| 1 |
Revision history for indirect |
| 2 |
|
| 3 |
0.39 2019-07-08 12:00 UTC |
| 4 |
+ Fix : [RT #127118] : Fails on perl >= v5.28.0 with -DDEBUGGING |
| 5 |
The module has been amended to accomodate with a change of |
| 6 |
behaviour of a core macro. |
| 7 |
+ Upd : Contact info. |
| 8 |
|
| 9 |
0.38 2017-11-04 15:40 UTC |
| 10 |
+ Fix : [RT #123374] : Compatibility with CV-in-stash optimisation |
| 11 |
Thanks Father Chrysostomos for reporting and contributing a |
| 12 |
patch. |
| 13 |
|
| 14 |
0.37 2016-07-12 16:40 UTC |
| 15 |
+ Chg : A large chunk of boilerplate XS code, which is also used in |
| 16 |
other XS modules, has been factored out of the main .xs file |
| 17 |
to a collection of .h files in the xsh subdirectory. |
| 18 |
+ Fix : [RT #115392] : Intermittent segfaults with heredocs |
| 19 |
Heredocs should now be handled correctly. |
| 20 |
Thanks Graham Knop for reporting. |
| 21 |
|
| 22 |
0.36 2015-07-17 22:15 UTC |
| 23 |
+ Fix : [RT #104312] : fatal hides perl errors in modules |
| 24 |
no indirect 'fatal' will no longer hide compilation errors |
| 25 |
occurring before indirect constructs. |
| 26 |
Thanks Lukas Mai for reporting. |
| 27 |
|
| 28 |
0.35 2015-04-06 22:20 UTC |
| 29 |
+ Fix : The module could end being disabled in one thread if it was |
| 30 |
first loaded in another thread and that thread was immediately |
| 31 |
terminated. This is now fixed and should address test failures |
| 32 |
of t//09-load-threads.t and t/42-threads-global.t. |
| 33 |
|
| 34 |
0.34 2015-04-02 19:50 UTC |
| 35 |
+ Chg : The new environment variable to enable thread tests on older |
| 36 |
perls is PERL_FORCE_TEST_THREADS. Note that this variable |
| 37 |
should only be turned on by authors. |
| 38 |
+ Fix : [RT #100068] : add link to historical tchrist post |
| 39 |
The link has been added to the documentation. Thanks Olivier |
| 40 |
Mengué for reporting. |
| 41 |
+ Fix : Segfaults when the module is loaded by several threads (or |
| 42 |
Windows emulated processes) ran in parallel. |
| 43 |
+ Fix : Update the Windows ActivePerl + gcc 3.4 workaround for |
| 44 |
ExtUtils::MakeMaker 7.04. Thanks Christian Walde for reporting |
| 45 |
and feedback on this issue. |
| 46 |
+ Fix : Be really compatible with the optional OP_PARENT feature. |
| 47 |
+ Tst : $ENV{$Config{ldlibpthname}} is now preserved on all platforms, |
| 48 |
which will address failures of t/41-threads-teardown.t and |
| 49 |
t/50-external.t with unusual compilers (like icc) that link all |
| 50 |
their compiled objects to their own libraries. |
| 51 |
|
| 52 |
0.33 2014-09-29 20:20 UTC |
| 53 |
+ Fix : [RT #99083] : Breaks eval in an END block in Win32 pseudo-forks. |
| 54 |
Thanks Graham Knop for reporting. |
| 55 |
+ Fix : Segfaults during global destruction of a thread or a |
| 56 |
pseudo-fork. |
| 57 |
|
| 58 |
0.32 2014-09-21 20:15 UTC |
| 59 |
+ Add : Support for the PERL_OP_PARENT optional feature introduced in |
| 60 |
perl 5.21.2. |
| 61 |
+ Fix : [RT #92806] : Tests that use run_perl() fail on Android |
| 62 |
Thanks Brian Fraser for the patch. |
| 63 |
+ Fix : indirect constructs will no longer segfault while inside |
| 64 |
the empty package on perl 5.8.x. This fix may also prevent |
| 65 |
some segfaults during global destruction. |
| 66 |
|
| 67 |
0.31 2013-09-05 16:45 UTC |
| 68 |
+ Fix : [RT #88428] : no indirect in eval can trigger for direct calls |
| 69 |
on __PACKAGE__ |
| 70 |
Thanks Graham Knop for reporting. |
| 71 |
+ Tst : Author tests are no longer bundled with this distribution. |
| 72 |
They are only made available to authors in the git repository. |
| 73 |
|
| 74 |
0.30 2013-05-16 15:55 UTC |
| 75 |
+ Fix : [RT #83806] : false positives with Devel::Declare |
| 76 |
[RT #83839] : false positive using ? : syntax |
| 77 |
Thanks Andrew Main for the patch. |
| 78 |
However, please note that the reason this patch seems to fix |
| 79 |
thinks has not been explained. |
| 80 |
+ Fix : [RT #84649] : incorrect RT link in metadata |
| 81 |
Thanks Karen Etheridge for reporting. |
| 82 |
|
| 83 |
0.29 2013-03-05 01:30 UTC |
| 84 |
+ Fix : [RT #83659] : false positives |
| 85 |
Proper method calls in string-like environments (like |
| 86 |
"@{[ $x->new ]}" will no longer be reported as indirect. |
| 87 |
This was a regression in 0.28. |
| 88 |
Thanks Andrew Main for reporting. |
| 89 |
+ Fix : Broken linkage on Windows with gcc 3.4, which appears in |
| 90 |
particular when using ActivePerl's default compiler suite. |
| 91 |
For those setups, the indirect shared library will now be |
| 92 |
linked against the perl dll directly (instead of the import |
| 93 |
library). |
| 94 |
|
| 95 |
0.28 2013-02-26 17:05 UTC |
| 96 |
+ Fix : [RT #83450] : newlines confuse indirect |
| 97 |
Perl sometimes resets the line buffer between the object and |
| 98 |
the method name (e.g. for "sort Class\n->method" outside of |
| 99 |
eval), and this could cause direct method calls to be reported |
| 100 |
as indirect. |
| 101 |
Thanks Gianni Ceccarelli for reporting. |
| 102 |
+ Fix : Check functions are now replaced and restored in a thread-safe |
| 103 |
manner, either by using the wrap_op_checker() function from perl |
| 104 |
when it is available (starting from perl 5.16) or by taking the |
| 105 |
OP_REFCNT mutex on older perls. |
| 106 |
|
| 107 |
0.27 2013-01-30 19:00 UTC |
| 108 |
+ Fix : [RT #82562] : indirect/Devel::CallParser interaction |
| 109 |
indirect has been taught to play nicely with Devel::CallParser. |
| 110 |
Thanks Andrew Main for the patch. |
| 111 |
+ Tst : Author tests overhaul. |
| 112 |
|
| 113 |
0.26 2011-10-23 14:25 UTC |
| 114 |
+ Add : "no indirect 'global'" enables the pragma for the whole program, |
| 115 |
except for lexical scopes that "use indirect" explicitely. |
| 116 |
+ Chg : Passing both the 'fatal' and 'hook' options to unimport() is |
| 117 |
now considered an error, and will result in an exception. |
| 118 |
unimport() used to consider only the first passed option of |
| 119 |
those two, and silently ignored the other. |
| 120 |
+ Tst : Test failures of t/41-threads-teardown.t and t/50-external.t on |
| 121 |
Cygwin should have been addressed. |
| 122 |
+ Tst : Threads tests will not fail anymore if resources constraints |
| 123 |
prevent the system from creating all the required threads. |
| 124 |
|
| 125 |
0.25 2011-08-24 15:40 UTC |
| 126 |
+ Fix : RT #69291 is now also fixed for perl 5.8. The pragma will no |
| 127 |
longer vivify the "indirect" entry in the hints hash %^H on |
| 128 |
perl 5.8. |
| 129 |
+ Tst : Attempt to make t/50-external.t pass on Cygwin. |
| 130 |
|
| 131 |
0.24 2011-07-17 23:15 UTC |
| 132 |
+ Fix : [RT #64521] : "no indirect" leaking into eval. |
| 133 |
This is currently only fixed for perl 5.10 (perl 5.12 and |
| 134 |
higher were never affected). It was caused by a very stupid |
| 135 |
mistake of mine that was introduced in indirect version 0.23. |
| 136 |
Thanks Michael G Schwern for reporting. |
| 137 |
+ Fix : [RT #69291] : indirect.pm breaks %^H. |
| 138 |
This was caused by the same mistake as for the previous bug, |
| 139 |
and as such it is also only fixed for perl 5.10 (and never |
| 140 |
affected perl 5.12). |
| 141 |
Thanks Andrew Main for reporting. |
| 142 |
+ Doc : C++ compilers are officially NOT supported. |
| 143 |
|
| 144 |
0.23 2010-10-03 00:15 UTC |
| 145 |
+ Fix : Some indirect constructs could be incorrectly reported when |
| 146 |
several modules were used in the same scope. This caused |
| 147 |
t/30-scope.t to fail randomly. |
| 148 |
+ Tst : Threads tests are now only run on perl 5.13.4 and higher. |
| 149 |
They could segfault randomly because of what seems to be an |
| 150 |
internal bug of Perl, which has been addressed in 5.13.4. |
| 151 |
There is also an environment variable that allows you to |
| 152 |
forcefully run those tests, but it should be set only for |
| 153 |
author testing and not for end users. |
| 154 |
|
| 155 |
0.22 2010-08-16 16:00 UTC |
| 156 |
+ Add : Indirect constructs are now reported for code interpolated |
| 157 |
in quote-like environments, like "${\( ... )}", "@{[ ... ]}", |
| 158 |
s/pattern/ ... /e, qr/(?{ ... })/ or qr/(??{ ... })/. |
| 159 |
+ Add : You can now make the pragma lethal by passing anything matching |
| 160 |
/^:?fatal$/i to import(), including "FATAL" and ":Fatal". |
| 161 |
+ Fix : [RT #60378] : segmentation fault on indirect_ck_method. |
| 162 |
This caused constructs like "@{[ $obj->$meth ]}" to segfault |
| 163 |
when $meth was a lexical. |
| 164 |
Thanks Tokuhiro Matsuno for reporting. |
| 165 |
|
| 166 |
0.21 2010-05-31 23:10 UTC |
| 167 |
+ Chg : perl 5.8.1 is now required (instead of 5.8.0). |
| 168 |
+ Fix : [RT #57699] : indirect fail with 64-bit int on 5.13.1. |
| 169 |
It was actually a problem with thread destructors segfaulting |
| 170 |
because they weren't called at the right time anymore. |
| 171 |
Thanks Andrew Main for reporting. |
| 172 |
+ Tst : A few more regression tests about the scope leak bug. |
| 173 |
|
| 174 |
0.20 2010-04-18 21:25 UTC |
| 175 |
+ Fix : [RT #50570] : "indirect" leaking into LWP. |
| 176 |
Thanks Andrew Main for reporting. |
| 177 |
More generally, the require propagation workaround on 5.8-5.10.0 |
| 178 |
has been overhauled, and other scope leaks should be fixed. |
| 179 |
+ Fix : Test failures with 5.12 on Windows where Strawberry Perl crashes |
| 180 |
because the SystemRoot environment variable is missing. |
| 181 |
+ Fix : Work around Kwalitee test misfailures. |
| 182 |
|
| 183 |
0.19 2009-08-28 18:40 UTC |
| 184 |
+ Add : The new constant I_FORKSAFE can be tested to know whether the |
| 185 |
module will behave nicely when fork()ing. It's currently always |
| 186 |
true except on Windows where you need perl 5.10.1 for it to be |
| 187 |
true. |
| 188 |
+ Fix : I_THREADSAFE and I_FORKSAFE ought to be true when |
| 189 |
PERL_INDIRECT_PM_DISABLE is set. |
| 190 |
+ Fix : The pragma could still leak if you passed to the "hook" option |
| 191 |
a reference to the same (named) subroutine from two different |
| 192 |
require scopes. The fix also provides a better solution for |
| 193 |
RT #47902. |
| 194 |
+ Fix : Segfaults when indirect is loaded for the first time from |
| 195 |
inside a thread. |
| 196 |
+ Fix : Leaks of memory associated with the root interpreter. |
| 197 |
+ Opt : Less memory will be used for non-threaded perls version 5.10.0 |
| 198 |
and below, and for threaded perls from version 5.10.1. |
| 199 |
|
| 200 |
0.18 2009-08-23 16:15 UTC |
| 201 |
+ Add : When set, the PERL_INDIRECT_PM_DISABLE environment variable |
| 202 |
disables the pragma globally. |
| 203 |
|
| 204 |
0.17 2009-07-16 12:10 UTC |
| 205 |
+ Fix : [RT #47902] : "no indirect" leaking again. |
| 206 |
This actually turned out to be a bug in how the hook coderefs |
| 207 |
were stored in the hints hash. |
| 208 |
Thanks Andrew Main for reporting once again. |
| 209 |
+ Fix : t/80-regressions.t failing on Windows. |
| 210 |
+ Tst : Yet more cleanups. |
| 211 |
|
| 212 |
0.16 2009-07-14 16:50 UTC |
| 213 |
+ Add : Indirect calls on blocks are now reported. For those, '{' is |
| 214 |
passed to the hook as the object description. |
| 215 |
+ Add : The new indirect::msg() function publicizes the default |
| 216 |
warning/exception message. |
| 217 |
+ Fix : [RT #47866] : Segfault with UTF-8 regexps. |
| 218 |
Thanks Andrew Main for reporting. |
| 219 |
+ Tst : Cleanups. |
| 220 |
|
| 221 |
0.15 2009-07-08 22:55 UTC |
| 222 |
+ Fix : Invalid constructs with the same method and package name were |
| 223 |
not reported. |
| 224 |
+ Fix : The error line number used to point to the end of the expression |
| 225 |
instead of its beginning. |
| 226 |
|
| 227 |
0.14 2009-06-04 21:55 UTC |
| 228 |
+ Fix : Prevent bogus invalid syntaxes caused by reallocated memory |
| 229 |
chunks. Thanks Andrew Main for reporting with a reproducible |
| 230 |
test case. |
| 231 |
|
| 232 |
0.13 2009-05-24 18:50 UTC |
| 233 |
+ Add : The hook now receives the file name and the line where the |
| 234 |
error happened in respectively $_[2] and $_[3]. |
| 235 |
+ Fix : Pass mortalized copies of the method name and object to the |
| 236 |
hook. This seems to fix some rare crashes. |
| 237 |
+ Fix : Work around a bug in perl 5.10.0 and lower. Thanks Andrew Main |
| 238 |
for teaching me about this issue. |
| 239 |
+ Fix : Report the correct file in error messages (a regression from the |
| 240 |
previous version). |
| 241 |
|
| 242 |
0.12 2009-05-03 14:30 UTC |
| 243 |
+ Add : You can specify the action to execute for each indirect |
| 244 |
construct encountered with the new "hook => $coderef" unimport() |
| 245 |
option. |
| 246 |
+ Chg : A ptable is now used internally for the op => position mapping. |
| 247 |
+ Fix : The pragma is now built with thread safety features enabled |
| 248 |
whenever possible (a notable exception is perl 5.8.x on Win32, |
| 249 |
as something seems wrong with its context handling). |
| 250 |
The new indirect::I_THREADSAFE() constant reflects this. |
| 251 |
+ Fix : A negation precedence nit in indirect_ck_entersub(). |
| 252 |
+ Tst : "use/no indirect" while parsing an indirect construct. |
| 253 |
+ Tst : Thread safety. |
| 254 |
|
| 255 |
0.11 2009-02-08 18:35 UTC |
| 256 |
+ Fix : Potential collisions by hashing pointers with a wrong format. |
| 257 |
+ Upd : Resources in META.yml. |
| 258 |
|
| 259 |
0.10 2009-01-17 12:40 UTC |
| 260 |
Re-release 0.09_01 as stable. |
| 261 |
|
| 262 |
0.09_01 2008-12-08 17:55 UTC |
| 263 |
+ Fix : Invalid memory read with "${\(new Foo)}" constructs. The |
| 264 |
corresponding test is turned back on. |
| 265 |
+ Tst : Refinements in t/30-scope.t |
| 266 |
|
| 267 |
0.09 2008-12-05 20:35 UTC |
| 268 |
+ Add : Support for perl 5.8. |
| 269 |
+ Tst : Skip a test in t/10-good.t that randomly segfaults for (I guess) |
| 270 |
systems stricter than linux in the way they manage their memory. |
| 271 |
|
| 272 |
0.08 2008-10-22 14:45 UTC |
| 273 |
+ Fix : A rare edge case for package whose names are prefix of 'main'. |
| 274 |
+ Tst : Test $$ as variable and state variables. |
| 275 |
|
| 276 |
0.07_03 2008-10-17 20:10 UTC |
| 277 |
+ Add : Support and tests for variables with spaces after the sigil. |
| 278 |
+ Upd : META.yml spec updated to 1.4. |
| 279 |
|
| 280 |
0.07_02 2008-10-15 21:10 UTC |
| 281 |
+ Add : Support and tests for package variables. |
| 282 |
+ Tst : Coverage improved by removing dead code. |
| 283 |
|
| 284 |
0.07_01 2008-10-15 16:00 UTC |
| 285 |
+ Fix : [RT #40055] : Not handling RV2SV => GV(SV) correctly, which |
| 286 |
could cause 'no indirect; print' segfaults. Thanks Goro Fuji for |
| 287 |
reporting. |
| 288 |
|
| 289 |
0.06 2008-10-11 16:45 UTC |
| 290 |
+ Doc : Nits. |
| 291 |
+ Tst : Test "no indirect 'anything'", "foo Class->bar", and indirect |
| 292 |
uses of exec() and system(). |
| 293 |
|
| 294 |
0.05 2008-10-02 14:40 UTC |
| 295 |
+ Chg : Now the old check function is always called before storing an op |
| 296 |
into the map. |
| 297 |
+ Fix : Misc code and docs refinements. |
| 298 |
|
| 299 |
0.04 2008-08-30 19:00 UTC |
| 300 |
+ Fix : Clean up the op->src hash when we're done with an entersub. |
| 301 |
+ Tst : No longer fork for testing. IPC::Cmd isn't required anymore. |
| 302 |
|
| 303 |
0.03 2008-08-12 15:25 UTC |
| 304 |
This release is kindly supported by Copenhagen Hotel Centrum WiFi. |
| 305 |
+ Fix : Tests used not to pass PERL5OPTS to their kids. This lead to |
| 306 |
failures under CPAN. I think. |
| 307 |
+ Tst : Refinements. |
| 308 |
|
| 309 |
0.02 2008-08-11 15:55 UTC |
| 310 |
+ Fix : Some building failures with old gcc versions that didn't seem to |
| 311 |
like the ((hint == 2) ? croak : warn)(msg) construct. I think. |
| 312 |
+ Rem : Unused cruft from a previous implementation. |
| 313 |
+ Tst : Fail more gracefully when we can't capture buffers or when the |
| 314 |
child returned an error. |
| 315 |
|
| 316 |
0.01 2008-08-10 20:40 UTC |
| 317 |
First version, released on an unsuspecting world. |
| 318 |
|