1 Revision history for Variable-Magic
3 0.48 2012-02-17 23:40 UTC
4 + Add : You can now pass a reference to undef as the magic callback in
5 order to install a no-op callback.
6 Thanks Florian Ragwitz for the suggestion.
8 0.47 2011-10-27 16:55 UTC
9 + Add : The new constant VMG_COMPAT_HASH_DELETE_NOUVAR_VOID evaluates
10 to true when "delete $hash{key}" does not call 'delete' uvar
11 magic in void context.
12 + Chg : The MAGIC tokens created by this module no longer use the
13 mg_private member for storing a "magical" signature.
14 + Fix : Triggering magic in a thread on a variable cloned from the main
15 interpreter, and when the wizard already went out of scope,
16 will not segfault anymore.
17 + Opt : The pointer table is no longer needed to ensure thread safety.
18 The size of the object code is about 8% smaller for threaded
20 + Tst : Threads tests will not fail anymore if resources constraints
21 prevent the system from creating all the required threads.
23 0.46 2011-01-23 16:45 UTC
24 + Fix : [RT #64866] : Assertion failure with perl 5.13.9.
25 The real problem was that dispell() wasn't properly resetting
26 the magical sv flags, which became visible in the test suite
27 only with perl 5.13.9.
28 Thanks Joshua ben Jore for reporting.
30 0.45 2010-11-21 23:15 UTC
31 This is a maintenance release. The code contains no functional change.
32 Users of 0.44 can skip this update.
33 + Doc : C++ compilers are officially NOT supported.
34 + Doc : The minimum perl 5.10.0 requirement for uvar magic has been
36 Thanks Peter Rabbitson for pointing this out and contributing a
38 + Tst : Tune for perl 5.13.7.
39 + Tst : Capture::Tiny will be used in t/17-ctl.t if and only if it can
42 0.44 2010-09-24 19:10 UTC
43 + Fix : Broken linkage on Windows with gcc 3.4, which appears in
44 particular when using ActivePerl's default compiler suite.
45 For those setups, the Variable::Magic shared library will now
46 be linked against the perl dll directly (instead of the import
47 library). This (should) fix RT #51483.
48 Thanks Christian Walde for helping to reproduce this failure
50 + Rem : Support for development perls from the 5.11 branch but older
51 than the 5.11.0 release was removed. This could cause more
52 recent setups to fail.
53 + Tst : Threads tests are now only run on perl 5.13.4 and higher.
54 They could segfault randomly because of what seems to be an
55 internal bug of Perl, which has been addressed in 5.13.4.
56 There is also an environment variable that allows you to
57 forcefully run those tests, but it should be set only for
58 author testing and not for end users.
60 0.43 2010-06-25 23:35 UTC
61 + Add : The new constant VMG_COMPAT_GLOB_GET tells you whether get magic
62 is called for globs. It's true starting perl 5.13.2.
63 + Chg : All callbacks are now called within an eval-like context.
64 Only free callbacks used to be called that way.
65 + Fix : Some exceptions thrown from a free callback could be lost.
66 + Fix : Croak messages could sometimes be repeated several times.
67 + Fix : t/41-clone.t segfaulting with perl 5.13.2.
69 0.42 2010-05-19 00:15 UTC
70 This is a maintenance release. The code contains no functional change.
71 Users of 0.41 can skip this update.
72 + Fix : Test failures with perl 5.13.
73 + Tst : Improve coverage.
75 0.41 2010-03-15 17:35 UTC
76 + Doc : Tweaks and fixups.
78 + Fix : Compatibility with the soon-to-be-released perl 5.12.0.
79 + Fix : Correctly propagate the errors thrown when variable destruction
80 happens at compile-time and not from inside eval STRING.
81 Thanks Florian Ragwitz and Ash Berlin for reporting.
83 0.40 2010-01-06 23:20 UTC
84 + Fix : Possible memory miswrites when passing data arguments to cast().
85 + Fix : Minor C portability tweaks.
87 0.39 2009-12-01 00:05 UTC
88 + Add : You can use a function name as a callback by passing a string
89 reference to wizard() instead of a code reference.
90 + Fix : Compatiblity with perl 5.11.1 and 5.11.2.
91 + Fix : Scalars stored into the data slot no longer leak.
92 + Fix : Thread destruction should not segfault anymore.
93 + Opt : As a result of removing the deprecated features, less memory is
94 used for both threaded (a pointer table is used instead of a
95 hash) and non-threaded (no global structure is needed anymore)
97 + Rem : As advertised in the precendent version, all the signature
98 related features were removed. This includes the 'sig' option
99 to wizard(), the getsig() and gensig() functions, and the SIG_*
102 0.38 2009-10-04 16:10 UTC
103 + Dep : All the signature-related features are DEPRECATED and will be
104 removed in december 2009. This includes the 'sig' option to
105 wizard(), the getsig() and gensig() functions, and the SIG_*
107 Afaik those features were never used in real life, and they make
108 the XS code slow and brittle. If you want your wizards to be
109 globally available, you'll have to handle that yourself by
110 storing them in a hash.
111 + Doc : More examples in a new COOKBOOK section.
112 + Fix : Compatiblity with the 5.11.0 release. The new compatibility
113 constant VMG_COMPAT_ARRAY_PUSH_NOLEN_VOID was added to cover
115 + Fix : Work around Kwalitee test misfailures.
117 0.37 2009-08-25 15:25 UTC
118 + Add : The new constant VMG_FORKSAFE can be tested to know whether the
119 module will behave nicely when fork()ing. It's currently always
120 true except on Windows where you need perl 5.10.1 for it to be
122 + Doc : Nits and clarifications.
124 0.36 2009-07-05 16:30 UTC
125 + Chg : getdata() now returns an empty list when no magic is present.
126 It used to return undef but did not croak as stated in the doc.
127 Thanks Matt S. Trout for pointing this out.
128 + Fix : Building and tests with ActiveState Perl 5.8 build >= 822.
130 0.35 2009-05-15 20:50 UTC
131 This is a maintenance release. Users of 0.34 can skip this update.
132 + Chg : uvar magic used to be enabled for 5.9.5. It now requires 5.10 or
134 + Fix : Silence compilation warnings for cxinc().
135 + Rem : The workaround introduced in 0.31 for the "Unknown errors" issue
136 was removed, as regression tests show that 0.34 provides a
138 + Tst : Skip t/40-threads.t on 5.8.x with old versions of threads and
141 0.34 2009-04-19 16:55 UTC
142 + Fix : Destruction of wizards in require.
143 + Fix : "panic: restartop" in 5.10 when dieing at compile time triggers
146 0.33 2009-03-26 00:00 UTC
147 + Chg : cast() and getsig() now croak too when an invalid signature is
149 + Fix : It was possible to generate with gensig() the same signature
150 twice before actually defining the wizards.
151 + Tst : More stash tests.
152 + Tst : Test when the magic takes place in some cases.
153 + Tst : Improved coverage for the op object feature.
155 0.32 2009-03-01 13:45 UTC
156 + Chg : dispell() and getdata() now croak when an invalid signature is
158 + Doc : More examples and a brand new synopsis.
159 + Fix : The signature is no longer stored in the mg_private member of
160 the MAGIC struct. This fixes possible clashes with magics from
162 + Fix : op info objects weren't blessed into the right class.
163 + Fix : Races when initializing the op names cache under threads.
165 0.31 2009-02-19 23:50 UTC
166 + Doc : A somewhat better introduction. The FUNCTION section was moved
167 before CONSTANTS and HISTORY since it's more important. And
168 'assignation' is really 'assignment', says Ricardo. :)
169 + Fix : The op_info features are now thread safe.
170 + Fix : Segfaults that occured when trying to get the op name during
172 + Fix : Segfaults and wrong "Unknown error" exceptions that happened
173 when dieing in require caused a free callback to fire.
174 Thanks Florian Ragwitz for reporting.
176 0.30 2009-02-12 17:05 UTC
177 + Add : You can now get in all the callbacks either the name or a B::OP
178 object representing the current op if you pass the new 'op_info'
179 option to wizard() with respective values VMG_OP_INFO_NAME and
181 + Doc : Document magic on array/hashes values.
182 + Fix : Completely skip free callbacks during global destruction.
183 + Tst : Tests now pass again on 5.8.3 and lower (they were broken since
185 + Tst : Test magic on symbol table and on array/hash values.
187 0.29 2009-02-08 11:10 UTC
188 + Add : Set $_[2] to the default length for len magic on scalars.
189 + Chg : Perl 5.8 is required.
190 + Chg : The CLONE method will no longer be defined for non-threaded
193 + Fix : The logic for turning off the GMAGICAL flag on hashes when no
194 uvar magic is involved could do nasty things on tied hashes.
195 + Upd : More resources in META.yml.
197 0.28 2009-01-24 17:05 UTC
198 + Add : The 'copy_key' option to wizard(), that enables the safe edition
199 of $_[2] in uvar callbacks to redirect actions to another key.
200 + Fix : Exception objects thrown from callbacks at compile time were not
202 + Fix : free callbacks triggered during global destruction could lead to
203 segfaults because the wizard was already freed.
204 + Fix : Refcount of scalars with free callback.
205 + Fix : Segfaults when using get or uvar magic simultaneously with clear
207 + Doc : The documentation about what the callbacks take and should
208 return is now somewhat clearer.
210 0.27 2009-01-18 23:40 UTC
211 + Fix : Offset of 1 with len magic on scalars.
212 + Fix : Segfaults that seem to happen when one croak in a callback at
213 compile time (thanks Ash Berlin for the testcase).
214 + Upd : Resources in META.yml (Florian Ragwitz)
215 + Tst : Large cleanup resulting in a wider set of tested behaviours.
217 0.26 2008-11-29 22:05 UTC
218 + Fix : Leaks of cloned coderefs that access lexicals.
219 + Fix : Building with maint-5.8 and blead (thanks Andreas Koenig).
221 0.25 2008-11-12 19:40 UTC
222 + Fix : Compatibility with 5.8.9.
223 + Fix : Old Pod::Coverage don't ignore CLONE.
224 + Upd : META.yml spec updated to 1.4.
226 0.24 2008-10-12 14:55 UTC
227 + Fix : Really fix it.
229 0.23 2008-10-11 17:25 UTC
230 + Fix : Building on perls with multiplicity but not ithreads.
232 0.22 2008-09-29 19:40 UTC
233 + Doc : Explain how magic differs from overloading.
234 + Fix : Disable thread safety for perl 5.8 on Windows. If you want it
235 enabled, reliable feedback is welcome.
237 0.21_02 2008-09-28 23:10 UTC
238 + Fix : Hide previously deleted magic tokens in vmg_svt_free (Yuval
241 0.21_01 2008-09-22 13:25 UTC
242 + Chg : Shrink the context data.
243 + Fix : Enable thread safety features for perls that just lack
245 + Tst : Output threads and threads::shared versions in threads tests.
247 0.20 2008-09-20 23:35 UTC
248 + Fix : The module is now thread safe. You can test if it had been
249 built with thread safety features enabled through the
250 VMG_THREADSAFE constant.
251 + Fix : Correct backtrace when wizard() croaks.
252 + Fix : 'cast $a, undef' should throw an 'invalid wizard' error instead
253 of an 'invalid signature'.
255 0.19 2008-07-05 15:05 UTC
256 + Add : The patchlevel with which the module has been built is now
257 exported with the VMG_PERL_PATCHLEVEL constant.
258 + Chg : Version checking macros overhaul.
259 + Fix : 'uvar' magic should only be enabled in maint-5.10 and higher
260 since it's never going to be backported to 5.8.
261 + Fix : Likewise, the copy callback API change should only be enabled in
263 + Fix : Instead of those ackward fake refcount increments in the wizard
264 destructor, flag the wizard with SVf_BREAK and force temps
265 cleanup. This should fix the wizard destruction for all perls,
266 and solves the leak introduced by the previous solution.
268 0.18 2008-07-04 12:20 UTC
269 + Fix : Enable uvar magic when patchlevel >= 28419.
270 + Fix : Set VMG_COMPAT_ARRAY_PUSH_NOLEN for Windows ActiveState Perl
273 0.17 2008-05-11 09:05 UTC
274 + Fix : 5.8.4 and lower don't have mPUSHi. Gah. Yeah, I know, I should
277 0.16 2008-05-10 22:05 UTC
278 + Add : The samples/copy.pl script.
279 + Chg : The sv_magical() + vmg_mg_magical() combo was simplified into
281 + Tst : t/33-code.t was testing scalars, not code.
282 + Tst : is() and like() are better than ok().
283 + Tst : 100% coverage reached.
285 0.15 2008-04-11 18:25 UTC
286 + Chg : Factor vmg_cb_call{1,2,3}() into one function.
287 + Fix : len magic is no longer called when taking the length() of a
288 magical scalar since p32969. The VMG_COMPAT_SCALAR_LENGTH_NOLEN
289 constant was added to cover this.
290 + Tst : More tests for t/22-len.t.
292 0.14 2008-03-24 12:35 UTC
293 + Fix : t/16-huf.t failures on Solaris and FreeBSD caused by not
294 updating mg->mg_ptr after Renew-ing it on dispell.
295 + Fix : Undefining MGf_DUP caused the wizard's magic to be wrongly
296 initialized. MGf_DUP now appears as true from the user point of
297 view, but the dup callback isn't actually set.
298 + Fix : Warnings with blead due to copy callbacks now taking an I32 (API
300 + Fix : vmg_svt_val() and vmg_uvar_del() aren't required if no uvar
302 + Tst : Tests now output the version of optional modules loaded.
304 0.13 2008-03-19 14:35 UTC
305 + Doc : Link to coverage report.
306 + Fix : Correct dependencies listing in META.yml.
307 + Tst : Improved test coverage.
308 + Tst : Print the patchlevel as a comment.
309 + Tst : Use is() where it's relevant.
310 + Tst : t/16-huf.t now really tests interaction with H::U::FH.
312 0.12 2008-02-07 18:15 UTC
313 + Fix : POD error. Thanks to Chris Williams (BinGOs) for the quick
316 0.11 2008-02-07 17:55 UTC
317 + Add : Copy callbacks now receive the current key/index in $_[2].
318 The current value/element is now in $_[3].
319 + Chg : The trigger of copy magic on list assignment with perls greater
320 than 5.10.0 was caused by uvar magic. Hence,
321 VMG_COMPAT_HASH_LISTASSIGN_COPY was removed. Use VMG_UVAR
323 + Fix : Build failures on patched perls.
324 + Tst : Added missing exported symbols to 01-import.t.
326 0.10 2008-02-04 11:30 UTC
327 + Add : New script : samples/vm_vs_tie.pl, that benchmarks our uvar
328 magic versus tied hashes.
329 + Add : The VMG_COMPAT_* constants can be used from userspace to check
330 perl magic abilities.
331 + Fix : Callbacks that returned undef made us croak, breaking the
332 variable behaviour (regression test in 14-callbacks.t).
333 + Fix : uvar callbacks weren't tested for non-NULL-ity before being
334 called (regression test in 28-uvar.t).
335 + Tst : Fix typo in 25-copy.t that prevented Tie::Hash tests to be ran.
337 0.09 2008-02-02 11:30 UTC
338 + Doc : Explicitely say that uvar callbacks are safely ignored for
340 + Doc : Document caveats and fix the usual set of typos.
341 + Fix : vmg_dispell() didn't check if the ext magic were ours when
342 counting wizards that have uvar callbacks, resulting in a
343 possible memory misread.
344 + Fix : getdata() now returns directly the data object, and no longer a
345 copy. This caused a leak.
346 + Tst : Prefix author tests by 9*-.
347 + Tst : New optional author test : 95-portability-files.t, that uses
348 Test::Portability::Files when it's present.
349 + Tst : New test : 14-self.t, that tests application of magic on the
351 + Tst : Move Hash::Util::FieldHash tests out of 11-multiple.t to
354 0.08 2008-02-01 16:55 UTC
355 + Add : copy magic for tied arrays/hashes.
357 + Add : uvar magics : fetch, store, exists, delete for hashes.
360 Internal development versions.
362 0.06 2007-11-20 10:10 UTC
363 + Chg : 5.7.3 is now officially required.
364 + Fix : "data" test failures on 5.8.{0,2}.
365 + Fix : Drand01() vs rand().
367 0.05 2007-11-19 09:10 UTC
368 + Fix : 5.10.0_RC1 compatibility fix.
370 0.04 2007-08-28 12:25 UTC
371 + Chg : s/require (XSLoader)/use $1/.
372 + Fix : Tests are now strict.
373 + Fix : Complete dependencies.
375 0.03 2007-08-01 17:20 UTC
376 + Add : Passing the signature of an already defined magic to wizard()
377 now returns the corresponding magic object.
378 + Add : You can pass the numeric signature as the wizard argument of
379 cast(), dispell() and getdata().
380 + Add : Any argument specified after the wizard (or the signature) in a
381 call to cast() is now passed to the private data constructor in
383 + Chg : $_[0] is now always a reference to the magic variable in all
384 callbacks. The reason for changing from the previous behaviour
385 is that one may want to apply the same magic to a plain scalar
386 and to a scalar reference, and needs a way to distinguish
387 between them in the callback (say, ref()).
388 + Fix : Wizard object destruction used not to free the signature.
390 0.02 2007-07-27 13:50 UTC
391 + Fix : In response to test report 548152 :
392 Newx() and SvMAGIC_set() not present on older perls.
393 + Fix : In response to test report 548275 :
394 Since perl 5.9.5, 'clear' magic is invoked when an array is
395 undefined (bug #43357). Moreover, 'len' magic is no longer
396 called by pushing an element since perl 5.9.3.
397 + Fix : Missing glob test in MANIFEST.
399 0.01 2007-07-25 16:15 UTC
400 First version, released on an unsuspecting world.