1 Revision history for Variable-Magic
3 0.57 2015-04-17 15:20 UTC
4 + Chg : The new environment variable to enable thread tests on older
5 perls is PERL_FORCE_TEST_THREADS. Note that this variable
6 should only be turned on by authors.
7 + Fix : Segfaults when the module is loaded by several threads (or
8 Windows emulated processes) ran in parallel.
9 + Fix : Segfaults when the module is loaded in a thread, which spawns
10 itself a new thread, and that child thread outlives its parent.
11 + Fix : Small memory leaks of structures required for thread safety.
12 + Fix : Update the Windows ActivePerl + gcc 3.4 workaround for
13 ExtUtils::MakeMaker 7.04. Thanks Christian Walde for reporting
14 and feedback on this issue.
15 + Tst : The global destruction test will now be exercised on any perl
16 that has DEBUGGING set.
17 + Tst : Optional capturing tests in t/17-ctl.t that were only run when
18 Capture::Tiny was present were converted to an IPC::Open3
19 based helper and will now be run everywhere.
21 0.56 2015-03-11 15:15 UTC
22 + Fix : [RT #101410] : Install fails in blead
23 Even though the change that caused this error was reverted from
24 blead, lvalue uses of ERRSV have been removed from this module
25 so that there will be no breakage when this change is possibly
26 reintroduced in the following months.
27 Thanks Dave Rolsky for reporting.
28 + Tst : $ENV{$Config{ldlibpthname}} is now preserved on all platforms,
29 which will address failures of t/17-ctl.t with unusual
30 compilers (like icc) that link all their compiled objects to
32 + Tst : The global destruction test is now only run on perl 5.13.4 and
33 higher, and only if either Perl::Destruct::Level is installed
34 or PERL_DESTRUCT_LEVEL is set and the perl is a debugging perl.
35 This will solve rare crashes of t/15-self.t on perl 5.13.3 and
38 0.55 2014-10-20 23:45 UTC
39 + Fix : [RT #99319] : t/18-opinfo.t: test failure due to change in
41 The new METHOP op class added in perl 5.21.5 is now supported.
42 Thanks James E Keenan for reporting.
44 0.54 2014-09-22 17:30 UTC
45 + Add : The new constant VMG_COMPAT_CODE_COPY_CLONE evaluates to true
46 if your perl calls 'copy' magic when a magical code prototype
47 is cloned, which is currently the case for perl 5.17.0 and
49 + Fix : [RT #90205] : copy magic on subs puts raw CV in $_[3]
50 $_[3] will now contain a reference to the cloned code when
51 'copy' magic is called for a coderef.
52 Thanks Lukas Mai for reporting.
53 + Fix : t/35-stash.t has been taught about perl 5.21.4.
54 + Fix : Tests using run_perl() in t/17-ctl.t will no longer fail on
57 0.53 2013-09-01 17:50 UTC
58 This is a maintenance release. The code contains no functional change.
59 Satisfied users of version 0.52 can skip this update.
60 + Fix : [RT #86338] : typo fix.
61 Thanks dsteinbrunner@pobox.com for the patch.
62 + Tst : Author tests are no longer bundled with this distribution.
63 They are only made available to authors in the git repository.
65 0.52 2012-11-05 02:30 UTC
66 + Add : The new constant VMG_COMPAT_SCALAR_NOLEN evaluates to true
67 when your perl does not call 'len' magic for scalars, which
68 is the case for perl 5.17.4 and above.
70 + Tst : t/35-stash.t has been taught about perl 5.17.4.
71 + Tst : Author tests overhaul.
73 0.51 2012-08-18 15:00 UTC
74 + Fix : It is now safe to call dispell() from inside 'free', 'copy' and
75 'uvar' callbacks to dispell the magic currently in use.
76 Thanks Clinton Gormley for reporting.
77 + Fix : Exceptions thrown from inside a 'free' callback are now always
78 consistently propagated outside of the callback. They used to
79 be lost when the 'free' callback was invoked at the end of an
81 + Fix : The 'reset RMG flag' workaroundn used to allow wizards with
82 both 'uvar' and 'clear' magics to be cast onto a hash, has been
85 0.50 2012-06-24 23:00 UTC
86 + Fix : Less memory is leaked when a wizard is freed during global
87 destruction, or when an exception is thrown from a 'free'
89 + Fix : [RT #77991] : t/17-ctl.t fails on perl 5.14 and 5.16.
90 This was actually an issue with ActivePerl, and this test
91 has learned to cope with it.
92 Thanks Gisle Aas for reporting.
93 + Tst : t/35-stash.t has been taught about perl 5.17.1.
94 + Doc : Many clarifications.
96 0.49 2012-06-05 21:40 UTC
97 This is a maintenance release. The code contains no functional change.
98 Satisfied users of version 0.48 can skip this update.
99 + Fix : [RT #77644] : t/17_ctl.t fails in test 44 since 5.17.0.
100 This test has been taught about perl 5.17.0.
101 Thanks Reini Urban for reporting.
102 + Tst : t/99-kwalitee.t will be skipped when only problematic versions
103 of its dependencies are available.
105 0.48 2012-02-17 23:40 UTC
106 + Add : You can now pass a reference to undef as the magic callback in
107 order to install a no-op callback.
108 Thanks Florian Ragwitz for the suggestion.
110 0.47 2011-10-27 16:55 UTC
111 + Add : The new constant VMG_COMPAT_HASH_DELETE_NOUVAR_VOID evaluates
112 to true when "delete $hash{key}" does not call 'delete' uvar
113 magic in void context.
114 + Chg : The MAGIC tokens created by this module no longer use the
115 mg_private member for storing a "magical" signature.
116 + Fix : Triggering magic in a thread on a variable cloned from the main
117 interpreter, and when the wizard already went out of scope,
118 will not segfault anymore.
119 + Opt : The pointer table is no longer needed to ensure thread safety.
120 The size of the object code is about 8% smaller for threaded
122 + Tst : Threads tests will not fail anymore if resources constraints
123 prevent the system from creating all the required threads.
125 0.46 2011-01-23 16:45 UTC
126 + Fix : [RT #64866] : Assertion failure with perl 5.13.9.
127 The real problem was that dispell() wasn't properly resetting
128 the magical sv flags, which became visible in the test suite
129 only with perl 5.13.9.
130 Thanks Joshua ben Jore for reporting.
132 0.45 2010-11-21 23:15 UTC
133 This is a maintenance release. The code contains no functional change.
134 Users of 0.44 can skip this update.
135 + Doc : C++ compilers are officially NOT supported.
136 + Doc : The minimum perl 5.10.0 requirement for uvar magic has been
138 Thanks Peter Rabbitson for pointing this out and contributing a
140 + Tst : Tune for perl 5.13.7.
141 + Tst : Capture::Tiny will be used in t/17-ctl.t if and only if it can
142 capture a simple run.
144 0.44 2010-09-24 19:10 UTC
145 + Fix : Broken linkage on Windows with gcc 3.4, which appears in
146 particular when using ActivePerl's default compiler suite.
147 For those setups, the Variable::Magic shared library will now
148 be linked against the perl dll directly (instead of the import
149 library). This (should) fix RT #51483.
150 Thanks Christian Walde for helping to reproduce this failure
152 + Rem : Support for development perls from the 5.11 branch but older
153 than the 5.11.0 release was removed. This could cause more
154 recent setups to fail.
155 + Tst : Threads tests are now only run on perl 5.13.4 and higher.
156 They could segfault randomly because of what seems to be an
157 internal bug of Perl, which has been addressed in 5.13.4.
158 There is also an environment variable that allows you to
159 forcefully run those tests, but it should be set only for
160 author testing and not for end users.
162 0.43 2010-06-25 23:35 UTC
163 + Add : The new constant VMG_COMPAT_GLOB_GET tells you whether get magic
164 is called for globs. It's true starting perl 5.13.2.
165 + Chg : All callbacks are now called within an eval-like context.
166 Only free callbacks used to be called that way.
167 + Fix : Some exceptions thrown from a free callback could be lost.
168 + Fix : Croak messages could sometimes be repeated several times.
169 + Fix : t/41-clone.t segfaulting with perl 5.13.2.
171 0.42 2010-05-19 00:15 UTC
172 This is a maintenance release. The code contains no functional change.
173 Users of 0.41 can skip this update.
174 + Fix : Test failures with perl 5.13.
175 + Tst : Improve coverage.
177 0.41 2010-03-15 17:35 UTC
178 + Doc : Tweaks and fixups.
180 + Fix : Compatibility with the soon-to-be-released perl 5.12.0.
181 + Fix : Correctly propagate the errors thrown when variable destruction
182 happens at compile-time and not from inside eval STRING.
183 Thanks Florian Ragwitz and Ash Berlin for reporting.
185 0.40 2010-01-06 23:20 UTC
186 + Fix : Possible memory miswrites when passing data arguments to cast().
187 + Fix : Minor C portability tweaks.
189 0.39 2009-12-01 00:05 UTC
190 + Add : You can use a function name as a callback by passing a string
191 reference to wizard() instead of a code reference.
192 + Fix : Compatiblity with perl 5.11.1 and 5.11.2.
193 + Fix : Scalars stored into the data slot no longer leak.
194 + Fix : Thread destruction should not segfault anymore.
195 + Opt : As a result of removing the deprecated features, less memory is
196 used for both threaded (a pointer table is used instead of a
197 hash) and non-threaded (no global structure is needed anymore)
199 + Rem : As advertised in the precendent version, all the signature
200 related features were removed. This includes the 'sig' option
201 to wizard(), the getsig() and gensig() functions, and the SIG_*
204 0.38 2009-10-04 16:10 UTC
205 + Dep : All the signature-related features are DEPRECATED and will be
206 removed in december 2009. This includes the 'sig' option to
207 wizard(), the getsig() and gensig() functions, and the SIG_*
209 Afaik those features were never used in real life, and they make
210 the XS code slow and brittle. If you want your wizards to be
211 globally available, you'll have to handle that yourself by
212 storing them in a hash.
213 + Doc : More examples in a new COOKBOOK section.
214 + Fix : Compatiblity with the 5.11.0 release. The new compatibility
215 constant VMG_COMPAT_ARRAY_PUSH_NOLEN_VOID was added to cover
217 + Fix : Work around Kwalitee test misfailures.
219 0.37 2009-08-25 15:25 UTC
220 + Add : The new constant VMG_FORKSAFE can be tested to know whether the
221 module will behave nicely when fork()ing. It's currently always
222 true except on Windows where you need perl 5.10.1 for it to be
224 + Doc : Nits and clarifications.
226 0.36 2009-07-05 16:30 UTC
227 + Chg : getdata() now returns an empty list when no magic is present.
228 It used to return undef but did not croak as stated in the doc.
229 Thanks Matt S. Trout for pointing this out.
230 + Fix : Building and tests with ActiveState Perl 5.8 build >= 822.
232 0.35 2009-05-15 20:50 UTC
233 This is a maintenance release. Users of 0.34 can skip this update.
234 + Chg : uvar magic used to be enabled for 5.9.5. It now requires 5.10 or
236 + Fix : Silence compilation warnings for cxinc().
237 + Rem : The workaround introduced in 0.31 for the "Unknown errors" issue
238 was removed, as regression tests show that 0.34 provides a
240 + Tst : Skip t/40-threads.t on 5.8.x with old versions of threads and
243 0.34 2009-04-19 16:55 UTC
244 + Fix : Destruction of wizards in require.
245 + Fix : "panic: restartop" in 5.10 when dieing at compile time triggers
248 0.33 2009-03-26 00:00 UTC
249 + Chg : cast() and getsig() now croak too when an invalid signature is
251 + Fix : It was possible to generate with gensig() the same signature
252 twice before actually defining the wizards.
253 + Tst : More stash tests.
254 + Tst : Test when the magic takes place in some cases.
255 + Tst : Improved coverage for the op object feature.
257 0.32 2009-03-01 13:45 UTC
258 + Chg : dispell() and getdata() now croak when an invalid signature is
260 + Doc : More examples and a brand new synopsis.
261 + Fix : The signature is no longer stored in the mg_private member of
262 the MAGIC struct. This fixes possible clashes with magics from
264 + Fix : op info objects weren't blessed into the right class.
265 + Fix : Races when initializing the op names cache under threads.
267 0.31 2009-02-19 23:50 UTC
268 + Doc : A somewhat better introduction. The FUNCTION section was moved
269 before CONSTANTS and HISTORY since it's more important. And
270 'assignation' is really 'assignment', says Ricardo. :)
271 + Fix : The op_info features are now thread safe.
272 + Fix : Segfaults that occured when trying to get the op name during
274 + Fix : Segfaults and wrong "Unknown error" exceptions that happened
275 when dieing in require caused a free callback to fire.
276 Thanks Florian Ragwitz for reporting.
278 0.30 2009-02-12 17:05 UTC
279 + Add : You can now get in all the callbacks either the name or a B::OP
280 object representing the current op if you pass the new 'op_info'
281 option to wizard() with respective values VMG_OP_INFO_NAME and
283 + Doc : Document magic on array/hashes values.
284 + Fix : Completely skip free callbacks during global destruction.
285 + Tst : Tests now pass again on 5.8.3 and lower (they were broken since
287 + Tst : Test magic on symbol table and on array/hash values.
289 0.29 2009-02-08 11:10 UTC
290 + Add : Set $_[2] to the default length for len magic on scalars.
291 + Chg : Perl 5.8 is required.
292 + Chg : The CLONE method will no longer be defined for non-threaded
295 + Fix : The logic for turning off the GMAGICAL flag on hashes when no
296 uvar magic is involved could do nasty things on tied hashes.
297 + Upd : More resources in META.yml.
299 0.28 2009-01-24 17:05 UTC
300 + Add : The 'copy_key' option to wizard(), that enables the safe edition
301 of $_[2] in uvar callbacks to redirect actions to another key.
302 + Fix : Exception objects thrown from callbacks at compile time were not
304 + Fix : free callbacks triggered during global destruction could lead to
305 segfaults because the wizard was already freed.
306 + Fix : Refcount of scalars with free callback.
307 + Fix : Segfaults when using get or uvar magic simultaneously with clear
309 + Doc : The documentation about what the callbacks take and should
310 return is now somewhat clearer.
312 0.27 2009-01-18 23:40 UTC
313 + Fix : Offset of 1 with len magic on scalars.
314 + Fix : Segfaults that seem to happen when one croak in a callback at
315 compile time (thanks Ash Berlin for the testcase).
316 + Upd : Resources in META.yml (Florian Ragwitz)
317 + Tst : Large cleanup resulting in a wider set of tested behaviours.
319 0.26 2008-11-29 22:05 UTC
320 + Fix : Leaks of cloned coderefs that access lexicals.
321 + Fix : Building with maint-5.8 and blead (thanks Andreas Koenig).
323 0.25 2008-11-12 19:40 UTC
324 + Fix : Compatibility with 5.8.9.
325 + Fix : Old Pod::Coverage don't ignore CLONE.
326 + Upd : META.yml spec updated to 1.4.
328 0.24 2008-10-12 14:55 UTC
329 + Fix : Really fix it.
331 0.23 2008-10-11 17:25 UTC
332 + Fix : Building on perls with multiplicity but not ithreads.
334 0.22 2008-09-29 19:40 UTC
335 + Doc : Explain how magic differs from overloading.
336 + Fix : Disable thread safety for perl 5.8 on Windows. If you want it
337 enabled, reliable feedback is welcome.
339 0.21_02 2008-09-28 23:10 UTC
340 + Fix : Hide previously deleted magic tokens in vmg_svt_free (Yuval
343 0.21_01 2008-09-22 13:25 UTC
344 + Chg : Shrink the context data.
345 + Fix : Enable thread safety features for perls that just lack
347 + Tst : Output threads and threads::shared versions in threads tests.
349 0.20 2008-09-20 23:35 UTC
350 + Fix : The module is now thread safe. You can test if it had been
351 built with thread safety features enabled through the
352 VMG_THREADSAFE constant.
353 + Fix : Correct backtrace when wizard() croaks.
354 + Fix : 'cast $a, undef' should throw an 'invalid wizard' error instead
355 of an 'invalid signature'.
357 0.19 2008-07-05 15:05 UTC
358 + Add : The patchlevel with which the module has been built is now
359 exported with the VMG_PERL_PATCHLEVEL constant.
360 + Chg : Version checking macros overhaul.
361 + Fix : 'uvar' magic should only be enabled in maint-5.10 and higher
362 since it's never going to be backported to 5.8.
363 + Fix : Likewise, the copy callback API change should only be enabled in
365 + Fix : Instead of those ackward fake refcount increments in the wizard
366 destructor, flag the wizard with SVf_BREAK and force temps
367 cleanup. This should fix the wizard destruction for all perls,
368 and solves the leak introduced by the previous solution.
370 0.18 2008-07-04 12:20 UTC
371 + Fix : Enable uvar magic when patchlevel >= 28419.
372 + Fix : Set VMG_COMPAT_ARRAY_PUSH_NOLEN for Windows ActiveState Perl
375 0.17 2008-05-11 09:05 UTC
376 + Fix : 5.8.4 and lower don't have mPUSHi. Gah. Yeah, I know, I should
379 0.16 2008-05-10 22:05 UTC
380 + Add : The samples/copy.pl script.
381 + Chg : The sv_magical() + vmg_mg_magical() combo was simplified into
383 + Tst : t/33-code.t was testing scalars, not code.
384 + Tst : is() and like() are better than ok().
385 + Tst : 100% coverage reached.
387 0.15 2008-04-11 18:25 UTC
388 + Chg : Factor vmg_cb_call{1,2,3}() into one function.
389 + Fix : len magic is no longer called when taking the length() of a
390 magical scalar since p32969. The VMG_COMPAT_SCALAR_LENGTH_NOLEN
391 constant was added to cover this.
392 + Tst : More tests for t/22-len.t.
394 0.14 2008-03-24 12:35 UTC
395 + Fix : t/16-huf.t failures on Solaris and FreeBSD caused by not
396 updating mg->mg_ptr after Renew-ing it on dispell.
397 + Fix : Undefining MGf_DUP caused the wizard's magic to be wrongly
398 initialized. MGf_DUP now appears as true from the user point of
399 view, but the dup callback isn't actually set.
400 + Fix : Warnings with blead due to copy callbacks now taking an I32 (API
402 + Fix : vmg_svt_val() and vmg_uvar_del() aren't required if no uvar
404 + Tst : Tests now output the version of optional modules loaded.
406 0.13 2008-03-19 14:35 UTC
407 + Doc : Link to coverage report.
408 + Fix : Correct dependencies listing in META.yml.
409 + Tst : Improved test coverage.
410 + Tst : Print the patchlevel as a comment.
411 + Tst : Use is() where it's relevant.
412 + Tst : t/16-huf.t now really tests interaction with H::U::FH.
414 0.12 2008-02-07 18:15 UTC
415 + Fix : POD error. Thanks to Chris Williams (BinGOs) for the quick
418 0.11 2008-02-07 17:55 UTC
419 + Add : Copy callbacks now receive the current key/index in $_[2].
420 The current value/element is now in $_[3].
421 + Chg : The trigger of copy magic on list assignment with perls greater
422 than 5.10.0 was caused by uvar magic. Hence,
423 VMG_COMPAT_HASH_LISTASSIGN_COPY was removed. Use VMG_UVAR
425 + Fix : Build failures on patched perls.
426 + Tst : Added missing exported symbols to 01-import.t.
428 0.10 2008-02-04 11:30 UTC
429 + Add : New script : samples/vm_vs_tie.pl, that benchmarks our uvar
430 magic versus tied hashes.
431 + Add : The VMG_COMPAT_* constants can be used from userspace to check
432 perl magic abilities.
433 + Fix : Callbacks that returned undef made us croak, breaking the
434 variable behaviour (regression test in 14-callbacks.t).
435 + Fix : uvar callbacks weren't tested for non-NULL-ity before being
436 called (regression test in 28-uvar.t).
437 + Tst : Fix typo in 25-copy.t that prevented Tie::Hash tests to be ran.
439 0.09 2008-02-02 11:30 UTC
440 + Doc : Explicitely say that uvar callbacks are safely ignored for
442 + Doc : Document caveats and fix the usual set of typos.
443 + Fix : vmg_dispell() didn't check if the ext magic were ours when
444 counting wizards that have uvar callbacks, resulting in a
445 possible memory misread.
446 + Fix : getdata() now returns directly the data object, and no longer a
447 copy. This caused a leak.
448 + Tst : Prefix author tests by 9*-.
449 + Tst : New optional author test : 95-portability-files.t, that uses
450 Test::Portability::Files when it's present.
451 + Tst : New test : 14-self.t, that tests application of magic on the
453 + Tst : Move Hash::Util::FieldHash tests out of 11-multiple.t to
456 0.08 2008-02-01 16:55 UTC
457 + Add : copy magic for tied arrays/hashes.
459 + Add : uvar magics : fetch, store, exists, delete for hashes.
462 Internal development versions.
464 0.06 2007-11-20 10:10 UTC
465 + Chg : 5.7.3 is now officially required.
466 + Fix : "data" test failures on 5.8.{0,2}.
467 + Fix : Drand01() vs rand().
469 0.05 2007-11-19 09:10 UTC
470 + Fix : 5.10.0_RC1 compatibility fix.
472 0.04 2007-08-28 12:25 UTC
473 + Chg : s/require (XSLoader)/use $1/.
474 + Fix : Tests are now strict.
475 + Fix : Complete dependencies.
477 0.03 2007-08-01 17:20 UTC
478 + Add : Passing the signature of an already defined magic to wizard()
479 now returns the corresponding magic object.
480 + Add : You can pass the numeric signature as the wizard argument of
481 cast(), dispell() and getdata().
482 + Add : Any argument specified after the wizard (or the signature) in a
483 call to cast() is now passed to the private data constructor in
485 + Chg : $_[0] is now always a reference to the magic variable in all
486 callbacks. The reason for changing from the previous behaviour
487 is that one may want to apply the same magic to a plain scalar
488 and to a scalar reference, and needs a way to distinguish
489 between them in the callback (say, ref()).
490 + Fix : Wizard object destruction used not to free the signature.
492 0.02 2007-07-27 13:50 UTC
493 + Fix : In response to test report 548152 :
494 Newx() and SvMAGIC_set() not present on older perls.
495 + Fix : In response to test report 548275 :
496 Since perl 5.9.5, 'clear' magic is invoked when an array is
497 undefined (bug #43357). Moreover, 'len' magic is no longer
498 called by pushing an element since perl 5.9.3.
499 + Fix : Missing glob test in MANIFEST.
501 0.01 2007-07-25 16:15 UTC
502 First version, released on an unsuspecting world.