1 Revision history for Variable-Magic
3 0.60 2016-09-06 22:30 UTC
4 + Chg : A large chunk of boilerplate XS code, which is also used in
5 other XS modules, has been factored out of the main .xs file
6 to a collection of .h files in the xsh subdirectory.
7 + Fix : [RT #128989] : Bleadperl v5.25.3-266-g1d7e644 breaks V::M.
8 Len magic is no longer called for negative array indices
9 fetches, so the relevant tests are now skipped.
10 Thanks Andreas Koenig for reporting.
12 0.59 2015-10-05 19:05 UTC
13 + Chg : The magic callbacks are now always executed in a new stack.
14 This should have no negative impact (performance-wise or
15 behaviour-wise) on your code, but may actually fix some obscure
17 + Fix : [RT #107294] : Crash during install of Variable-Magic-0.58
18 running "t\18-opinfo.t".
19 A recent change in the core implementation of transliterations
20 caused an undefined behaviour when it interacted with this
21 module, which resulted in visible stack corruption on Win32
22 perls compiled with MSVC when the op_info feature was enabled.
23 This was fixed by using a separate stack for the magic
25 Thanks Peter Stokes for reporting and providing greatly
27 + Fix : The object op_info feature will now handle correctly Unicode
28 transliteration, multideref and custom ops.
30 0.58 2015-07-21 16:00 UTC
31 + Add : If a non-len magic callback returns a reference, it will now
32 only be freed at the end of the statement that caused the
33 magic to trigger. This allows the user to attach free magic
34 (or a plain destructor) to a token returned from the callbacks
35 in order to defer an action after the magic is processed by
37 + Fix : Test failures of threads tests on systems with harsh resource
38 constraints causing the threads to exit() during run.
40 0.57 2015-04-17 15:20 UTC
41 + Chg : The new environment variable to enable thread tests on older
42 perls is PERL_FORCE_TEST_THREADS. Note that this variable
43 should only be turned on by authors.
44 + Fix : Segfaults when the module is loaded by several threads (or
45 Windows emulated processes) ran in parallel.
46 + Fix : Segfaults when the module is loaded in a thread, which spawns
47 itself a new thread, and that child thread outlives its parent.
48 + Fix : Small memory leaks of structures required for thread safety.
49 + Fix : Update the Windows ActivePerl + gcc 3.4 workaround for
50 ExtUtils::MakeMaker 7.04. Thanks Christian Walde for reporting
51 and feedback on this issue.
52 + Tst : The global destruction test will now be exercised on any perl
53 that has DEBUGGING set.
54 + Tst : Optional capturing tests in t/17-ctl.t that were only run when
55 Capture::Tiny was present were converted to an IPC::Open3
56 based helper and will now be run everywhere.
58 0.56 2015-03-11 15:15 UTC
59 + Fix : [RT #101410] : Install fails in blead
60 Even though the change that caused this error was reverted from
61 blead, lvalue uses of ERRSV have been removed from this module
62 so that there will be no breakage when this change is possibly
63 reintroduced in the following months.
64 Thanks Dave Rolsky for reporting.
65 + Tst : $ENV{$Config{ldlibpthname}} is now preserved on all platforms,
66 which will address failures of t/17-ctl.t with unusual
67 compilers (like icc) that link all their compiled objects to
69 + Tst : The global destruction test is now only run on perl 5.13.4 and
70 higher, and only if either Perl::Destruct::Level is installed
71 or PERL_DESTRUCT_LEVEL is set and the perl is a debugging perl.
72 This will solve rare crashes of t/15-self.t on perl 5.13.3 and
75 0.55 2014-10-20 23:45 UTC
76 + Fix : [RT #99319] : t/18-opinfo.t: test failure due to change in
78 The new METHOP op class added in perl 5.21.5 is now supported.
79 Thanks James E Keenan for reporting.
81 0.54 2014-09-22 17:30 UTC
82 + Add : The new constant VMG_COMPAT_CODE_COPY_CLONE evaluates to true
83 if your perl calls 'copy' magic when a magical code prototype
84 is cloned, which is currently the case for perl 5.17.0 and
86 + Fix : [RT #90205] : copy magic on subs puts raw CV in $_[3]
87 $_[3] will now contain a reference to the cloned code when
88 'copy' magic is called for a coderef.
89 Thanks Lukas Mai for reporting.
90 + Fix : t/35-stash.t has been taught about perl 5.21.4.
91 + Fix : Tests using run_perl() in t/17-ctl.t will no longer fail on
94 0.53 2013-09-01 17:50 UTC
95 This is a maintenance release. The code contains no functional change.
96 Satisfied users of version 0.52 can skip this update.
97 + Fix : [RT #86338] : typo fix.
98 Thanks dsteinbrunner@pobox.com for the patch.
99 + Tst : Author tests are no longer bundled with this distribution.
100 They are only made available to authors in the git repository.
102 0.52 2012-11-05 02:30 UTC
103 + Add : The new constant VMG_COMPAT_SCALAR_NOLEN evaluates to true
104 when your perl does not call 'len' magic for scalars, which
105 is the case for perl 5.17.4 and above.
106 This fixes RT #80388.
107 + Tst : t/35-stash.t has been taught about perl 5.17.4.
108 + Tst : Author tests overhaul.
110 0.51 2012-08-18 15:00 UTC
111 + Fix : It is now safe to call dispell() from inside 'free', 'copy' and
112 'uvar' callbacks to dispell the magic currently in use.
113 Thanks Clinton Gormley for reporting.
114 + Fix : Exceptions thrown from inside a 'free' callback are now always
115 consistently propagated outside of the callback. They used to
116 be lost when the 'free' callback was invoked at the end of an
117 eval block or string.
118 + Fix : The 'reset RMG flag' workaroundn used to allow wizards with
119 both 'uvar' and 'clear' magics to be cast onto a hash, has been
122 0.50 2012-06-24 23:00 UTC
123 + Fix : Less memory is leaked when a wizard is freed during global
124 destruction, or when an exception is thrown from a 'free'
126 + Fix : [RT #77991] : t/17-ctl.t fails on perl 5.14 and 5.16.
127 This was actually an issue with ActivePerl, and this test
128 has learned to cope with it.
129 Thanks Gisle Aas for reporting.
130 + Tst : t/35-stash.t has been taught about perl 5.17.1.
131 + Doc : Many clarifications.
133 0.49 2012-06-05 21:40 UTC
134 This is a maintenance release. The code contains no functional change.
135 Satisfied users of version 0.48 can skip this update.
136 + Fix : [RT #77644] : t/17_ctl.t fails in test 44 since 5.17.0.
137 This test has been taught about perl 5.17.0.
138 Thanks Reini Urban for reporting.
139 + Tst : t/99-kwalitee.t will be skipped when only problematic versions
140 of its dependencies are available.
142 0.48 2012-02-17 23:40 UTC
143 + Add : You can now pass a reference to undef as the magic callback in
144 order to install a no-op callback.
145 Thanks Florian Ragwitz for the suggestion.
147 0.47 2011-10-27 16:55 UTC
148 + Add : The new constant VMG_COMPAT_HASH_DELETE_NOUVAR_VOID evaluates
149 to true when "delete $hash{key}" does not call 'delete' uvar
150 magic in void context.
151 + Chg : The MAGIC tokens created by this module no longer use the
152 mg_private member for storing a "magical" signature.
153 + Fix : Triggering magic in a thread on a variable cloned from the main
154 interpreter, and when the wizard already went out of scope,
155 will not segfault anymore.
156 + Opt : The pointer table is no longer needed to ensure thread safety.
157 The size of the object code is about 8% smaller for threaded
159 + Tst : Threads tests will not fail anymore if resources constraints
160 prevent the system from creating all the required threads.
162 0.46 2011-01-23 16:45 UTC
163 + Fix : [RT #64866] : Assertion failure with perl 5.13.9.
164 The real problem was that dispell() wasn't properly resetting
165 the magical sv flags, which became visible in the test suite
166 only with perl 5.13.9.
167 Thanks Joshua ben Jore for reporting.
169 0.45 2010-11-21 23:15 UTC
170 This is a maintenance release. The code contains no functional change.
171 Users of 0.44 can skip this update.
172 + Doc : C++ compilers are officially NOT supported.
173 + Doc : The minimum perl 5.10.0 requirement for uvar magic has been
175 Thanks Peter Rabbitson for pointing this out and contributing a
177 + Tst : Tune for perl 5.13.7.
178 + Tst : Capture::Tiny will be used in t/17-ctl.t if and only if it can
179 capture a simple run.
181 0.44 2010-09-24 19:10 UTC
182 + Fix : Broken linkage on Windows with gcc 3.4, which appears in
183 particular when using ActivePerl's default compiler suite.
184 For those setups, the Variable::Magic shared library will now
185 be linked against the perl dll directly (instead of the import
186 library). This (should) fix RT #51483.
187 Thanks Christian Walde for helping to reproduce this failure
189 + Rem : Support for development perls from the 5.11 branch but older
190 than the 5.11.0 release was removed. This could cause more
191 recent setups to fail.
192 + Tst : Threads tests are now only run on perl 5.13.4 and higher.
193 They could segfault randomly because of what seems to be an
194 internal bug of Perl, which has been addressed in 5.13.4.
195 There is also an environment variable that allows you to
196 forcefully run those tests, but it should be set only for
197 author testing and not for end users.
199 0.43 2010-06-25 23:35 UTC
200 + Add : The new constant VMG_COMPAT_GLOB_GET tells you whether get magic
201 is called for globs. It's true starting perl 5.13.2.
202 + Chg : All callbacks are now called within an eval-like context.
203 Only free callbacks used to be called that way.
204 + Fix : Some exceptions thrown from a free callback could be lost.
205 + Fix : Croak messages could sometimes be repeated several times.
206 + Fix : t/41-clone.t segfaulting with perl 5.13.2.
208 0.42 2010-05-19 00:15 UTC
209 This is a maintenance release. The code contains no functional change.
210 Users of 0.41 can skip this update.
211 + Fix : Test failures with perl 5.13.
212 + Tst : Improve coverage.
214 0.41 2010-03-15 17:35 UTC
215 + Doc : Tweaks and fixups.
217 + Fix : Compatibility with the soon-to-be-released perl 5.12.0.
218 + Fix : Correctly propagate the errors thrown when variable destruction
219 happens at compile-time and not from inside eval STRING.
220 Thanks Florian Ragwitz and Ash Berlin for reporting.
222 0.40 2010-01-06 23:20 UTC
223 + Fix : Possible memory miswrites when passing data arguments to cast().
224 + Fix : Minor C portability tweaks.
226 0.39 2009-12-01 00:05 UTC
227 + Add : You can use a function name as a callback by passing a string
228 reference to wizard() instead of a code reference.
229 + Fix : Compatiblity with perl 5.11.1 and 5.11.2.
230 + Fix : Scalars stored into the data slot no longer leak.
231 + Fix : Thread destruction should not segfault anymore.
232 + Opt : As a result of removing the deprecated features, less memory is
233 used for both threaded (a pointer table is used instead of a
234 hash) and non-threaded (no global structure is needed anymore)
236 + Rem : As advertised in the precendent version, all the signature
237 related features were removed. This includes the 'sig' option
238 to wizard(), the getsig() and gensig() functions, and the SIG_*
241 0.38 2009-10-04 16:10 UTC
242 + Dep : All the signature-related features are DEPRECATED and will be
243 removed in december 2009. This includes the 'sig' option to
244 wizard(), the getsig() and gensig() functions, and the SIG_*
246 Afaik those features were never used in real life, and they make
247 the XS code slow and brittle. If you want your wizards to be
248 globally available, you'll have to handle that yourself by
249 storing them in a hash.
250 + Doc : More examples in a new COOKBOOK section.
251 + Fix : Compatiblity with the 5.11.0 release. The new compatibility
252 constant VMG_COMPAT_ARRAY_PUSH_NOLEN_VOID was added to cover
254 + Fix : Work around Kwalitee test misfailures.
256 0.37 2009-08-25 15:25 UTC
257 + Add : The new constant VMG_FORKSAFE can be tested to know whether the
258 module will behave nicely when fork()ing. It's currently always
259 true except on Windows where you need perl 5.10.1 for it to be
261 + Doc : Nits and clarifications.
263 0.36 2009-07-05 16:30 UTC
264 + Chg : getdata() now returns an empty list when no magic is present.
265 It used to return undef but did not croak as stated in the doc.
266 Thanks Matt S. Trout for pointing this out.
267 + Fix : Building and tests with ActiveState Perl 5.8 build >= 822.
269 0.35 2009-05-15 20:50 UTC
270 This is a maintenance release. Users of 0.34 can skip this update.
271 + Chg : uvar magic used to be enabled for 5.9.5. It now requires 5.10 or
273 + Fix : Silence compilation warnings for cxinc().
274 + Rem : The workaround introduced in 0.31 for the "Unknown errors" issue
275 was removed, as regression tests show that 0.34 provides a
277 + Tst : Skip t/40-threads.t on 5.8.x with old versions of threads and
280 0.34 2009-04-19 16:55 UTC
281 + Fix : Destruction of wizards in require.
282 + Fix : "panic: restartop" in 5.10 when dieing at compile time triggers
285 0.33 2009-03-26 00:00 UTC
286 + Chg : cast() and getsig() now croak too when an invalid signature is
288 + Fix : It was possible to generate with gensig() the same signature
289 twice before actually defining the wizards.
290 + Tst : More stash tests.
291 + Tst : Test when the magic takes place in some cases.
292 + Tst : Improved coverage for the op object feature.
294 0.32 2009-03-01 13:45 UTC
295 + Chg : dispell() and getdata() now croak when an invalid signature is
297 + Doc : More examples and a brand new synopsis.
298 + Fix : The signature is no longer stored in the mg_private member of
299 the MAGIC struct. This fixes possible clashes with magics from
301 + Fix : op info objects weren't blessed into the right class.
302 + Fix : Races when initializing the op names cache under threads.
304 0.31 2009-02-19 23:50 UTC
305 + Doc : A somewhat better introduction. The FUNCTION section was moved
306 before CONSTANTS and HISTORY since it's more important. And
307 'assignation' is really 'assignment', says Ricardo. :)
308 + Fix : The op_info features are now thread safe.
309 + Fix : Segfaults that occured when trying to get the op name during
311 + Fix : Segfaults and wrong "Unknown error" exceptions that happened
312 when dieing in require caused a free callback to fire.
313 Thanks Florian Ragwitz for reporting.
315 0.30 2009-02-12 17:05 UTC
316 + Add : You can now get in all the callbacks either the name or a B::OP
317 object representing the current op if you pass the new 'op_info'
318 option to wizard() with respective values VMG_OP_INFO_NAME and
320 + Doc : Document magic on array/hashes values.
321 + Fix : Completely skip free callbacks during global destruction.
322 + Tst : Tests now pass again on 5.8.3 and lower (they were broken since
324 + Tst : Test magic on symbol table and on array/hash values.
326 0.29 2009-02-08 11:10 UTC
327 + Add : Set $_[2] to the default length for len magic on scalars.
328 + Chg : Perl 5.8 is required.
329 + Chg : The CLONE method will no longer be defined for non-threaded
332 + Fix : The logic for turning off the GMAGICAL flag on hashes when no
333 uvar magic is involved could do nasty things on tied hashes.
334 + Upd : More resources in META.yml.
336 0.28 2009-01-24 17:05 UTC
337 + Add : The 'copy_key' option to wizard(), that enables the safe edition
338 of $_[2] in uvar callbacks to redirect actions to another key.
339 + Fix : Exception objects thrown from callbacks at compile time were not
341 + Fix : free callbacks triggered during global destruction could lead to
342 segfaults because the wizard was already freed.
343 + Fix : Refcount of scalars with free callback.
344 + Fix : Segfaults when using get or uvar magic simultaneously with clear
346 + Doc : The documentation about what the callbacks take and should
347 return is now somewhat clearer.
349 0.27 2009-01-18 23:40 UTC
350 + Fix : Offset of 1 with len magic on scalars.
351 + Fix : Segfaults that seem to happen when one croak in a callback at
352 compile time (thanks Ash Berlin for the testcase).
353 + Upd : Resources in META.yml (Florian Ragwitz)
354 + Tst : Large cleanup resulting in a wider set of tested behaviours.
356 0.26 2008-11-29 22:05 UTC
357 + Fix : Leaks of cloned coderefs that access lexicals.
358 + Fix : Building with maint-5.8 and blead (thanks Andreas Koenig).
360 0.25 2008-11-12 19:40 UTC
361 + Fix : Compatibility with 5.8.9.
362 + Fix : Old Pod::Coverage don't ignore CLONE.
363 + Upd : META.yml spec updated to 1.4.
365 0.24 2008-10-12 14:55 UTC
366 + Fix : Really fix it.
368 0.23 2008-10-11 17:25 UTC
369 + Fix : Building on perls with multiplicity but not ithreads.
371 0.22 2008-09-29 19:40 UTC
372 + Doc : Explain how magic differs from overloading.
373 + Fix : Disable thread safety for perl 5.8 on Windows. If you want it
374 enabled, reliable feedback is welcome.
376 0.21_02 2008-09-28 23:10 UTC
377 + Fix : Hide previously deleted magic tokens in vmg_svt_free (Yuval
380 0.21_01 2008-09-22 13:25 UTC
381 + Chg : Shrink the context data.
382 + Fix : Enable thread safety features for perls that just lack
384 + Tst : Output threads and threads::shared versions in threads tests.
386 0.20 2008-09-20 23:35 UTC
387 + Fix : The module is now thread safe. You can test if it had been
388 built with thread safety features enabled through the
389 VMG_THREADSAFE constant.
390 + Fix : Correct backtrace when wizard() croaks.
391 + Fix : 'cast $a, undef' should throw an 'invalid wizard' error instead
392 of an 'invalid signature'.
394 0.19 2008-07-05 15:05 UTC
395 + Add : The patchlevel with which the module has been built is now
396 exported with the VMG_PERL_PATCHLEVEL constant.
397 + Chg : Version checking macros overhaul.
398 + Fix : 'uvar' magic should only be enabled in maint-5.10 and higher
399 since it's never going to be backported to 5.8.
400 + Fix : Likewise, the copy callback API change should only be enabled in
402 + Fix : Instead of those ackward fake refcount increments in the wizard
403 destructor, flag the wizard with SVf_BREAK and force temps
404 cleanup. This should fix the wizard destruction for all perls,
405 and solves the leak introduced by the previous solution.
407 0.18 2008-07-04 12:20 UTC
408 + Fix : Enable uvar magic when patchlevel >= 28419.
409 + Fix : Set VMG_COMPAT_ARRAY_PUSH_NOLEN for Windows ActiveState Perl
412 0.17 2008-05-11 09:05 UTC
413 + Fix : 5.8.4 and lower don't have mPUSHi. Gah. Yeah, I know, I should
416 0.16 2008-05-10 22:05 UTC
417 + Add : The samples/copy.pl script.
418 + Chg : The sv_magical() + vmg_mg_magical() combo was simplified into
420 + Tst : t/33-code.t was testing scalars, not code.
421 + Tst : is() and like() are better than ok().
422 + Tst : 100% coverage reached.
424 0.15 2008-04-11 18:25 UTC
425 + Chg : Factor vmg_cb_call{1,2,3}() into one function.
426 + Fix : len magic is no longer called when taking the length() of a
427 magical scalar since p32969. The VMG_COMPAT_SCALAR_LENGTH_NOLEN
428 constant was added to cover this.
429 + Tst : More tests for t/22-len.t.
431 0.14 2008-03-24 12:35 UTC
432 + Fix : t/16-huf.t failures on Solaris and FreeBSD caused by not
433 updating mg->mg_ptr after Renew-ing it on dispell.
434 + Fix : Undefining MGf_DUP caused the wizard's magic to be wrongly
435 initialized. MGf_DUP now appears as true from the user point of
436 view, but the dup callback isn't actually set.
437 + Fix : Warnings with blead due to copy callbacks now taking an I32 (API
439 + Fix : vmg_svt_val() and vmg_uvar_del() aren't required if no uvar
441 + Tst : Tests now output the version of optional modules loaded.
443 0.13 2008-03-19 14:35 UTC
444 + Doc : Link to coverage report.
445 + Fix : Correct dependencies listing in META.yml.
446 + Tst : Improved test coverage.
447 + Tst : Print the patchlevel as a comment.
448 + Tst : Use is() where it's relevant.
449 + Tst : t/16-huf.t now really tests interaction with H::U::FH.
451 0.12 2008-02-07 18:15 UTC
452 + Fix : POD error. Thanks to Chris Williams (BinGOs) for the quick
455 0.11 2008-02-07 17:55 UTC
456 + Add : Copy callbacks now receive the current key/index in $_[2].
457 The current value/element is now in $_[3].
458 + Chg : The trigger of copy magic on list assignment with perls greater
459 than 5.10.0 was caused by uvar magic. Hence,
460 VMG_COMPAT_HASH_LISTASSIGN_COPY was removed. Use VMG_UVAR
462 + Fix : Build failures on patched perls.
463 + Tst : Added missing exported symbols to 01-import.t.
465 0.10 2008-02-04 11:30 UTC
466 + Add : New script : samples/vm_vs_tie.pl, that benchmarks our uvar
467 magic versus tied hashes.
468 + Add : The VMG_COMPAT_* constants can be used from userspace to check
469 perl magic abilities.
470 + Fix : Callbacks that returned undef made us croak, breaking the
471 variable behaviour (regression test in 14-callbacks.t).
472 + Fix : uvar callbacks weren't tested for non-NULL-ity before being
473 called (regression test in 28-uvar.t).
474 + Tst : Fix typo in 25-copy.t that prevented Tie::Hash tests to be ran.
476 0.09 2008-02-02 11:30 UTC
477 + Doc : Explicitely say that uvar callbacks are safely ignored for
479 + Doc : Document caveats and fix the usual set of typos.
480 + Fix : vmg_dispell() didn't check if the ext magic were ours when
481 counting wizards that have uvar callbacks, resulting in a
482 possible memory misread.
483 + Fix : getdata() now returns directly the data object, and no longer a
484 copy. This caused a leak.
485 + Tst : Prefix author tests by 9*-.
486 + Tst : New optional author test : 95-portability-files.t, that uses
487 Test::Portability::Files when it's present.
488 + Tst : New test : 14-self.t, that tests application of magic on the
490 + Tst : Move Hash::Util::FieldHash tests out of 11-multiple.t to
493 0.08 2008-02-01 16:55 UTC
494 + Add : copy magic for tied arrays/hashes.
496 + Add : uvar magics : fetch, store, exists, delete for hashes.
499 Internal development versions.
501 0.06 2007-11-20 10:10 UTC
502 + Chg : 5.7.3 is now officially required.
503 + Fix : "data" test failures on 5.8.{0,2}.
504 + Fix : Drand01() vs rand().
506 0.05 2007-11-19 09:10 UTC
507 + Fix : 5.10.0_RC1 compatibility fix.
509 0.04 2007-08-28 12:25 UTC
510 + Chg : s/require (XSLoader)/use $1/.
511 + Fix : Tests are now strict.
512 + Fix : Complete dependencies.
514 0.03 2007-08-01 17:20 UTC
515 + Add : Passing the signature of an already defined magic to wizard()
516 now returns the corresponding magic object.
517 + Add : You can pass the numeric signature as the wizard argument of
518 cast(), dispell() and getdata().
519 + Add : Any argument specified after the wizard (or the signature) in a
520 call to cast() is now passed to the private data constructor in
522 + Chg : $_[0] is now always a reference to the magic variable in all
523 callbacks. The reason for changing from the previous behaviour
524 is that one may want to apply the same magic to a plain scalar
525 and to a scalar reference, and needs a way to distinguish
526 between them in the callback (say, ref()).
527 + Fix : Wizard object destruction used not to free the signature.
529 0.02 2007-07-27 13:50 UTC
530 + Fix : In response to test report 548152 :
531 Newx() and SvMAGIC_set() not present on older perls.
532 + Fix : In response to test report 548275 :
533 Since perl 5.9.5, 'clear' magic is invoked when an array is
534 undefined (bug #43357). Moreover, 'len' magic is no longer
535 called by pushing an element since perl 5.9.3.
536 + Fix : Missing glob test in MANIFEST.
538 0.01 2007-07-25 16:15 UTC
539 First version, released on an unsuspecting world.