1 Revision history for Variable-Magic
3 0.58 2015-07-21 16:00 UTC
4 + Add : If a non-len magic callback returns a reference, it will now
5 only be freed at the end of the statement that caused the
6 magic to trigger. This allows the user to attach free magic
7 (or a plain destructor) to a token returned from the callbacks
8 in order to defer an action after the magic is processed by
10 + Fix : Test failures of threads tests on systems with harsh resource
11 constraints causing the threads to exit() during run.
13 0.57 2015-04-17 15:20 UTC
14 + Chg : The new environment variable to enable thread tests on older
15 perls is PERL_FORCE_TEST_THREADS. Note that this variable
16 should only be turned on by authors.
17 + Fix : Segfaults when the module is loaded by several threads (or
18 Windows emulated processes) ran in parallel.
19 + Fix : Segfaults when the module is loaded in a thread, which spawns
20 itself a new thread, and that child thread outlives its parent.
21 + Fix : Small memory leaks of structures required for thread safety.
22 + Fix : Update the Windows ActivePerl + gcc 3.4 workaround for
23 ExtUtils::MakeMaker 7.04. Thanks Christian Walde for reporting
24 and feedback on this issue.
25 + Tst : The global destruction test will now be exercised on any perl
26 that has DEBUGGING set.
27 + Tst : Optional capturing tests in t/17-ctl.t that were only run when
28 Capture::Tiny was present were converted to an IPC::Open3
29 based helper and will now be run everywhere.
31 0.56 2015-03-11 15:15 UTC
32 + Fix : [RT #101410] : Install fails in blead
33 Even though the change that caused this error was reverted from
34 blead, lvalue uses of ERRSV have been removed from this module
35 so that there will be no breakage when this change is possibly
36 reintroduced in the following months.
37 Thanks Dave Rolsky for reporting.
38 + Tst : $ENV{$Config{ldlibpthname}} is now preserved on all platforms,
39 which will address failures of t/17-ctl.t with unusual
40 compilers (like icc) that link all their compiled objects to
42 + Tst : The global destruction test is now only run on perl 5.13.4 and
43 higher, and only if either Perl::Destruct::Level is installed
44 or PERL_DESTRUCT_LEVEL is set and the perl is a debugging perl.
45 This will solve rare crashes of t/15-self.t on perl 5.13.3 and
48 0.55 2014-10-20 23:45 UTC
49 + Fix : [RT #99319] : t/18-opinfo.t: test failure due to change in
51 The new METHOP op class added in perl 5.21.5 is now supported.
52 Thanks James E Keenan for reporting.
54 0.54 2014-09-22 17:30 UTC
55 + Add : The new constant VMG_COMPAT_CODE_COPY_CLONE evaluates to true
56 if your perl calls 'copy' magic when a magical code prototype
57 is cloned, which is currently the case for perl 5.17.0 and
59 + Fix : [RT #90205] : copy magic on subs puts raw CV in $_[3]
60 $_[3] will now contain a reference to the cloned code when
61 'copy' magic is called for a coderef.
62 Thanks Lukas Mai for reporting.
63 + Fix : t/35-stash.t has been taught about perl 5.21.4.
64 + Fix : Tests using run_perl() in t/17-ctl.t will no longer fail on
67 0.53 2013-09-01 17:50 UTC
68 This is a maintenance release. The code contains no functional change.
69 Satisfied users of version 0.52 can skip this update.
70 + Fix : [RT #86338] : typo fix.
71 Thanks dsteinbrunner@pobox.com for the patch.
72 + Tst : Author tests are no longer bundled with this distribution.
73 They are only made available to authors in the git repository.
75 0.52 2012-11-05 02:30 UTC
76 + Add : The new constant VMG_COMPAT_SCALAR_NOLEN evaluates to true
77 when your perl does not call 'len' magic for scalars, which
78 is the case for perl 5.17.4 and above.
80 + Tst : t/35-stash.t has been taught about perl 5.17.4.
81 + Tst : Author tests overhaul.
83 0.51 2012-08-18 15:00 UTC
84 + Fix : It is now safe to call dispell() from inside 'free', 'copy' and
85 'uvar' callbacks to dispell the magic currently in use.
86 Thanks Clinton Gormley for reporting.
87 + Fix : Exceptions thrown from inside a 'free' callback are now always
88 consistently propagated outside of the callback. They used to
89 be lost when the 'free' callback was invoked at the end of an
91 + Fix : The 'reset RMG flag' workaroundn used to allow wizards with
92 both 'uvar' and 'clear' magics to be cast onto a hash, has been
95 0.50 2012-06-24 23:00 UTC
96 + Fix : Less memory is leaked when a wizard is freed during global
97 destruction, or when an exception is thrown from a 'free'
99 + Fix : [RT #77991] : t/17-ctl.t fails on perl 5.14 and 5.16.
100 This was actually an issue with ActivePerl, and this test
101 has learned to cope with it.
102 Thanks Gisle Aas for reporting.
103 + Tst : t/35-stash.t has been taught about perl 5.17.1.
104 + Doc : Many clarifications.
106 0.49 2012-06-05 21:40 UTC
107 This is a maintenance release. The code contains no functional change.
108 Satisfied users of version 0.48 can skip this update.
109 + Fix : [RT #77644] : t/17_ctl.t fails in test 44 since 5.17.0.
110 This test has been taught about perl 5.17.0.
111 Thanks Reini Urban for reporting.
112 + Tst : t/99-kwalitee.t will be skipped when only problematic versions
113 of its dependencies are available.
115 0.48 2012-02-17 23:40 UTC
116 + Add : You can now pass a reference to undef as the magic callback in
117 order to install a no-op callback.
118 Thanks Florian Ragwitz for the suggestion.
120 0.47 2011-10-27 16:55 UTC
121 + Add : The new constant VMG_COMPAT_HASH_DELETE_NOUVAR_VOID evaluates
122 to true when "delete $hash{key}" does not call 'delete' uvar
123 magic in void context.
124 + Chg : The MAGIC tokens created by this module no longer use the
125 mg_private member for storing a "magical" signature.
126 + Fix : Triggering magic in a thread on a variable cloned from the main
127 interpreter, and when the wizard already went out of scope,
128 will not segfault anymore.
129 + Opt : The pointer table is no longer needed to ensure thread safety.
130 The size of the object code is about 8% smaller for threaded
132 + Tst : Threads tests will not fail anymore if resources constraints
133 prevent the system from creating all the required threads.
135 0.46 2011-01-23 16:45 UTC
136 + Fix : [RT #64866] : Assertion failure with perl 5.13.9.
137 The real problem was that dispell() wasn't properly resetting
138 the magical sv flags, which became visible in the test suite
139 only with perl 5.13.9.
140 Thanks Joshua ben Jore for reporting.
142 0.45 2010-11-21 23:15 UTC
143 This is a maintenance release. The code contains no functional change.
144 Users of 0.44 can skip this update.
145 + Doc : C++ compilers are officially NOT supported.
146 + Doc : The minimum perl 5.10.0 requirement for uvar magic has been
148 Thanks Peter Rabbitson for pointing this out and contributing a
150 + Tst : Tune for perl 5.13.7.
151 + Tst : Capture::Tiny will be used in t/17-ctl.t if and only if it can
152 capture a simple run.
154 0.44 2010-09-24 19:10 UTC
155 + Fix : Broken linkage on Windows with gcc 3.4, which appears in
156 particular when using ActivePerl's default compiler suite.
157 For those setups, the Variable::Magic shared library will now
158 be linked against the perl dll directly (instead of the import
159 library). This (should) fix RT #51483.
160 Thanks Christian Walde for helping to reproduce this failure
162 + Rem : Support for development perls from the 5.11 branch but older
163 than the 5.11.0 release was removed. This could cause more
164 recent setups to fail.
165 + Tst : Threads tests are now only run on perl 5.13.4 and higher.
166 They could segfault randomly because of what seems to be an
167 internal bug of Perl, which has been addressed in 5.13.4.
168 There is also an environment variable that allows you to
169 forcefully run those tests, but it should be set only for
170 author testing and not for end users.
172 0.43 2010-06-25 23:35 UTC
173 + Add : The new constant VMG_COMPAT_GLOB_GET tells you whether get magic
174 is called for globs. It's true starting perl 5.13.2.
175 + Chg : All callbacks are now called within an eval-like context.
176 Only free callbacks used to be called that way.
177 + Fix : Some exceptions thrown from a free callback could be lost.
178 + Fix : Croak messages could sometimes be repeated several times.
179 + Fix : t/41-clone.t segfaulting with perl 5.13.2.
181 0.42 2010-05-19 00:15 UTC
182 This is a maintenance release. The code contains no functional change.
183 Users of 0.41 can skip this update.
184 + Fix : Test failures with perl 5.13.
185 + Tst : Improve coverage.
187 0.41 2010-03-15 17:35 UTC
188 + Doc : Tweaks and fixups.
190 + Fix : Compatibility with the soon-to-be-released perl 5.12.0.
191 + Fix : Correctly propagate the errors thrown when variable destruction
192 happens at compile-time and not from inside eval STRING.
193 Thanks Florian Ragwitz and Ash Berlin for reporting.
195 0.40 2010-01-06 23:20 UTC
196 + Fix : Possible memory miswrites when passing data arguments to cast().
197 + Fix : Minor C portability tweaks.
199 0.39 2009-12-01 00:05 UTC
200 + Add : You can use a function name as a callback by passing a string
201 reference to wizard() instead of a code reference.
202 + Fix : Compatiblity with perl 5.11.1 and 5.11.2.
203 + Fix : Scalars stored into the data slot no longer leak.
204 + Fix : Thread destruction should not segfault anymore.
205 + Opt : As a result of removing the deprecated features, less memory is
206 used for both threaded (a pointer table is used instead of a
207 hash) and non-threaded (no global structure is needed anymore)
209 + Rem : As advertised in the precendent version, all the signature
210 related features were removed. This includes the 'sig' option
211 to wizard(), the getsig() and gensig() functions, and the SIG_*
214 0.38 2009-10-04 16:10 UTC
215 + Dep : All the signature-related features are DEPRECATED and will be
216 removed in december 2009. This includes the 'sig' option to
217 wizard(), the getsig() and gensig() functions, and the SIG_*
219 Afaik those features were never used in real life, and they make
220 the XS code slow and brittle. If you want your wizards to be
221 globally available, you'll have to handle that yourself by
222 storing them in a hash.
223 + Doc : More examples in a new COOKBOOK section.
224 + Fix : Compatiblity with the 5.11.0 release. The new compatibility
225 constant VMG_COMPAT_ARRAY_PUSH_NOLEN_VOID was added to cover
227 + Fix : Work around Kwalitee test misfailures.
229 0.37 2009-08-25 15:25 UTC
230 + Add : The new constant VMG_FORKSAFE can be tested to know whether the
231 module will behave nicely when fork()ing. It's currently always
232 true except on Windows where you need perl 5.10.1 for it to be
234 + Doc : Nits and clarifications.
236 0.36 2009-07-05 16:30 UTC
237 + Chg : getdata() now returns an empty list when no magic is present.
238 It used to return undef but did not croak as stated in the doc.
239 Thanks Matt S. Trout for pointing this out.
240 + Fix : Building and tests with ActiveState Perl 5.8 build >= 822.
242 0.35 2009-05-15 20:50 UTC
243 This is a maintenance release. Users of 0.34 can skip this update.
244 + Chg : uvar magic used to be enabled for 5.9.5. It now requires 5.10 or
246 + Fix : Silence compilation warnings for cxinc().
247 + Rem : The workaround introduced in 0.31 for the "Unknown errors" issue
248 was removed, as regression tests show that 0.34 provides a
250 + Tst : Skip t/40-threads.t on 5.8.x with old versions of threads and
253 0.34 2009-04-19 16:55 UTC
254 + Fix : Destruction of wizards in require.
255 + Fix : "panic: restartop" in 5.10 when dieing at compile time triggers
258 0.33 2009-03-26 00:00 UTC
259 + Chg : cast() and getsig() now croak too when an invalid signature is
261 + Fix : It was possible to generate with gensig() the same signature
262 twice before actually defining the wizards.
263 + Tst : More stash tests.
264 + Tst : Test when the magic takes place in some cases.
265 + Tst : Improved coverage for the op object feature.
267 0.32 2009-03-01 13:45 UTC
268 + Chg : dispell() and getdata() now croak when an invalid signature is
270 + Doc : More examples and a brand new synopsis.
271 + Fix : The signature is no longer stored in the mg_private member of
272 the MAGIC struct. This fixes possible clashes with magics from
274 + Fix : op info objects weren't blessed into the right class.
275 + Fix : Races when initializing the op names cache under threads.
277 0.31 2009-02-19 23:50 UTC
278 + Doc : A somewhat better introduction. The FUNCTION section was moved
279 before CONSTANTS and HISTORY since it's more important. And
280 'assignation' is really 'assignment', says Ricardo. :)
281 + Fix : The op_info features are now thread safe.
282 + Fix : Segfaults that occured when trying to get the op name during
284 + Fix : Segfaults and wrong "Unknown error" exceptions that happened
285 when dieing in require caused a free callback to fire.
286 Thanks Florian Ragwitz for reporting.
288 0.30 2009-02-12 17:05 UTC
289 + Add : You can now get in all the callbacks either the name or a B::OP
290 object representing the current op if you pass the new 'op_info'
291 option to wizard() with respective values VMG_OP_INFO_NAME and
293 + Doc : Document magic on array/hashes values.
294 + Fix : Completely skip free callbacks during global destruction.
295 + Tst : Tests now pass again on 5.8.3 and lower (they were broken since
297 + Tst : Test magic on symbol table and on array/hash values.
299 0.29 2009-02-08 11:10 UTC
300 + Add : Set $_[2] to the default length for len magic on scalars.
301 + Chg : Perl 5.8 is required.
302 + Chg : The CLONE method will no longer be defined for non-threaded
305 + Fix : The logic for turning off the GMAGICAL flag on hashes when no
306 uvar magic is involved could do nasty things on tied hashes.
307 + Upd : More resources in META.yml.
309 0.28 2009-01-24 17:05 UTC
310 + Add : The 'copy_key' option to wizard(), that enables the safe edition
311 of $_[2] in uvar callbacks to redirect actions to another key.
312 + Fix : Exception objects thrown from callbacks at compile time were not
314 + Fix : free callbacks triggered during global destruction could lead to
315 segfaults because the wizard was already freed.
316 + Fix : Refcount of scalars with free callback.
317 + Fix : Segfaults when using get or uvar magic simultaneously with clear
319 + Doc : The documentation about what the callbacks take and should
320 return is now somewhat clearer.
322 0.27 2009-01-18 23:40 UTC
323 + Fix : Offset of 1 with len magic on scalars.
324 + Fix : Segfaults that seem to happen when one croak in a callback at
325 compile time (thanks Ash Berlin for the testcase).
326 + Upd : Resources in META.yml (Florian Ragwitz)
327 + Tst : Large cleanup resulting in a wider set of tested behaviours.
329 0.26 2008-11-29 22:05 UTC
330 + Fix : Leaks of cloned coderefs that access lexicals.
331 + Fix : Building with maint-5.8 and blead (thanks Andreas Koenig).
333 0.25 2008-11-12 19:40 UTC
334 + Fix : Compatibility with 5.8.9.
335 + Fix : Old Pod::Coverage don't ignore CLONE.
336 + Upd : META.yml spec updated to 1.4.
338 0.24 2008-10-12 14:55 UTC
339 + Fix : Really fix it.
341 0.23 2008-10-11 17:25 UTC
342 + Fix : Building on perls with multiplicity but not ithreads.
344 0.22 2008-09-29 19:40 UTC
345 + Doc : Explain how magic differs from overloading.
346 + Fix : Disable thread safety for perl 5.8 on Windows. If you want it
347 enabled, reliable feedback is welcome.
349 0.21_02 2008-09-28 23:10 UTC
350 + Fix : Hide previously deleted magic tokens in vmg_svt_free (Yuval
353 0.21_01 2008-09-22 13:25 UTC
354 + Chg : Shrink the context data.
355 + Fix : Enable thread safety features for perls that just lack
357 + Tst : Output threads and threads::shared versions in threads tests.
359 0.20 2008-09-20 23:35 UTC
360 + Fix : The module is now thread safe. You can test if it had been
361 built with thread safety features enabled through the
362 VMG_THREADSAFE constant.
363 + Fix : Correct backtrace when wizard() croaks.
364 + Fix : 'cast $a, undef' should throw an 'invalid wizard' error instead
365 of an 'invalid signature'.
367 0.19 2008-07-05 15:05 UTC
368 + Add : The patchlevel with which the module has been built is now
369 exported with the VMG_PERL_PATCHLEVEL constant.
370 + Chg : Version checking macros overhaul.
371 + Fix : 'uvar' magic should only be enabled in maint-5.10 and higher
372 since it's never going to be backported to 5.8.
373 + Fix : Likewise, the copy callback API change should only be enabled in
375 + Fix : Instead of those ackward fake refcount increments in the wizard
376 destructor, flag the wizard with SVf_BREAK and force temps
377 cleanup. This should fix the wizard destruction for all perls,
378 and solves the leak introduced by the previous solution.
380 0.18 2008-07-04 12:20 UTC
381 + Fix : Enable uvar magic when patchlevel >= 28419.
382 + Fix : Set VMG_COMPAT_ARRAY_PUSH_NOLEN for Windows ActiveState Perl
385 0.17 2008-05-11 09:05 UTC
386 + Fix : 5.8.4 and lower don't have mPUSHi. Gah. Yeah, I know, I should
389 0.16 2008-05-10 22:05 UTC
390 + Add : The samples/copy.pl script.
391 + Chg : The sv_magical() + vmg_mg_magical() combo was simplified into
393 + Tst : t/33-code.t was testing scalars, not code.
394 + Tst : is() and like() are better than ok().
395 + Tst : 100% coverage reached.
397 0.15 2008-04-11 18:25 UTC
398 + Chg : Factor vmg_cb_call{1,2,3}() into one function.
399 + Fix : len magic is no longer called when taking the length() of a
400 magical scalar since p32969. The VMG_COMPAT_SCALAR_LENGTH_NOLEN
401 constant was added to cover this.
402 + Tst : More tests for t/22-len.t.
404 0.14 2008-03-24 12:35 UTC
405 + Fix : t/16-huf.t failures on Solaris and FreeBSD caused by not
406 updating mg->mg_ptr after Renew-ing it on dispell.
407 + Fix : Undefining MGf_DUP caused the wizard's magic to be wrongly
408 initialized. MGf_DUP now appears as true from the user point of
409 view, but the dup callback isn't actually set.
410 + Fix : Warnings with blead due to copy callbacks now taking an I32 (API
412 + Fix : vmg_svt_val() and vmg_uvar_del() aren't required if no uvar
414 + Tst : Tests now output the version of optional modules loaded.
416 0.13 2008-03-19 14:35 UTC
417 + Doc : Link to coverage report.
418 + Fix : Correct dependencies listing in META.yml.
419 + Tst : Improved test coverage.
420 + Tst : Print the patchlevel as a comment.
421 + Tst : Use is() where it's relevant.
422 + Tst : t/16-huf.t now really tests interaction with H::U::FH.
424 0.12 2008-02-07 18:15 UTC
425 + Fix : POD error. Thanks to Chris Williams (BinGOs) for the quick
428 0.11 2008-02-07 17:55 UTC
429 + Add : Copy callbacks now receive the current key/index in $_[2].
430 The current value/element is now in $_[3].
431 + Chg : The trigger of copy magic on list assignment with perls greater
432 than 5.10.0 was caused by uvar magic. Hence,
433 VMG_COMPAT_HASH_LISTASSIGN_COPY was removed. Use VMG_UVAR
435 + Fix : Build failures on patched perls.
436 + Tst : Added missing exported symbols to 01-import.t.
438 0.10 2008-02-04 11:30 UTC
439 + Add : New script : samples/vm_vs_tie.pl, that benchmarks our uvar
440 magic versus tied hashes.
441 + Add : The VMG_COMPAT_* constants can be used from userspace to check
442 perl magic abilities.
443 + Fix : Callbacks that returned undef made us croak, breaking the
444 variable behaviour (regression test in 14-callbacks.t).
445 + Fix : uvar callbacks weren't tested for non-NULL-ity before being
446 called (regression test in 28-uvar.t).
447 + Tst : Fix typo in 25-copy.t that prevented Tie::Hash tests to be ran.
449 0.09 2008-02-02 11:30 UTC
450 + Doc : Explicitely say that uvar callbacks are safely ignored for
452 + Doc : Document caveats and fix the usual set of typos.
453 + Fix : vmg_dispell() didn't check if the ext magic were ours when
454 counting wizards that have uvar callbacks, resulting in a
455 possible memory misread.
456 + Fix : getdata() now returns directly the data object, and no longer a
457 copy. This caused a leak.
458 + Tst : Prefix author tests by 9*-.
459 + Tst : New optional author test : 95-portability-files.t, that uses
460 Test::Portability::Files when it's present.
461 + Tst : New test : 14-self.t, that tests application of magic on the
463 + Tst : Move Hash::Util::FieldHash tests out of 11-multiple.t to
466 0.08 2008-02-01 16:55 UTC
467 + Add : copy magic for tied arrays/hashes.
469 + Add : uvar magics : fetch, store, exists, delete for hashes.
472 Internal development versions.
474 0.06 2007-11-20 10:10 UTC
475 + Chg : 5.7.3 is now officially required.
476 + Fix : "data" test failures on 5.8.{0,2}.
477 + Fix : Drand01() vs rand().
479 0.05 2007-11-19 09:10 UTC
480 + Fix : 5.10.0_RC1 compatibility fix.
482 0.04 2007-08-28 12:25 UTC
483 + Chg : s/require (XSLoader)/use $1/.
484 + Fix : Tests are now strict.
485 + Fix : Complete dependencies.
487 0.03 2007-08-01 17:20 UTC
488 + Add : Passing the signature of an already defined magic to wizard()
489 now returns the corresponding magic object.
490 + Add : You can pass the numeric signature as the wizard argument of
491 cast(), dispell() and getdata().
492 + Add : Any argument specified after the wizard (or the signature) in a
493 call to cast() is now passed to the private data constructor in
495 + Chg : $_[0] is now always a reference to the magic variable in all
496 callbacks. The reason for changing from the previous behaviour
497 is that one may want to apply the same magic to a plain scalar
498 and to a scalar reference, and needs a way to distinguish
499 between them in the callback (say, ref()).
500 + Fix : Wizard object destruction used not to free the signature.
502 0.02 2007-07-27 13:50 UTC
503 + Fix : In response to test report 548152 :
504 Newx() and SvMAGIC_set() not present on older perls.
505 + Fix : In response to test report 548275 :
506 Since perl 5.9.5, 'clear' magic is invoked when an array is
507 undefined (bug #43357). Moreover, 'len' magic is no longer
508 called by pushing an element since perl 5.9.3.
509 + Fix : Missing glob test in MANIFEST.
511 0.01 2007-07-25 16:15 UTC
512 First version, released on an unsuspecting world.