1 Revision history for Variable-Magic
3 0.59 2015-10-05 19:05 UTC
4 + Chg : The magic callbacks are now always executed in a new stack.
5 This should have no negative impact (performance-wise or
6 behaviour-wise) on your code, but may actually fix some obscure
8 + Fix : [RT #107294] : Crash during install of Variable-Magic-0.58
9 running "t\18-opinfo.t".
10 A recent change in the core implementation of transliterations
11 caused an undefined behaviour when it interacted with this
12 module, which resulted in visible stack corruption on Win32
13 perls compiled with MSVC when the op_info feature was enabled.
14 This was fixed by using a separate stack for the magic
16 Thanks Peter Stokes for reporting and providing greatly
18 + Fix : The object op_info feature will now handle correctly Unicode
19 transliteration, multideref and custom ops.
21 0.58 2015-07-21 16:00 UTC
22 + Add : If a non-len magic callback returns a reference, it will now
23 only be freed at the end of the statement that caused the
24 magic to trigger. This allows the user to attach free magic
25 (or a plain destructor) to a token returned from the callbacks
26 in order to defer an action after the magic is processed by
28 + Fix : Test failures of threads tests on systems with harsh resource
29 constraints causing the threads to exit() during run.
31 0.57 2015-04-17 15:20 UTC
32 + Chg : The new environment variable to enable thread tests on older
33 perls is PERL_FORCE_TEST_THREADS. Note that this variable
34 should only be turned on by authors.
35 + Fix : Segfaults when the module is loaded by several threads (or
36 Windows emulated processes) ran in parallel.
37 + Fix : Segfaults when the module is loaded in a thread, which spawns
38 itself a new thread, and that child thread outlives its parent.
39 + Fix : Small memory leaks of structures required for thread safety.
40 + Fix : Update the Windows ActivePerl + gcc 3.4 workaround for
41 ExtUtils::MakeMaker 7.04. Thanks Christian Walde for reporting
42 and feedback on this issue.
43 + Tst : The global destruction test will now be exercised on any perl
44 that has DEBUGGING set.
45 + Tst : Optional capturing tests in t/17-ctl.t that were only run when
46 Capture::Tiny was present were converted to an IPC::Open3
47 based helper and will now be run everywhere.
49 0.56 2015-03-11 15:15 UTC
50 + Fix : [RT #101410] : Install fails in blead
51 Even though the change that caused this error was reverted from
52 blead, lvalue uses of ERRSV have been removed from this module
53 so that there will be no breakage when this change is possibly
54 reintroduced in the following months.
55 Thanks Dave Rolsky for reporting.
56 + Tst : $ENV{$Config{ldlibpthname}} is now preserved on all platforms,
57 which will address failures of t/17-ctl.t with unusual
58 compilers (like icc) that link all their compiled objects to
60 + Tst : The global destruction test is now only run on perl 5.13.4 and
61 higher, and only if either Perl::Destruct::Level is installed
62 or PERL_DESTRUCT_LEVEL is set and the perl is a debugging perl.
63 This will solve rare crashes of t/15-self.t on perl 5.13.3 and
66 0.55 2014-10-20 23:45 UTC
67 + Fix : [RT #99319] : t/18-opinfo.t: test failure due to change in
69 The new METHOP op class added in perl 5.21.5 is now supported.
70 Thanks James E Keenan for reporting.
72 0.54 2014-09-22 17:30 UTC
73 + Add : The new constant VMG_COMPAT_CODE_COPY_CLONE evaluates to true
74 if your perl calls 'copy' magic when a magical code prototype
75 is cloned, which is currently the case for perl 5.17.0 and
77 + Fix : [RT #90205] : copy magic on subs puts raw CV in $_[3]
78 $_[3] will now contain a reference to the cloned code when
79 'copy' magic is called for a coderef.
80 Thanks Lukas Mai for reporting.
81 + Fix : t/35-stash.t has been taught about perl 5.21.4.
82 + Fix : Tests using run_perl() in t/17-ctl.t will no longer fail on
85 0.53 2013-09-01 17:50 UTC
86 This is a maintenance release. The code contains no functional change.
87 Satisfied users of version 0.52 can skip this update.
88 + Fix : [RT #86338] : typo fix.
89 Thanks dsteinbrunner@pobox.com for the patch.
90 + Tst : Author tests are no longer bundled with this distribution.
91 They are only made available to authors in the git repository.
93 0.52 2012-11-05 02:30 UTC
94 + Add : The new constant VMG_COMPAT_SCALAR_NOLEN evaluates to true
95 when your perl does not call 'len' magic for scalars, which
96 is the case for perl 5.17.4 and above.
98 + Tst : t/35-stash.t has been taught about perl 5.17.4.
99 + Tst : Author tests overhaul.
101 0.51 2012-08-18 15:00 UTC
102 + Fix : It is now safe to call dispell() from inside 'free', 'copy' and
103 'uvar' callbacks to dispell the magic currently in use.
104 Thanks Clinton Gormley for reporting.
105 + Fix : Exceptions thrown from inside a 'free' callback are now always
106 consistently propagated outside of the callback. They used to
107 be lost when the 'free' callback was invoked at the end of an
108 eval block or string.
109 + Fix : The 'reset RMG flag' workaroundn used to allow wizards with
110 both 'uvar' and 'clear' magics to be cast onto a hash, has been
113 0.50 2012-06-24 23:00 UTC
114 + Fix : Less memory is leaked when a wizard is freed during global
115 destruction, or when an exception is thrown from a 'free'
117 + Fix : [RT #77991] : t/17-ctl.t fails on perl 5.14 and 5.16.
118 This was actually an issue with ActivePerl, and this test
119 has learned to cope with it.
120 Thanks Gisle Aas for reporting.
121 + Tst : t/35-stash.t has been taught about perl 5.17.1.
122 + Doc : Many clarifications.
124 0.49 2012-06-05 21:40 UTC
125 This is a maintenance release. The code contains no functional change.
126 Satisfied users of version 0.48 can skip this update.
127 + Fix : [RT #77644] : t/17_ctl.t fails in test 44 since 5.17.0.
128 This test has been taught about perl 5.17.0.
129 Thanks Reini Urban for reporting.
130 + Tst : t/99-kwalitee.t will be skipped when only problematic versions
131 of its dependencies are available.
133 0.48 2012-02-17 23:40 UTC
134 + Add : You can now pass a reference to undef as the magic callback in
135 order to install a no-op callback.
136 Thanks Florian Ragwitz for the suggestion.
138 0.47 2011-10-27 16:55 UTC
139 + Add : The new constant VMG_COMPAT_HASH_DELETE_NOUVAR_VOID evaluates
140 to true when "delete $hash{key}" does not call 'delete' uvar
141 magic in void context.
142 + Chg : The MAGIC tokens created by this module no longer use the
143 mg_private member for storing a "magical" signature.
144 + Fix : Triggering magic in a thread on a variable cloned from the main
145 interpreter, and when the wizard already went out of scope,
146 will not segfault anymore.
147 + Opt : The pointer table is no longer needed to ensure thread safety.
148 The size of the object code is about 8% smaller for threaded
150 + Tst : Threads tests will not fail anymore if resources constraints
151 prevent the system from creating all the required threads.
153 0.46 2011-01-23 16:45 UTC
154 + Fix : [RT #64866] : Assertion failure with perl 5.13.9.
155 The real problem was that dispell() wasn't properly resetting
156 the magical sv flags, which became visible in the test suite
157 only with perl 5.13.9.
158 Thanks Joshua ben Jore for reporting.
160 0.45 2010-11-21 23:15 UTC
161 This is a maintenance release. The code contains no functional change.
162 Users of 0.44 can skip this update.
163 + Doc : C++ compilers are officially NOT supported.
164 + Doc : The minimum perl 5.10.0 requirement for uvar magic has been
166 Thanks Peter Rabbitson for pointing this out and contributing a
168 + Tst : Tune for perl 5.13.7.
169 + Tst : Capture::Tiny will be used in t/17-ctl.t if and only if it can
170 capture a simple run.
172 0.44 2010-09-24 19:10 UTC
173 + Fix : Broken linkage on Windows with gcc 3.4, which appears in
174 particular when using ActivePerl's default compiler suite.
175 For those setups, the Variable::Magic shared library will now
176 be linked against the perl dll directly (instead of the import
177 library). This (should) fix RT #51483.
178 Thanks Christian Walde for helping to reproduce this failure
180 + Rem : Support for development perls from the 5.11 branch but older
181 than the 5.11.0 release was removed. This could cause more
182 recent setups to fail.
183 + Tst : Threads tests are now only run on perl 5.13.4 and higher.
184 They could segfault randomly because of what seems to be an
185 internal bug of Perl, which has been addressed in 5.13.4.
186 There is also an environment variable that allows you to
187 forcefully run those tests, but it should be set only for
188 author testing and not for end users.
190 0.43 2010-06-25 23:35 UTC
191 + Add : The new constant VMG_COMPAT_GLOB_GET tells you whether get magic
192 is called for globs. It's true starting perl 5.13.2.
193 + Chg : All callbacks are now called within an eval-like context.
194 Only free callbacks used to be called that way.
195 + Fix : Some exceptions thrown from a free callback could be lost.
196 + Fix : Croak messages could sometimes be repeated several times.
197 + Fix : t/41-clone.t segfaulting with perl 5.13.2.
199 0.42 2010-05-19 00:15 UTC
200 This is a maintenance release. The code contains no functional change.
201 Users of 0.41 can skip this update.
202 + Fix : Test failures with perl 5.13.
203 + Tst : Improve coverage.
205 0.41 2010-03-15 17:35 UTC
206 + Doc : Tweaks and fixups.
208 + Fix : Compatibility with the soon-to-be-released perl 5.12.0.
209 + Fix : Correctly propagate the errors thrown when variable destruction
210 happens at compile-time and not from inside eval STRING.
211 Thanks Florian Ragwitz and Ash Berlin for reporting.
213 0.40 2010-01-06 23:20 UTC
214 + Fix : Possible memory miswrites when passing data arguments to cast().
215 + Fix : Minor C portability tweaks.
217 0.39 2009-12-01 00:05 UTC
218 + Add : You can use a function name as a callback by passing a string
219 reference to wizard() instead of a code reference.
220 + Fix : Compatiblity with perl 5.11.1 and 5.11.2.
221 + Fix : Scalars stored into the data slot no longer leak.
222 + Fix : Thread destruction should not segfault anymore.
223 + Opt : As a result of removing the deprecated features, less memory is
224 used for both threaded (a pointer table is used instead of a
225 hash) and non-threaded (no global structure is needed anymore)
227 + Rem : As advertised in the precendent version, all the signature
228 related features were removed. This includes the 'sig' option
229 to wizard(), the getsig() and gensig() functions, and the SIG_*
232 0.38 2009-10-04 16:10 UTC
233 + Dep : All the signature-related features are DEPRECATED and will be
234 removed in december 2009. This includes the 'sig' option to
235 wizard(), the getsig() and gensig() functions, and the SIG_*
237 Afaik those features were never used in real life, and they make
238 the XS code slow and brittle. If you want your wizards to be
239 globally available, you'll have to handle that yourself by
240 storing them in a hash.
241 + Doc : More examples in a new COOKBOOK section.
242 + Fix : Compatiblity with the 5.11.0 release. The new compatibility
243 constant VMG_COMPAT_ARRAY_PUSH_NOLEN_VOID was added to cover
245 + Fix : Work around Kwalitee test misfailures.
247 0.37 2009-08-25 15:25 UTC
248 + Add : The new constant VMG_FORKSAFE can be tested to know whether the
249 module will behave nicely when fork()ing. It's currently always
250 true except on Windows where you need perl 5.10.1 for it to be
252 + Doc : Nits and clarifications.
254 0.36 2009-07-05 16:30 UTC
255 + Chg : getdata() now returns an empty list when no magic is present.
256 It used to return undef but did not croak as stated in the doc.
257 Thanks Matt S. Trout for pointing this out.
258 + Fix : Building and tests with ActiveState Perl 5.8 build >= 822.
260 0.35 2009-05-15 20:50 UTC
261 This is a maintenance release. Users of 0.34 can skip this update.
262 + Chg : uvar magic used to be enabled for 5.9.5. It now requires 5.10 or
264 + Fix : Silence compilation warnings for cxinc().
265 + Rem : The workaround introduced in 0.31 for the "Unknown errors" issue
266 was removed, as regression tests show that 0.34 provides a
268 + Tst : Skip t/40-threads.t on 5.8.x with old versions of threads and
271 0.34 2009-04-19 16:55 UTC
272 + Fix : Destruction of wizards in require.
273 + Fix : "panic: restartop" in 5.10 when dieing at compile time triggers
276 0.33 2009-03-26 00:00 UTC
277 + Chg : cast() and getsig() now croak too when an invalid signature is
279 + Fix : It was possible to generate with gensig() the same signature
280 twice before actually defining the wizards.
281 + Tst : More stash tests.
282 + Tst : Test when the magic takes place in some cases.
283 + Tst : Improved coverage for the op object feature.
285 0.32 2009-03-01 13:45 UTC
286 + Chg : dispell() and getdata() now croak when an invalid signature is
288 + Doc : More examples and a brand new synopsis.
289 + Fix : The signature is no longer stored in the mg_private member of
290 the MAGIC struct. This fixes possible clashes with magics from
292 + Fix : op info objects weren't blessed into the right class.
293 + Fix : Races when initializing the op names cache under threads.
295 0.31 2009-02-19 23:50 UTC
296 + Doc : A somewhat better introduction. The FUNCTION section was moved
297 before CONSTANTS and HISTORY since it's more important. And
298 'assignation' is really 'assignment', says Ricardo. :)
299 + Fix : The op_info features are now thread safe.
300 + Fix : Segfaults that occured when trying to get the op name during
302 + Fix : Segfaults and wrong "Unknown error" exceptions that happened
303 when dieing in require caused a free callback to fire.
304 Thanks Florian Ragwitz for reporting.
306 0.30 2009-02-12 17:05 UTC
307 + Add : You can now get in all the callbacks either the name or a B::OP
308 object representing the current op if you pass the new 'op_info'
309 option to wizard() with respective values VMG_OP_INFO_NAME and
311 + Doc : Document magic on array/hashes values.
312 + Fix : Completely skip free callbacks during global destruction.
313 + Tst : Tests now pass again on 5.8.3 and lower (they were broken since
315 + Tst : Test magic on symbol table and on array/hash values.
317 0.29 2009-02-08 11:10 UTC
318 + Add : Set $_[2] to the default length for len magic on scalars.
319 + Chg : Perl 5.8 is required.
320 + Chg : The CLONE method will no longer be defined for non-threaded
323 + Fix : The logic for turning off the GMAGICAL flag on hashes when no
324 uvar magic is involved could do nasty things on tied hashes.
325 + Upd : More resources in META.yml.
327 0.28 2009-01-24 17:05 UTC
328 + Add : The 'copy_key' option to wizard(), that enables the safe edition
329 of $_[2] in uvar callbacks to redirect actions to another key.
330 + Fix : Exception objects thrown from callbacks at compile time were not
332 + Fix : free callbacks triggered during global destruction could lead to
333 segfaults because the wizard was already freed.
334 + Fix : Refcount of scalars with free callback.
335 + Fix : Segfaults when using get or uvar magic simultaneously with clear
337 + Doc : The documentation about what the callbacks take and should
338 return is now somewhat clearer.
340 0.27 2009-01-18 23:40 UTC
341 + Fix : Offset of 1 with len magic on scalars.
342 + Fix : Segfaults that seem to happen when one croak in a callback at
343 compile time (thanks Ash Berlin for the testcase).
344 + Upd : Resources in META.yml (Florian Ragwitz)
345 + Tst : Large cleanup resulting in a wider set of tested behaviours.
347 0.26 2008-11-29 22:05 UTC
348 + Fix : Leaks of cloned coderefs that access lexicals.
349 + Fix : Building with maint-5.8 and blead (thanks Andreas Koenig).
351 0.25 2008-11-12 19:40 UTC
352 + Fix : Compatibility with 5.8.9.
353 + Fix : Old Pod::Coverage don't ignore CLONE.
354 + Upd : META.yml spec updated to 1.4.
356 0.24 2008-10-12 14:55 UTC
357 + Fix : Really fix it.
359 0.23 2008-10-11 17:25 UTC
360 + Fix : Building on perls with multiplicity but not ithreads.
362 0.22 2008-09-29 19:40 UTC
363 + Doc : Explain how magic differs from overloading.
364 + Fix : Disable thread safety for perl 5.8 on Windows. If you want it
365 enabled, reliable feedback is welcome.
367 0.21_02 2008-09-28 23:10 UTC
368 + Fix : Hide previously deleted magic tokens in vmg_svt_free (Yuval
371 0.21_01 2008-09-22 13:25 UTC
372 + Chg : Shrink the context data.
373 + Fix : Enable thread safety features for perls that just lack
375 + Tst : Output threads and threads::shared versions in threads tests.
377 0.20 2008-09-20 23:35 UTC
378 + Fix : The module is now thread safe. You can test if it had been
379 built with thread safety features enabled through the
380 VMG_THREADSAFE constant.
381 + Fix : Correct backtrace when wizard() croaks.
382 + Fix : 'cast $a, undef' should throw an 'invalid wizard' error instead
383 of an 'invalid signature'.
385 0.19 2008-07-05 15:05 UTC
386 + Add : The patchlevel with which the module has been built is now
387 exported with the VMG_PERL_PATCHLEVEL constant.
388 + Chg : Version checking macros overhaul.
389 + Fix : 'uvar' magic should only be enabled in maint-5.10 and higher
390 since it's never going to be backported to 5.8.
391 + Fix : Likewise, the copy callback API change should only be enabled in
393 + Fix : Instead of those ackward fake refcount increments in the wizard
394 destructor, flag the wizard with SVf_BREAK and force temps
395 cleanup. This should fix the wizard destruction for all perls,
396 and solves the leak introduced by the previous solution.
398 0.18 2008-07-04 12:20 UTC
399 + Fix : Enable uvar magic when patchlevel >= 28419.
400 + Fix : Set VMG_COMPAT_ARRAY_PUSH_NOLEN for Windows ActiveState Perl
403 0.17 2008-05-11 09:05 UTC
404 + Fix : 5.8.4 and lower don't have mPUSHi. Gah. Yeah, I know, I should
407 0.16 2008-05-10 22:05 UTC
408 + Add : The samples/copy.pl script.
409 + Chg : The sv_magical() + vmg_mg_magical() combo was simplified into
411 + Tst : t/33-code.t was testing scalars, not code.
412 + Tst : is() and like() are better than ok().
413 + Tst : 100% coverage reached.
415 0.15 2008-04-11 18:25 UTC
416 + Chg : Factor vmg_cb_call{1,2,3}() into one function.
417 + Fix : len magic is no longer called when taking the length() of a
418 magical scalar since p32969. The VMG_COMPAT_SCALAR_LENGTH_NOLEN
419 constant was added to cover this.
420 + Tst : More tests for t/22-len.t.
422 0.14 2008-03-24 12:35 UTC
423 + Fix : t/16-huf.t failures on Solaris and FreeBSD caused by not
424 updating mg->mg_ptr after Renew-ing it on dispell.
425 + Fix : Undefining MGf_DUP caused the wizard's magic to be wrongly
426 initialized. MGf_DUP now appears as true from the user point of
427 view, but the dup callback isn't actually set.
428 + Fix : Warnings with blead due to copy callbacks now taking an I32 (API
430 + Fix : vmg_svt_val() and vmg_uvar_del() aren't required if no uvar
432 + Tst : Tests now output the version of optional modules loaded.
434 0.13 2008-03-19 14:35 UTC
435 + Doc : Link to coverage report.
436 + Fix : Correct dependencies listing in META.yml.
437 + Tst : Improved test coverage.
438 + Tst : Print the patchlevel as a comment.
439 + Tst : Use is() where it's relevant.
440 + Tst : t/16-huf.t now really tests interaction with H::U::FH.
442 0.12 2008-02-07 18:15 UTC
443 + Fix : POD error. Thanks to Chris Williams (BinGOs) for the quick
446 0.11 2008-02-07 17:55 UTC
447 + Add : Copy callbacks now receive the current key/index in $_[2].
448 The current value/element is now in $_[3].
449 + Chg : The trigger of copy magic on list assignment with perls greater
450 than 5.10.0 was caused by uvar magic. Hence,
451 VMG_COMPAT_HASH_LISTASSIGN_COPY was removed. Use VMG_UVAR
453 + Fix : Build failures on patched perls.
454 + Tst : Added missing exported symbols to 01-import.t.
456 0.10 2008-02-04 11:30 UTC
457 + Add : New script : samples/vm_vs_tie.pl, that benchmarks our uvar
458 magic versus tied hashes.
459 + Add : The VMG_COMPAT_* constants can be used from userspace to check
460 perl magic abilities.
461 + Fix : Callbacks that returned undef made us croak, breaking the
462 variable behaviour (regression test in 14-callbacks.t).
463 + Fix : uvar callbacks weren't tested for non-NULL-ity before being
464 called (regression test in 28-uvar.t).
465 + Tst : Fix typo in 25-copy.t that prevented Tie::Hash tests to be ran.
467 0.09 2008-02-02 11:30 UTC
468 + Doc : Explicitely say that uvar callbacks are safely ignored for
470 + Doc : Document caveats and fix the usual set of typos.
471 + Fix : vmg_dispell() didn't check if the ext magic were ours when
472 counting wizards that have uvar callbacks, resulting in a
473 possible memory misread.
474 + Fix : getdata() now returns directly the data object, and no longer a
475 copy. This caused a leak.
476 + Tst : Prefix author tests by 9*-.
477 + Tst : New optional author test : 95-portability-files.t, that uses
478 Test::Portability::Files when it's present.
479 + Tst : New test : 14-self.t, that tests application of magic on the
481 + Tst : Move Hash::Util::FieldHash tests out of 11-multiple.t to
484 0.08 2008-02-01 16:55 UTC
485 + Add : copy magic for tied arrays/hashes.
487 + Add : uvar magics : fetch, store, exists, delete for hashes.
490 Internal development versions.
492 0.06 2007-11-20 10:10 UTC
493 + Chg : 5.7.3 is now officially required.
494 + Fix : "data" test failures on 5.8.{0,2}.
495 + Fix : Drand01() vs rand().
497 0.05 2007-11-19 09:10 UTC
498 + Fix : 5.10.0_RC1 compatibility fix.
500 0.04 2007-08-28 12:25 UTC
501 + Chg : s/require (XSLoader)/use $1/.
502 + Fix : Tests are now strict.
503 + Fix : Complete dependencies.
505 0.03 2007-08-01 17:20 UTC
506 + Add : Passing the signature of an already defined magic to wizard()
507 now returns the corresponding magic object.
508 + Add : You can pass the numeric signature as the wizard argument of
509 cast(), dispell() and getdata().
510 + Add : Any argument specified after the wizard (or the signature) in a
511 call to cast() is now passed to the private data constructor in
513 + Chg : $_[0] is now always a reference to the magic variable in all
514 callbacks. The reason for changing from the previous behaviour
515 is that one may want to apply the same magic to a plain scalar
516 and to a scalar reference, and needs a way to distinguish
517 between them in the callback (say, ref()).
518 + Fix : Wizard object destruction used not to free the signature.
520 0.02 2007-07-27 13:50 UTC
521 + Fix : In response to test report 548152 :
522 Newx() and SvMAGIC_set() not present on older perls.
523 + Fix : In response to test report 548275 :
524 Since perl 5.9.5, 'clear' magic is invoked when an array is
525 undefined (bug #43357). Moreover, 'len' magic is no longer
526 called by pushing an element since perl 5.9.3.
527 + Fix : Missing glob test in MANIFEST.
529 0.01 2007-07-25 16:15 UTC
530 First version, released on an unsuspecting world.