1 Revision history for Variable-Magic
3 0.61 2017-01-24 15:30 UTC
4 + Fix : [RT #115792] : Latest blead stack unwind work causes infloop
6 A change in how the core frees temporary values after eval
7 interacted badly with Variable::Magic and caused infinite loops
8 when a free callback was attached to a wizard. This change was
9 temporarily reverted, but is going to be reinstated, and
10 Variable::Magic is now able to play nicely when this happens.
11 Thanks Peter Rabbitson for reporting and David Mitchell for
15 0.60 2016-09-06 22:30 UTC
16 + Chg : A large chunk of boilerplate XS code, which is also used in
17 other XS modules, has been factored out of the main .xs file
18 to a collection of .h files in the xsh subdirectory.
19 + Fix : [RT #128989] : Bleadperl v5.25.3-266-g1d7e644 breaks V::M.
20 Len magic is no longer called for negative array indices
21 fetches, so the relevant tests are now skipped.
22 Thanks Andreas Koenig for reporting.
24 0.59 2015-10-05 19:05 UTC
25 + Chg : The magic callbacks are now always executed in a new stack.
26 This should have no negative impact (performance-wise or
27 behaviour-wise) on your code, but may actually fix some obscure
29 + Fix : [RT #107294] : Crash during install of Variable-Magic-0.58
30 running "t\18-opinfo.t".
31 A recent change in the core implementation of transliterations
32 caused an undefined behaviour when it interacted with this
33 module, which resulted in visible stack corruption on Win32
34 perls compiled with MSVC when the op_info feature was enabled.
35 This was fixed by using a separate stack for the magic
37 Thanks Peter Stokes for reporting and providing greatly
39 + Fix : The object op_info feature will now handle correctly Unicode
40 transliteration, multideref and custom ops.
42 0.58 2015-07-21 16:00 UTC
43 + Add : If a non-len magic callback returns a reference, it will now
44 only be freed at the end of the statement that caused the
45 magic to trigger. This allows the user to attach free magic
46 (or a plain destructor) to a token returned from the callbacks
47 in order to defer an action after the magic is processed by
49 + Fix : Test failures of threads tests on systems with harsh resource
50 constraints causing the threads to exit() during run.
52 0.57 2015-04-17 15:20 UTC
53 + Chg : The new environment variable to enable thread tests on older
54 perls is PERL_FORCE_TEST_THREADS. Note that this variable
55 should only be turned on by authors.
56 + Fix : Segfaults when the module is loaded by several threads (or
57 Windows emulated processes) ran in parallel.
58 + Fix : Segfaults when the module is loaded in a thread, which spawns
59 itself a new thread, and that child thread outlives its parent.
60 + Fix : Small memory leaks of structures required for thread safety.
61 + Fix : Update the Windows ActivePerl + gcc 3.4 workaround for
62 ExtUtils::MakeMaker 7.04. Thanks Christian Walde for reporting
63 and feedback on this issue.
64 + Tst : The global destruction test will now be exercised on any perl
65 that has DEBUGGING set.
66 + Tst : Optional capturing tests in t/17-ctl.t that were only run when
67 Capture::Tiny was present were converted to an IPC::Open3
68 based helper and will now be run everywhere.
70 0.56 2015-03-11 15:15 UTC
71 + Fix : [RT #101410] : Install fails in blead
72 Even though the change that caused this error was reverted from
73 blead, lvalue uses of ERRSV have been removed from this module
74 so that there will be no breakage when this change is possibly
75 reintroduced in the following months.
76 Thanks Dave Rolsky for reporting.
77 + Tst : $ENV{$Config{ldlibpthname}} is now preserved on all platforms,
78 which will address failures of t/17-ctl.t with unusual
79 compilers (like icc) that link all their compiled objects to
81 + Tst : The global destruction test is now only run on perl 5.13.4 and
82 higher, and only if either Perl::Destruct::Level is installed
83 or PERL_DESTRUCT_LEVEL is set and the perl is a debugging perl.
84 This will solve rare crashes of t/15-self.t on perl 5.13.3 and
87 0.55 2014-10-20 23:45 UTC
88 + Fix : [RT #99319] : t/18-opinfo.t: test failure due to change in
90 The new METHOP op class added in perl 5.21.5 is now supported.
91 Thanks James E Keenan for reporting.
93 0.54 2014-09-22 17:30 UTC
94 + Add : The new constant VMG_COMPAT_CODE_COPY_CLONE evaluates to true
95 if your perl calls 'copy' magic when a magical code prototype
96 is cloned, which is currently the case for perl 5.17.0 and
98 + Fix : [RT #90205] : copy magic on subs puts raw CV in $_[3]
99 $_[3] will now contain a reference to the cloned code when
100 'copy' magic is called for a coderef.
101 Thanks Lukas Mai for reporting.
102 + Fix : t/35-stash.t has been taught about perl 5.21.4.
103 + Fix : Tests using run_perl() in t/17-ctl.t will no longer fail on
106 0.53 2013-09-01 17:50 UTC
107 This is a maintenance release. The code contains no functional change.
108 Satisfied users of version 0.52 can skip this update.
109 + Fix : [RT #86338] : typo fix.
110 Thanks dsteinbrunner@pobox.com for the patch.
111 + Tst : Author tests are no longer bundled with this distribution.
112 They are only made available to authors in the git repository.
114 0.52 2012-11-05 02:30 UTC
115 + Add : The new constant VMG_COMPAT_SCALAR_NOLEN evaluates to true
116 when your perl does not call 'len' magic for scalars, which
117 is the case for perl 5.17.4 and above.
118 This fixes RT #80388.
119 + Tst : t/35-stash.t has been taught about perl 5.17.4.
120 + Tst : Author tests overhaul.
122 0.51 2012-08-18 15:00 UTC
123 + Fix : It is now safe to call dispell() from inside 'free', 'copy' and
124 'uvar' callbacks to dispell the magic currently in use.
125 Thanks Clinton Gormley for reporting.
126 + Fix : Exceptions thrown from inside a 'free' callback are now always
127 consistently propagated outside of the callback. They used to
128 be lost when the 'free' callback was invoked at the end of an
129 eval block or string.
130 + Fix : The 'reset RMG flag' workaroundn used to allow wizards with
131 both 'uvar' and 'clear' magics to be cast onto a hash, has been
134 0.50 2012-06-24 23:00 UTC
135 + Fix : Less memory is leaked when a wizard is freed during global
136 destruction, or when an exception is thrown from a 'free'
138 + Fix : [RT #77991] : t/17-ctl.t fails on perl 5.14 and 5.16.
139 This was actually an issue with ActivePerl, and this test
140 has learned to cope with it.
141 Thanks Gisle Aas for reporting.
142 + Tst : t/35-stash.t has been taught about perl 5.17.1.
143 + Doc : Many clarifications.
145 0.49 2012-06-05 21:40 UTC
146 This is a maintenance release. The code contains no functional change.
147 Satisfied users of version 0.48 can skip this update.
148 + Fix : [RT #77644] : t/17_ctl.t fails in test 44 since 5.17.0.
149 This test has been taught about perl 5.17.0.
150 Thanks Reini Urban for reporting.
151 + Tst : t/99-kwalitee.t will be skipped when only problematic versions
152 of its dependencies are available.
154 0.48 2012-02-17 23:40 UTC
155 + Add : You can now pass a reference to undef as the magic callback in
156 order to install a no-op callback.
157 Thanks Florian Ragwitz for the suggestion.
159 0.47 2011-10-27 16:55 UTC
160 + Add : The new constant VMG_COMPAT_HASH_DELETE_NOUVAR_VOID evaluates
161 to true when "delete $hash{key}" does not call 'delete' uvar
162 magic in void context.
163 + Chg : The MAGIC tokens created by this module no longer use the
164 mg_private member for storing a "magical" signature.
165 + Fix : Triggering magic in a thread on a variable cloned from the main
166 interpreter, and when the wizard already went out of scope,
167 will not segfault anymore.
168 + Opt : The pointer table is no longer needed to ensure thread safety.
169 The size of the object code is about 8% smaller for threaded
171 + Tst : Threads tests will not fail anymore if resources constraints
172 prevent the system from creating all the required threads.
174 0.46 2011-01-23 16:45 UTC
175 + Fix : [RT #64866] : Assertion failure with perl 5.13.9.
176 The real problem was that dispell() wasn't properly resetting
177 the magical sv flags, which became visible in the test suite
178 only with perl 5.13.9.
179 Thanks Joshua ben Jore for reporting.
181 0.45 2010-11-21 23:15 UTC
182 This is a maintenance release. The code contains no functional change.
183 Users of 0.44 can skip this update.
184 + Doc : C++ compilers are officially NOT supported.
185 + Doc : The minimum perl 5.10.0 requirement for uvar magic has been
187 Thanks Peter Rabbitson for pointing this out and contributing a
189 + Tst : Tune for perl 5.13.7.
190 + Tst : Capture::Tiny will be used in t/17-ctl.t if and only if it can
191 capture a simple run.
193 0.44 2010-09-24 19:10 UTC
194 + Fix : Broken linkage on Windows with gcc 3.4, which appears in
195 particular when using ActivePerl's default compiler suite.
196 For those setups, the Variable::Magic shared library will now
197 be linked against the perl dll directly (instead of the import
198 library). This (should) fix RT #51483.
199 Thanks Christian Walde for helping to reproduce this failure
201 + Rem : Support for development perls from the 5.11 branch but older
202 than the 5.11.0 release was removed. This could cause more
203 recent setups to fail.
204 + Tst : Threads tests are now only run on perl 5.13.4 and higher.
205 They could segfault randomly because of what seems to be an
206 internal bug of Perl, which has been addressed in 5.13.4.
207 There is also an environment variable that allows you to
208 forcefully run those tests, but it should be set only for
209 author testing and not for end users.
211 0.43 2010-06-25 23:35 UTC
212 + Add : The new constant VMG_COMPAT_GLOB_GET tells you whether get magic
213 is called for globs. It's true starting perl 5.13.2.
214 + Chg : All callbacks are now called within an eval-like context.
215 Only free callbacks used to be called that way.
216 + Fix : Some exceptions thrown from a free callback could be lost.
217 + Fix : Croak messages could sometimes be repeated several times.
218 + Fix : t/41-clone.t segfaulting with perl 5.13.2.
220 0.42 2010-05-19 00:15 UTC
221 This is a maintenance release. The code contains no functional change.
222 Users of 0.41 can skip this update.
223 + Fix : Test failures with perl 5.13.
224 + Tst : Improve coverage.
226 0.41 2010-03-15 17:35 UTC
227 + Doc : Tweaks and fixups.
229 + Fix : Compatibility with the soon-to-be-released perl 5.12.0.
230 + Fix : Correctly propagate the errors thrown when variable destruction
231 happens at compile-time and not from inside eval STRING.
232 Thanks Florian Ragwitz and Ash Berlin for reporting.
234 0.40 2010-01-06 23:20 UTC
235 + Fix : Possible memory miswrites when passing data arguments to cast().
236 + Fix : Minor C portability tweaks.
238 0.39 2009-12-01 00:05 UTC
239 + Add : You can use a function name as a callback by passing a string
240 reference to wizard() instead of a code reference.
241 + Fix : Compatiblity with perl 5.11.1 and 5.11.2.
242 + Fix : Scalars stored into the data slot no longer leak.
243 + Fix : Thread destruction should not segfault anymore.
244 + Opt : As a result of removing the deprecated features, less memory is
245 used for both threaded (a pointer table is used instead of a
246 hash) and non-threaded (no global structure is needed anymore)
248 + Rem : As advertised in the precendent version, all the signature
249 related features were removed. This includes the 'sig' option
250 to wizard(), the getsig() and gensig() functions, and the SIG_*
253 0.38 2009-10-04 16:10 UTC
254 + Dep : All the signature-related features are DEPRECATED and will be
255 removed in december 2009. This includes the 'sig' option to
256 wizard(), the getsig() and gensig() functions, and the SIG_*
258 Afaik those features were never used in real life, and they make
259 the XS code slow and brittle. If you want your wizards to be
260 globally available, you'll have to handle that yourself by
261 storing them in a hash.
262 + Doc : More examples in a new COOKBOOK section.
263 + Fix : Compatiblity with the 5.11.0 release. The new compatibility
264 constant VMG_COMPAT_ARRAY_PUSH_NOLEN_VOID was added to cover
266 + Fix : Work around Kwalitee test misfailures.
268 0.37 2009-08-25 15:25 UTC
269 + Add : The new constant VMG_FORKSAFE can be tested to know whether the
270 module will behave nicely when fork()ing. It's currently always
271 true except on Windows where you need perl 5.10.1 for it to be
273 + Doc : Nits and clarifications.
275 0.36 2009-07-05 16:30 UTC
276 + Chg : getdata() now returns an empty list when no magic is present.
277 It used to return undef but did not croak as stated in the doc.
278 Thanks Matt S. Trout for pointing this out.
279 + Fix : Building and tests with ActiveState Perl 5.8 build >= 822.
281 0.35 2009-05-15 20:50 UTC
282 This is a maintenance release. Users of 0.34 can skip this update.
283 + Chg : uvar magic used to be enabled for 5.9.5. It now requires 5.10 or
285 + Fix : Silence compilation warnings for cxinc().
286 + Rem : The workaround introduced in 0.31 for the "Unknown errors" issue
287 was removed, as regression tests show that 0.34 provides a
289 + Tst : Skip t/40-threads.t on 5.8.x with old versions of threads and
292 0.34 2009-04-19 16:55 UTC
293 + Fix : Destruction of wizards in require.
294 + Fix : "panic: restartop" in 5.10 when dieing at compile time triggers
297 0.33 2009-03-26 00:00 UTC
298 + Chg : cast() and getsig() now croak too when an invalid signature is
300 + Fix : It was possible to generate with gensig() the same signature
301 twice before actually defining the wizards.
302 + Tst : More stash tests.
303 + Tst : Test when the magic takes place in some cases.
304 + Tst : Improved coverage for the op object feature.
306 0.32 2009-03-01 13:45 UTC
307 + Chg : dispell() and getdata() now croak when an invalid signature is
309 + Doc : More examples and a brand new synopsis.
310 + Fix : The signature is no longer stored in the mg_private member of
311 the MAGIC struct. This fixes possible clashes with magics from
313 + Fix : op info objects weren't blessed into the right class.
314 + Fix : Races when initializing the op names cache under threads.
316 0.31 2009-02-19 23:50 UTC
317 + Doc : A somewhat better introduction. The FUNCTION section was moved
318 before CONSTANTS and HISTORY since it's more important. And
319 'assignation' is really 'assignment', says Ricardo. :)
320 + Fix : The op_info features are now thread safe.
321 + Fix : Segfaults that occured when trying to get the op name during
323 + Fix : Segfaults and wrong "Unknown error" exceptions that happened
324 when dieing in require caused a free callback to fire.
325 Thanks Florian Ragwitz for reporting.
327 0.30 2009-02-12 17:05 UTC
328 + Add : You can now get in all the callbacks either the name or a B::OP
329 object representing the current op if you pass the new 'op_info'
330 option to wizard() with respective values VMG_OP_INFO_NAME and
332 + Doc : Document magic on array/hashes values.
333 + Fix : Completely skip free callbacks during global destruction.
334 + Tst : Tests now pass again on 5.8.3 and lower (they were broken since
336 + Tst : Test magic on symbol table and on array/hash values.
338 0.29 2009-02-08 11:10 UTC
339 + Add : Set $_[2] to the default length for len magic on scalars.
340 + Chg : Perl 5.8 is required.
341 + Chg : The CLONE method will no longer be defined for non-threaded
344 + Fix : The logic for turning off the GMAGICAL flag on hashes when no
345 uvar magic is involved could do nasty things on tied hashes.
346 + Upd : More resources in META.yml.
348 0.28 2009-01-24 17:05 UTC
349 + Add : The 'copy_key' option to wizard(), that enables the safe edition
350 of $_[2] in uvar callbacks to redirect actions to another key.
351 + Fix : Exception objects thrown from callbacks at compile time were not
353 + Fix : free callbacks triggered during global destruction could lead to
354 segfaults because the wizard was already freed.
355 + Fix : Refcount of scalars with free callback.
356 + Fix : Segfaults when using get or uvar magic simultaneously with clear
358 + Doc : The documentation about what the callbacks take and should
359 return is now somewhat clearer.
361 0.27 2009-01-18 23:40 UTC
362 + Fix : Offset of 1 with len magic on scalars.
363 + Fix : Segfaults that seem to happen when one croak in a callback at
364 compile time (thanks Ash Berlin for the testcase).
365 + Upd : Resources in META.yml (Florian Ragwitz)
366 + Tst : Large cleanup resulting in a wider set of tested behaviours.
368 0.26 2008-11-29 22:05 UTC
369 + Fix : Leaks of cloned coderefs that access lexicals.
370 + Fix : Building with maint-5.8 and blead (thanks Andreas Koenig).
372 0.25 2008-11-12 19:40 UTC
373 + Fix : Compatibility with 5.8.9.
374 + Fix : Old Pod::Coverage don't ignore CLONE.
375 + Upd : META.yml spec updated to 1.4.
377 0.24 2008-10-12 14:55 UTC
378 + Fix : Really fix it.
380 0.23 2008-10-11 17:25 UTC
381 + Fix : Building on perls with multiplicity but not ithreads.
383 0.22 2008-09-29 19:40 UTC
384 + Doc : Explain how magic differs from overloading.
385 + Fix : Disable thread safety for perl 5.8 on Windows. If you want it
386 enabled, reliable feedback is welcome.
388 0.21_02 2008-09-28 23:10 UTC
389 + Fix : Hide previously deleted magic tokens in vmg_svt_free (Yuval
392 0.21_01 2008-09-22 13:25 UTC
393 + Chg : Shrink the context data.
394 + Fix : Enable thread safety features for perls that just lack
396 + Tst : Output threads and threads::shared versions in threads tests.
398 0.20 2008-09-20 23:35 UTC
399 + Fix : The module is now thread safe. You can test if it had been
400 built with thread safety features enabled through the
401 VMG_THREADSAFE constant.
402 + Fix : Correct backtrace when wizard() croaks.
403 + Fix : 'cast $a, undef' should throw an 'invalid wizard' error instead
404 of an 'invalid signature'.
406 0.19 2008-07-05 15:05 UTC
407 + Add : The patchlevel with which the module has been built is now
408 exported with the VMG_PERL_PATCHLEVEL constant.
409 + Chg : Version checking macros overhaul.
410 + Fix : 'uvar' magic should only be enabled in maint-5.10 and higher
411 since it's never going to be backported to 5.8.
412 + Fix : Likewise, the copy callback API change should only be enabled in
414 + Fix : Instead of those ackward fake refcount increments in the wizard
415 destructor, flag the wizard with SVf_BREAK and force temps
416 cleanup. This should fix the wizard destruction for all perls,
417 and solves the leak introduced by the previous solution.
419 0.18 2008-07-04 12:20 UTC
420 + Fix : Enable uvar magic when patchlevel >= 28419.
421 + Fix : Set VMG_COMPAT_ARRAY_PUSH_NOLEN for Windows ActiveState Perl
424 0.17 2008-05-11 09:05 UTC
425 + Fix : 5.8.4 and lower don't have mPUSHi. Gah. Yeah, I know, I should
428 0.16 2008-05-10 22:05 UTC
429 + Add : The samples/copy.pl script.
430 + Chg : The sv_magical() + vmg_mg_magical() combo was simplified into
432 + Tst : t/33-code.t was testing scalars, not code.
433 + Tst : is() and like() are better than ok().
434 + Tst : 100% coverage reached.
436 0.15 2008-04-11 18:25 UTC
437 + Chg : Factor vmg_cb_call{1,2,3}() into one function.
438 + Fix : len magic is no longer called when taking the length() of a
439 magical scalar since p32969. The VMG_COMPAT_SCALAR_LENGTH_NOLEN
440 constant was added to cover this.
441 + Tst : More tests for t/22-len.t.
443 0.14 2008-03-24 12:35 UTC
444 + Fix : t/16-huf.t failures on Solaris and FreeBSD caused by not
445 updating mg->mg_ptr after Renew-ing it on dispell.
446 + Fix : Undefining MGf_DUP caused the wizard's magic to be wrongly
447 initialized. MGf_DUP now appears as true from the user point of
448 view, but the dup callback isn't actually set.
449 + Fix : Warnings with blead due to copy callbacks now taking an I32 (API
451 + Fix : vmg_svt_val() and vmg_uvar_del() aren't required if no uvar
453 + Tst : Tests now output the version of optional modules loaded.
455 0.13 2008-03-19 14:35 UTC
456 + Doc : Link to coverage report.
457 + Fix : Correct dependencies listing in META.yml.
458 + Tst : Improved test coverage.
459 + Tst : Print the patchlevel as a comment.
460 + Tst : Use is() where it's relevant.
461 + Tst : t/16-huf.t now really tests interaction with H::U::FH.
463 0.12 2008-02-07 18:15 UTC
464 + Fix : POD error. Thanks to Chris Williams (BinGOs) for the quick
467 0.11 2008-02-07 17:55 UTC
468 + Add : Copy callbacks now receive the current key/index in $_[2].
469 The current value/element is now in $_[3].
470 + Chg : The trigger of copy magic on list assignment with perls greater
471 than 5.10.0 was caused by uvar magic. Hence,
472 VMG_COMPAT_HASH_LISTASSIGN_COPY was removed. Use VMG_UVAR
474 + Fix : Build failures on patched perls.
475 + Tst : Added missing exported symbols to 01-import.t.
477 0.10 2008-02-04 11:30 UTC
478 + Add : New script : samples/vm_vs_tie.pl, that benchmarks our uvar
479 magic versus tied hashes.
480 + Add : The VMG_COMPAT_* constants can be used from userspace to check
481 perl magic abilities.
482 + Fix : Callbacks that returned undef made us croak, breaking the
483 variable behaviour (regression test in 14-callbacks.t).
484 + Fix : uvar callbacks weren't tested for non-NULL-ity before being
485 called (regression test in 28-uvar.t).
486 + Tst : Fix typo in 25-copy.t that prevented Tie::Hash tests to be ran.
488 0.09 2008-02-02 11:30 UTC
489 + Doc : Explicitely say that uvar callbacks are safely ignored for
491 + Doc : Document caveats and fix the usual set of typos.
492 + Fix : vmg_dispell() didn't check if the ext magic were ours when
493 counting wizards that have uvar callbacks, resulting in a
494 possible memory misread.
495 + Fix : getdata() now returns directly the data object, and no longer a
496 copy. This caused a leak.
497 + Tst : Prefix author tests by 9*-.
498 + Tst : New optional author test : 95-portability-files.t, that uses
499 Test::Portability::Files when it's present.
500 + Tst : New test : 14-self.t, that tests application of magic on the
502 + Tst : Move Hash::Util::FieldHash tests out of 11-multiple.t to
505 0.08 2008-02-01 16:55 UTC
506 + Add : copy magic for tied arrays/hashes.
508 + Add : uvar magics : fetch, store, exists, delete for hashes.
511 Internal development versions.
513 0.06 2007-11-20 10:10 UTC
514 + Chg : 5.7.3 is now officially required.
515 + Fix : "data" test failures on 5.8.{0,2}.
516 + Fix : Drand01() vs rand().
518 0.05 2007-11-19 09:10 UTC
519 + Fix : 5.10.0_RC1 compatibility fix.
521 0.04 2007-08-28 12:25 UTC
522 + Chg : s/require (XSLoader)/use $1/.
523 + Fix : Tests are now strict.
524 + Fix : Complete dependencies.
526 0.03 2007-08-01 17:20 UTC
527 + Add : Passing the signature of an already defined magic to wizard()
528 now returns the corresponding magic object.
529 + Add : You can pass the numeric signature as the wizard argument of
530 cast(), dispell() and getdata().
531 + Add : Any argument specified after the wizard (or the signature) in a
532 call to cast() is now passed to the private data constructor in
534 + Chg : $_[0] is now always a reference to the magic variable in all
535 callbacks. The reason for changing from the previous behaviour
536 is that one may want to apply the same magic to a plain scalar
537 and to a scalar reference, and needs a way to distinguish
538 between them in the callback (say, ref()).
539 + Fix : Wizard object destruction used not to free the signature.
541 0.02 2007-07-27 13:50 UTC
542 + Fix : In response to test report 548152 :
543 Newx() and SvMAGIC_set() not present on older perls.
544 + Fix : In response to test report 548275 :
545 Since perl 5.9.5, 'clear' magic is invoked when an array is
546 undefined (bug #43357). Moreover, 'len' magic is no longer
547 called by pushing an element since perl 5.9.3.
548 + Fix : Missing glob test in MANIFEST.
550 0.01 2007-07-25 16:15 UTC
551 First version, released on an unsuspecting world.