...caught", as it has always done with |
| 2693 |
\&\f(CW\*(C`$@="3"\*(C'\fR. |
| 2694 |
.IP \(bu 4 |
| 2695 |
Numeric operators on magical variables (e.g., \f(CW\*(C`$1\ +\ 1\*(C'\fR) used to use |
| 2696 |
floating point operations even where integer operations were more appropriate, |
| 2697 |
resulting in loss of accuracy on 64\-bit platforms [perl #109542]. |
| 2698 |
.IP \(bu 4 |
| 2699 |
Unary negation no longer treats a string as a number if the string happened |
| 2700 |
to be used as a number at some point. So, if \f(CW$x\fR contains the string "dogs", |
| 2701 |
\&\f(CW\*(C`\-$x\*(C'\fR returns "\-dogs" even if \f(CW\*(C`$y=0+$x\*(C'\fR has happened at some point. |
| 2702 |
.IP \(bu 4 |
| 2703 |
In Perl v5.14, \f(CW\*(C`\-\*(Aq\-10\*(Aq\*(C'\fR was fixed to return "10", not "+10". But magical |
| 2704 |
variables (\f(CW$1\fR, ties) were not fixed till now [perl #57706]. |
| 2705 |
.IP \(bu 4 |
| 2706 |
Unary negation now treats strings consistently, regardless of the internal |
| 2707 |
\&\f(CW\*(C`UTF8\*(C'\fR flag. |
| 2708 |
.IP \(bu 4 |
| 2709 |
A regression introduced in Perl v5.16.0 involving |
| 2710 |
\&\f(CW\*(C`tr/\fR\f(CISEARCHLIST\fR\f(CW/\fR\f(CIREPLACEMENTLIST\fR\f(CW/\*(C'\fR has been fixed. Only the first |
| 2711 |
instance is supposed to be meaningful if a character appears more than |
| 2712 |
once in \f(CW\*(C`\fR\f(CISEARCHLIST\fR\f(CW\*(C'\fR. Under some circumstances, the final instance |
| 2713 |
was overriding all earlier ones. [perl #113584] |
| 2714 |
.IP \(bu 4 |
| 2715 |
Regular expressions like \f(CW\*(C`qr/\e87/\*(C'\fR previously silently inserted a NUL |
| 2716 |
character, thus matching as if it had been written \f(CW\*(C`qr/\e00087/\*(C'\fR. Now it |
| 2717 |
matches as if it had been written as \f(CW\*(C`qr/87/\*(C'\fR, with a message that the |
| 2718 |
sequence \f(CW"\e8"\fR is unrecognized. |
| 2719 |
.IP \(bu 4 |
| 2720 |
\&\f(CW\*(C`_\|_SUB_\|_\*(C'\fR now works in special blocks (\f(CW\*(C`BEGIN\*(C'\fR, \f(CW\*(C`END\*(C'\fR, etc.). |
| 2721 |
.IP \(bu 4 |
| 2722 |
Thread creation on Windows could theoretically result in a crash if done |
| 2723 |
inside a \f(CW\*(C`BEGIN\*(C'\fR block. It still does not work properly, but it no longer |
| 2724 |
crashes [perl #111610]. |
| 2725 |
.IP \(bu 4 |
| 2726 |
\&\f(CW\*(C`\e&{\*(Aq\*(Aq}\*(C'\fR (with the empty string) now autovivifies a stub like any other |
| 2727 |
sub name, and no longer produces the "Unable to create sub" error |
| 2728 |
[perl #94476]. |
| 2729 |
.IP \(bu 4 |
| 2730 |
A regression introduced in v5.14.0 has been fixed, in which some calls |
| 2731 |
to the \f(CW\*(C`re\*(C'\fR module would clobber \f(CW$_\fR [perl #113750]. |
| 2732 |
.IP \(bu 4 |
| 2733 |
\&\f(CW\*(C`do FILE\*(C'\fR now always either sets or clears \f(CW$@\fR, even when the file can't be |
| 2734 |
read. This ensures that testing \f(CW$@\fR first (as recommended by the |
| 2735 |
documentation) always returns the correct result. |
| 2736 |
.IP \(bu 4 |
| 2737 |
The array iterator used for the \f(CW\*(C`each @array\*(C'\fR construct is now correctly |
| 2738 |
reset when \f(CW@array\fR is cleared [perl #75596]. This happens, for example, when |
| 2739 |
the array is globally assigned to, as in \f(CW\*(C`@array = (...)\*(C'\fR, but not when its |
| 2740 |
\&\fBvalues\fR are assigned to. In terms of the XS API, it means that \f(CWav_clear()\fR |
| 2741 |
will now reset the iterator. |
| 2742 |
.Sp |
| 2743 |
This mirrors the behaviour of the hash iterator when the hash is cleared. |
| 2744 |
.IP \(bu 4 |
| 2745 |
\&\f(CW\*(C`$class\->can\*(C'\fR, \f(CW\*(C`$class\->isa\*(C'\fR, and \f(CW\*(C`$class\->DOES\*(C'\fR now return |
| 2746 |
correct results, regardless of whether that package referred to by \f(CW$class\fR |
| 2747 |
exists [perl #47113]. |
| 2748 |
.IP \(bu 4 |
| 2749 |
Arriving signals no longer clear \f(CW$@\fR [perl #45173]. |
| 2750 |
.IP \(bu 4 |
| 2751 |
Allow \f(CW\*(C`my ()\*(C'\fR declarations with an empty variable list [perl #113554]. |
| 2752 |
.IP \(bu 4 |
| 2753 |
During parsing, subs declared after errors no longer leave stubs |
| 2754 |
[perl #113712]. |
| 2755 |
.IP \(bu 4 |
| 2756 |
Closures containing no string evals no longer hang on to their containing |
| 2757 |
subroutines, allowing variables closed over by outer subroutines to be |
| 2758 |
freed when the outer sub is freed, even if the inner sub still exists |
| 2759 |
[perl #89544]. |
| 2760 |
.IP \(bu 4 |
| 2761 |
Duplication of in-memory filehandles by opening with a "<&=" or ">&=" mode |
| 2762 |
stopped working properly in v5.16.0. It was causing the new handle to |
| 2763 |
reference a different scalar variable. This has been fixed [perl #113764]. |
| 2764 |
.IP \(bu 4 |
| 2765 |
\&\f(CW\*(C`qr//\*(C'\fR expressions no longer crash with custom regular expression engines |
| 2766 |
that do not set \f(CW\*(C`offs\*(C'\fR at regular expression compilation time |
| 2767 |
[perl #112962]. |
| 2768 |
.IP \(bu 4 |
| 2769 |
\&\f(CW\*(C`delete local\*(C'\fR no longer crashes with certain magical arrays and hashes |
| 2770 |
[perl #112966]. |
| 2771 |
.IP \(bu 4 |
| 2772 |
\&\f(CW\*(C`local\*(C'\fR on elements of certain magical arrays and hashes used not to |
| 2773 |
arrange to have the element deleted on scope exit, even if the element did |
| 2774 |
not exist before \f(CW\*(C`local\*(C'\fR. |
| 2775 |
.IP \(bu 4 |
| 2776 |
\&\f(CWscalar(write)\fR no longer returns multiple items [perl #73690]. |
| 2777 |
.IP \(bu 4 |
| 2778 |
String to floating point conversions no longer misparse certain strings under |
| 2779 |
\&\f(CW\*(C`use locale\*(C'\fR [perl #109318]. |
| 2780 |
.IP \(bu 4 |
| 2781 |
\&\f(CW@INC\fR filters that die no longer leak memory [perl #92252]. |
| 2782 |
.IP \(bu 4 |
| 2783 |
The implementations of overloaded operations are now called in the correct |
| 2784 |
context. This allows, among other things, being able to properly override |
| 2785 |
\&\f(CW\*(C`<>\*(C'\fR [perl #47119]. |
| 2786 |
.IP \(bu 4 |
| 2787 |
Specifying only the \f(CW\*(C`fallback\*(C'\fR key when calling \f(CW\*(C`use overload\*(C'\fR now behaves |
| 2788 |
properly [perl #113010]. |
| 2789 |
.IP \(bu 4 |
| 2790 |
\&\f(CW\*(C`sub foo { my $a = 0; while ($a) { ... } }\*(C'\fR and |
| 2791 |
\&\f(CW\*(C`sub foo { while (0) { ... } }\*(C'\fR now return the same thing [perl #73618]. |
| 2792 |
.IP \(bu 4 |
| 2793 |
String negation now behaves the same under \f(CW\*(C`use integer;\*(C'\fR as it does |
| 2794 |
without [perl #113012]. |
| 2795 |
.IP \(bu 4 |
| 2796 |
\&\f(CW\*(C`chr\*(C'\fR now returns the Unicode replacement character (U+FFFD) for \-1, |
| 2797 |
regardless of the internal representation. \-1 used to wrap if the argument |
| 2798 |
was tied or a string internally. |
| 2799 |
.IP \(bu 4 |
| 2800 |
Using a \f(CW\*(C`format\*(C'\fR after its enclosing sub was freed could crash as of |
| 2801 |
perl v5.12.0, if the format referenced lexical variables from the outer sub. |
| 2802 |
.IP \(bu 4 |
| 2803 |
Using a \f(CW\*(C`format\*(C'\fR after its enclosing sub was undefined could crash as of |
| 2804 |
perl v5.10.0, if the format referenced lexical variables from the outer sub. |
| 2805 |
.IP \(bu 4 |
| 2806 |
Using a \f(CW\*(C`format\*(C'\fR defined inside a closure, which format references |
| 2807 |
lexical variables from outside, never really worked unless the \f(CW\*(C`write\*(C'\fR |
| 2808 |
call was directly inside the closure. In v5.10.0 it even started crashing. |
| 2809 |
Now the copy of that closure nearest the top of the call stack is used to |
| 2810 |
find those variables. |
| 2811 |
.IP \(bu 4 |
| 2812 |
Formats that close over variables in special blocks no longer crash if a |
| 2813 |
stub exists with the same name as the special block before the special |
| 2814 |
block is compiled. |
| 2815 |
.IP \(bu 4 |
| 2816 |
The parser no longer gets confused, treating \f(CW\*(C`eval foo ()\*(C'\fR as a syntax |
| 2817 |
error if preceded by \f(CW\*(C`print;\*(C'\fR [perl #16249]. |
| 2818 |
.IP \(bu 4 |
| 2819 |
The return value of \f(CW\*(C`syscall\*(C'\fR is no longer truncated on 64\-bit platforms |
| 2820 |
[perl #113980]. |
| 2821 |
.IP \(bu 4 |
| 2822 |
Constant folding no longer causes \f(CW\*(C`print 1 ? FOO : BAR\*(C'\fR to print to the |
| 2823 |
FOO handle [perl #78064]. |
| 2824 |
.IP \(bu 4 |
| 2825 |
\&\f(CW\*(C`do subname\*(C'\fR now calls the named subroutine and uses the file name it |
| 2826 |
returns, instead of opening a file named "subname". |
| 2827 |
.IP \(bu 4 |
| 2828 |
Subroutines looked up by rv2cv check hooks (registered by XS modules) are |
| 2829 |
now taken into consideration when determining whether \f(CW\*(C`foo bar\*(C'\fR should be |
| 2830 |
the sub call \f(CWfoo(bar)\fR or the method call \f(CW\*(C`"bar"\->foo\*(C'\fR. |
| 2831 |
.IP \(bu 4 |
| 2832 |
\&\f(CW\*(C`CORE::foo::bar\*(C'\fR is no longer treated specially, allowing global overrides |
| 2833 |
to be called directly via \f(CWCORE::GLOBAL::uc(...)\fR [perl #113016]. |
| 2834 |
.IP \(bu 4 |
| 2835 |
Calling an undefined sub whose typeglob has been undefined now produces the |
| 2836 |
customary "Undefined subroutine called" error, instead of "Not a CODE |
| 2837 |
reference". |
| 2838 |
.IP \(bu 4 |
| 2839 |
Two bugs involving \f(CW@ISA\fR have been fixed. \f(CW\*(C`*ISA = *glob_without_array\*(C'\fR and |
| 2840 |
\&\f(CW\*(C`undef *ISA; @{*ISA}\*(C'\fR would prevent future modifications to \f(CW@ISA\fR from |
| 2841 |
updating the internal caches used to look up methods. The |
| 2842 |
*glob_without_array case was a regression from Perl v5.12. |
| 2843 |
.IP \(bu 4 |
| 2844 |
Regular expression optimisations sometimes caused \f(CW\*(C`$\*(C'\fR with \f(CW\*(C`/m\*(C'\fR to |
| 2845 |
produce failed or incorrect matches [perl #114068]. |
| 2846 |
.IP \(bu 4 |
| 2847 |
\&\f(CW\*(C`_\|_SUB_\|_\*(C'\fR now works in a \f(CW\*(C`sort\*(C'\fR block when the enclosing subroutine is |
| 2848 |
predeclared with \f(CW\*(C`sub foo;\*(C'\fR syntax [perl #113710]. |
| 2849 |
.IP \(bu 4 |
| 2850 |
Unicode properties only apply to Unicode code points, which leads to |
| 2851 |
some subtleties when regular expressions are matched against |
| 2852 |
above-Unicode code points. There is a warning generated to draw your |
| 2853 |
attention to this. However, this warning was being generated |
| 2854 |
inappropriately in some cases, such as when a program was being parsed. |
| 2855 |
Non-Unicode matches such as \f(CW\*(C`\ew\*(C'\fR and \f(CW\*(C`[:word:]\*(C'\fR should not generate the |
| 2856 |
warning, as their definitions don't limit them to apply to only Unicode |
| 2857 |
code points. Now the message is only generated when matching against |
| 2858 |
\&\f(CW\*(C`\ep{}\*(C'\fR and \f(CW\*(C`\eP{}\*(C'\fR. There remains a bug, [perl #114148], for the very |
| 2859 |
few properties in Unicode that match just a single code point. The |
| 2860 |
warning is not generated if they are matched against an above-Unicode |
| 2861 |
code point. |
| 2862 |
.IP \(bu 4 |
| 2863 |
Uninitialized warnings mentioning hash elements would only mention the |
| 2864 |
element name if it was not in the first bucket of the hash, due to an |
| 2865 |
off-by-one error. |
| 2866 |
.IP \(bu 4 |
| 2867 |
A regular expression optimizer bug could cause multiline "^" to behave |
| 2868 |
incorrectly in the presence of line breaks, such that |
| 2869 |
\&\f(CW\*(C`"/\en\en" =~ m#\eA(?:^/$)#im\*(C'\fR would not match [perl #115242]. |
| 2870 |
.IP \(bu 4 |
| 2871 |
Failed \f(CW\*(C`fork\*(C'\fR in list context no longer corrupts the stack. |
| 2872 |
\&\f(CW\*(C`@a = (1, 2, fork, 3)\*(C'\fR used to gobble up the 2 and assign \f(CW\*(C`(1, undef, 3)\*(C'\fR |
| 2873 |
if the \f(CW\*(C`fork\*(C'\fR call failed. |
| 2874 |
.IP \(bu 4 |
| 2875 |
Numerous memory leaks have been fixed, mostly involving tied variables that |
| 2876 |
die, regular expression character classes and code blocks, and syntax |
| 2877 |
errors. |
| 2878 |
.IP \(bu 4 |
| 2879 |
Assigning a regular expression (\f(CW\*(C`${qr//}\*(C'\fR) to a variable that happens to |
| 2880 |
hold a floating point number no longer causes assertion failures on |
| 2881 |
debugging builds. |
| 2882 |
.IP \(bu 4 |
| 2883 |
Assigning a regular expression to a scalar containing a number no longer |
| 2884 |
causes subsequent numification to produce random numbers. |
| 2885 |
.IP \(bu 4 |
| 2886 |
Assigning a regular expression to a magic variable no longer wipes away the |
| 2887 |
magic. This was a regression from v5.10. |
| 2888 |
.IP \(bu 4 |
| 2889 |
Assigning a regular expression to a blessed scalar no longer results in |
| 2890 |
crashes. This was also a regression from v5.10. |
| 2891 |
.IP \(bu 4 |
| 2892 |
Regular expression can now be assigned to tied hash and array elements with |
| 2893 |
flattening into strings. |
| 2894 |
.IP \(bu 4 |
| 2895 |
Numifying a regular expression no longer results in an uninitialized |
| 2896 |
warning. |
| 2897 |
.IP \(bu 4 |
| 2898 |
Negative array indices no longer cause EXISTS methods of tied variables to |
| 2899 |
be ignored. This was a regression from v5.12. |
| 2900 |
.IP \(bu 4 |
| 2901 |
Negative array indices no longer result in crashes on arrays tied to |
| 2902 |
non-objects. |
| 2903 |
.IP \(bu 4 |
| 2904 |
\&\f(CW\*(C`$byte_overload .= $utf8\*(C'\fR no longer results in doubly-encoded UTF\-8 if the |
| 2905 |
left-hand scalar happened to have produced a UTF\-8 string the last time |
| 2906 |
overloading was invoked. |
| 2907 |
.IP \(bu 4 |
| 2908 |
\&\f(CW\*(C`goto &sub\*(C'\fR now uses the current value of \f(CW@_\fR, instead of using the array |
| 2909 |
the subroutine was originally called with. This means |
| 2910 |
\&\f(CW\*(C`local @_ = (...); goto &sub\*(C'\fR now works [perl #43077]. |
| 2911 |
.IP \(bu 4 |
| 2912 |
If a debugger is invoked recursively, it no longer stomps on its own |
| 2913 |
lexical variables. Formerly under recursion all calls would share the same |
| 2914 |
set of lexical variables [perl #115742]. |
| 2915 |
.IP \(bu 4 |
| 2916 |
\&\f(CW*_{ARRAY}\fR returned from a subroutine no longer spontaneously |
| 2917 |
becomes empty. |
| 2918 |
.IP \(bu 4 |
| 2919 |
When using \f(CW\*(C`say\*(C'\fR to print to a tied filehandle, the value of \f(CW\*(C`$\e\*(C'\fR is |
| 2920 |
correctly localized, even if it was previously undef. [perl #119927] |
| 2921 |
.SH "Known Problems" |
| 2922 |
.IX Header "Known Problems" |
| 2923 |
.IP \(bu 4 |
| 2924 |
UTF8\-flagged strings in \f(CW%ENV\fR on HP-UX 11.00 are buggy |
| 2925 |
.Sp |
| 2926 |
The interaction of UTF8\-flagged strings and \f(CW%ENV\fR on HP-UX 11.00 is |
| 2927 |
currently dodgy in some not-yet-fully-diagnosed way. Expect test |
| 2928 |
failures in \fIt/op/magic.t\fR, followed by unknown behavior when storing |
| 2929 |
wide characters in the environment. |
| 2930 |
.SH Obituary |
| 2931 |
.IX Header "Obituary" |
| 2932 |
Hojung Yoon (AMORETTE), 24, of Seoul, South Korea, went to his long rest |
| 2933 |
on May 8, 2013 with llama figurine and autographed TIMTOADY card. He |
| 2934 |
was a brilliant young Perl 5 & 6 hacker and a devoted member of |
| 2935 |
Seoul.pm. He programmed Perl, talked Perl, ate Perl, and loved Perl. We |
| 2936 |
believe that he is still programming in Perl with his broken IBM laptop |
| 2937 |
somewhere. He will be missed. |
| 2938 |
.SH Acknowledgements |
| 2939 |
.IX Header "Acknowledgements" |
| 2940 |
Perl v5.18.0 represents approximately 12 months of development since |
| 2941 |
Perl v5.16.0 and contains approximately 400,000 lines of changes across |
| 2942 |
2,100 files from 113 authors. |
| 2943 |
.PP |
| 2944 |
Perl continues to flourish into its third decade thanks to a vibrant |
| 2945 |
community of users and developers. The following people are known to |
| 2946 |
have contributed the improvements that became Perl v5.18.0: |
| 2947 |
.PP |
| 2948 |
Aaron Crane, Aaron Trevena, Abhijit Menon-Sen, Adrian M. Enache, Alan |
| 2949 |
Haggai Alavi, Alexandr Ciornii, Andrew Tam, Andy Dougherty, Anton Nikishaev, |
| 2950 |
Aristotle Pagaltzis, Augustina Blair, Bob Ernst, Brad Gilbert, Breno G. de |
| 2951 |
Oliveira, Brian Carlson, Brian Fraser, Charlie Gonzalez, Chip Salzenberg, Chris |
| 2952 |
\&'BinGOs' Williams, Christian Hansen, Colin Kuskie, Craig A. Berry, Dagfinn |
| 2953 |
Ilmari Mannsåker, Daniel Dragan, Daniel Perrett, Darin McBride, Dave Rolsky, |
| 2954 |
David Golden, David Leadbeater, David Mitchell, David Nicol, Dominic |
| 2955 |
Hargreaves, E. Choroba, Eric Brine, Evan Miller, Father Chrysostomos, Florian |
| 2956 |
Ragwitz, François Perrad, George Greer, Goro Fuji, H.Merijn Brand, Herbert |
| 2957 |
Breunung, Hugo van der Sanden, Igor Zaytsev, James E Keenan, Jan Dubois, |
| 2958 |
Jasmine Ahuja, Jerry D. Hedden, Jess Robinson, Jesse Luehrs, Joaquin Ferrero, |
| 2959 |
Joel Berger, John Goodyear, John Peacock, Karen Etheridge, Karl Williamson, |
| 2960 |
Karthik Rajagopalan, Kent Fredric, Leon Timmermans, Lucas Holt, Lukas Mai, |
| 2961 |
Marcus Holland-Moritz, Markus Jansen, Martin Hasch, Matthew Horsfall, Max |
| 2962 |
Maischein, Michael G Schwern, Michael Schroeder, Moritz Lenz, Nicholas Clark, |
| 2963 |
Niko Tyni, Oleg Nesterov, Patrik Hägglund, Paul Green, Paul Johnson, Paul |
| 2964 |
Marquess, Peter Martini, Rafael Garcia-Suarez, Reini Urban, Renee Baecker, |
| 2965 |
Rhesa Rozendaal, Ricardo Signes, Robin Barker, Ronald J. Kimball, Ruslan |
| 2966 |
Zakirov, Salvador Fandiño, Sawyer X, Scott Lanning, Sergey Alekseev, Shawn M |
| 2967 |
Moore, Shirakata Kentaro, Shlomi Fish, Sisyphus, Smylers, Steffen Müller, |
| 2968 |
Steve Hay, Steve Peters, Steven Schubiger, Sullivan Beck, Sven Strickroth, |
| 2969 |
Sébastien Aperghis-Tramoni, Thomas Sibley, Tobias Leich, Tom Wyant, Tony Cook, |
| 2970 |
Vadim Konovalov, Vincent Pit, Volker Schatz, Walt Mankowski, Yves Orton, |
| 2971 |
Zefram. |
| 2972 |
.PP |
| 2973 |
The list above is almost certainly incomplete as it is automatically generated |
| 2974 |
from version control history. In particular, it does not include the names of |
| 2975 |
the (very much appreciated) contributors who reported issues to the Perl bug |
| 2976 |
tracker. |
| 2977 |
.PP |
| 2978 |
Many of the changes included in this version originated in the CPAN modules |
| 2979 |
included in Perl's core. We're grateful to the entire CPAN community for |
| 2980 |
helping Perl to flourish. |
| 2981 |
.PP |
| 2982 |
For a more complete list of all of Perl's historical contributors, please see |
| 2983 |
the \fIAUTHORS\fR file in the Perl source distribution. |
| 2984 |
.SH "Reporting Bugs" |
| 2985 |
.IX Header "Reporting Bugs" |
| 2986 |
If you find what you think is a bug, you might check the articles recently |
| 2987 |
posted to the comp.lang.perl.misc newsgroup and the perl bug database at |
| 2988 |
http://rt.perl.org/perlbug/ . There may also be information at |
| 2989 |
http://www.perl.org/ , the Perl Home Page. |
| 2990 |
.PP |
| 2991 |
If you believe you have an unreported bug, please run the perlbug program |
| 2992 |
included with your release. Be sure to trim your bug down to a tiny but |
| 2993 |
sufficient test case. Your bug report, along with the output of \f(CW\*(C`perl \-V\*(C'\fR, |
| 2994 |
will be sent off to [email protected] to be analysed by the Perl porting team. |
| 2995 |
.PP |
| 2996 |
If the bug you are reporting has security implications, which make it |
| 2997 |
inappropriate to send to a publicly archived mailing list, then please send it |
| 2998 |
to perl5\-security\[email protected]. This points to a closed subscription |
| 2999 |
unarchived mailing list, which includes all the core committers, who will be |
| 3000 |
able to help assess the impact of issues, figure out a resolution, and help |
| 3001 |
co-ordinate the release of patches to mitigate or fix the problem across all |
| 3002 |
platforms on which Perl is supported. Please only use this address for |
| 3003 |
security issues in the Perl core, not for modules independently distributed on |
| 3004 |
CPAN. |
| 3005 |
.SH "SEE ALSO" |
| 3006 |
.IX Header "SEE ALSO" |
| 3007 |
The \fIChanges\fR file for an explanation of how to view exhaustive details on |
| 3008 |
what changed. |
| 3009 |
.PP |
| 3010 |
The \fIINSTALL\fR file for how to build Perl. |
| 3011 |
.PP |
| 3012 |
The \fIREADME\fR file for general stuff. |
| 3013 |
.PP |
| 3014 |
The \fIArtistic\fR and \fICopying\fR files for copyright information. |