1 Revision history for Variable-Magic
3 0.47 2011-10-27 16:55 UTC
4 + Add : The new constant VMG_COMPAT_HASH_DELETE_NOUVAR_VOID evaluates
5 to true when "delete $hash{key}" does not call 'delete' uvar
7 + Chg : The MAGIC tokens created by this module no longer use the
8 mg_private member for storing a "magical" signature.
9 + Fix : Triggering magic in a thread on a variable cloned from the main
10 interpreter, and when the wizard already went out of scope,
11 will not segfault anymore.
12 + Opt : The pointer table is no longer needed to ensure thread safety.
13 The size of the object code is about 8% smaller for threaded
15 + Tst : Threads tests will not fail anymore if resources constraints
16 prevent the system from creating all the required threads.
18 0.46 2011-01-23 16:45 UTC
19 + Fix : [RT #64866] : Assertion failure with perl 5.13.9.
20 The real problem was that dispell() wasn't properly resetting
21 the magical sv flags, which became visible in the test suite
22 only with perl 5.13.9.
23 Thanks Joshua ben Jore for reporting.
25 0.45 2010-11-21 23:15 UTC
26 This is a maintenance release. The code contains no functional change.
27 Users of 0.44 can skip this update.
28 + Doc : C++ compilers are officially NOT supported.
29 + Doc : The minimum perl 5.10.0 requirement for uvar magic has been
31 Thanks Peter Rabbitson for pointing this out and contributing a
33 + Tst : Tune for perl 5.13.7.
34 + Tst : Capture::Tiny will be used in t/17-ctl.t if and only if it can
37 0.44 2010-09-24 19:10 UTC
38 + Fix : Broken linkage on Windows with gcc 3.4, which appears in
39 particular when using ActivePerl's default compiler suite.
40 For those setups, the Variable::Magic shared library will now
41 be linked against the perl dll directly (instead of the import
42 library). This (should) fix RT #51483.
43 Thanks Christian Walde for helping to reproduce this failure
45 + Rem : Support for development perls from the 5.11 branch but older
46 than the 5.11.0 release was removed. This could cause more
47 recent setups to fail.
48 + Tst : Threads tests are now only run on perl 5.13.4 and higher.
49 They could segfault randomly because of what seems to be an
50 internal bug of Perl, which has been addressed in 5.13.4.
51 There is also an environment variable that allows you to
52 forcefully run those tests, but it should be set only for
53 author testing and not for end users.
55 0.43 2010-06-25 23:35 UTC
56 + Add : The new constant VMG_COMPAT_GLOB_GET tells you whether get magic
57 is called for globs. It's true starting perl 5.13.2.
58 + Chg : All callbacks are now called within an eval-like context.
59 Only free callbacks used to be called that way.
60 + Fix : Some exceptions thrown from a free callback could be lost.
61 + Fix : Croak messages could sometimes be repeated several times.
62 + Fix : t/41-clone.t segfaulting with perl 5.13.2.
64 0.42 2010-05-19 00:15 UTC
65 This is a maintenance release. The code contains no functional change.
66 Users of 0.41 can skip this update.
67 + Fix : Test failures with perl 5.13.
68 + Tst : Improve coverage.
70 0.41 2010-03-15 17:35 UTC
71 + Doc : Tweaks and fixups.
73 + Fix : Compatibility with the soon-to-be-released perl 5.12.0.
74 + Fix : Correctly propagate the errors thrown when variable destruction
75 happens at compile-time and not from inside eval STRING.
76 Thanks Florian Ragwitz and Ash Berlin for reporting.
78 0.40 2010-01-06 23:20 UTC
79 + Fix : Possible memory miswrites when passing data arguments to cast().
80 + Fix : Minor C portability tweaks.
82 0.39 2009-12-01 00:05 UTC
83 + Add : You can use a function name as a callback by passing a string
84 reference to wizard() instead of a code reference.
85 + Fix : Compatiblity with perl 5.11.1 and 5.11.2.
86 + Fix : Scalars stored into the data slot no longer leak.
87 + Fix : Thread destruction should not segfault anymore.
88 + Opt : As a result of removing the deprecated features, less memory is
89 used for both threaded (a pointer table is used instead of a
90 hash) and non-threaded (no global structure is needed anymore)
92 + Rem : As advertised in the precendent version, all the signature
93 related features were removed. This includes the 'sig' option
94 to wizard(), the getsig() and gensig() functions, and the SIG_*
97 0.38 2009-10-04 16:10 UTC
98 + Dep : All the signature-related features are DEPRECATED and will be
99 removed in december 2009. This includes the 'sig' option to
100 wizard(), the getsig() and gensig() functions, and the SIG_*
102 Afaik those features were never used in real life, and they make
103 the XS code slow and brittle. If you want your wizards to be
104 globally available, you'll have to handle that yourself by
105 storing them in a hash.
106 + Doc : More examples in a new COOKBOOK section.
107 + Fix : Compatiblity with the 5.11.0 release. The new compatibility
108 constant VMG_COMPAT_ARRAY_PUSH_NOLEN_VOID was added to cover
110 + Fix : Work around Kwalitee test misfailures.
112 0.37 2009-08-25 15:25 UTC
113 + Add : The new constant VMG_FORKSAFE can be tested to know whether the
114 module will behave nicely when fork()ing. It's currently always
115 true except on Windows where you need perl 5.10.1 for it to be
117 + Doc : Nits and clarifications.
119 0.36 2009-07-05 16:30 UTC
120 + Chg : getdata() now returns an empty list when no magic is present.
121 It used to return undef but did not croak as stated in the doc.
122 Thanks Matt S. Trout for pointing this out.
123 + Fix : Building and tests with ActiveState Perl 5.8 build >= 822.
125 0.35 2009-05-15 20:50 UTC
126 This is a maintenance release. Users of 0.34 can skip this update.
127 + Chg : uvar magic used to be enabled for 5.9.5. It now requires 5.10 or
129 + Fix : Silence compilation warnings for cxinc().
130 + Rem : The workaround introduced in 0.31 for the "Unknown errors" issue
131 was removed, as regression tests show that 0.34 provides a
133 + Tst : Skip t/40-threads.t on 5.8.x with old versions of threads and
136 0.34 2009-04-19 16:55 UTC
137 + Fix : Destruction of wizards in require.
138 + Fix : "panic: restartop" in 5.10 when dieing at compile time triggers
141 0.33 2009-03-26 00:00 UTC
142 + Chg : cast() and getsig() now croak too when an invalid signature is
144 + Fix : It was possible to generate with gensig() the same signature
145 twice before actually defining the wizards.
146 + Tst : More stash tests.
147 + Tst : Test when the magic takes place in some cases.
148 + Tst : Improved coverage for the op object feature.
150 0.32 2009-03-01 13:45 UTC
151 + Chg : dispell() and getdata() now croak when an invalid signature is
153 + Doc : More examples and a brand new synopsis.
154 + Fix : The signature is no longer stored in the mg_private member of
155 the MAGIC struct. This fixes possible clashes with magics from
157 + Fix : op info objects weren't blessed into the right class.
158 + Fix : Races when initializing the op names cache under threads.
160 0.31 2009-02-19 23:50 UTC
161 + Doc : A somewhat better introduction. The FUNCTION section was moved
162 before CONSTANTS and HISTORY since it's more important. And
163 'assignation' is really 'assignment', says Ricardo. :)
164 + Fix : The op_info features are now thread safe.
165 + Fix : Segfaults that occured when trying to get the op name during
167 + Fix : Segfaults and wrong "Unknown error" exceptions that happened
168 when dieing in require caused a free callback to fire.
169 Thanks Florian Ragwitz for reporting.
171 0.30 2009-02-12 17:05 UTC
172 + Add : You can now get in all the callbacks either the name or a B::OP
173 object representing the current op if you pass the new 'op_info'
174 option to wizard() with respective values VMG_OP_INFO_NAME and
176 + Doc : Document magic on array/hashes values.
177 + Fix : Completely skip free callbacks during global destruction.
178 + Tst : Tests now pass again on 5.8.3 and lower (they were broken since
180 + Tst : Test magic on symbol table and on array/hash values.
182 0.29 2009-02-08 11:10 UTC
183 + Add : Set $_[2] to the default length for len magic on scalars.
184 + Chg : Perl 5.8 is required.
185 + Chg : The CLONE method will no longer be defined for non-threaded
188 + Fix : The logic for turning off the GMAGICAL flag on hashes when no
189 uvar magic is involved could do nasty things on tied hashes.
190 + Upd : More resources in META.yml.
192 0.28 2009-01-24 17:05 UTC
193 + Add : The 'copy_key' option to wizard(), that enables the safe edition
194 of $_[2] in uvar callbacks to redirect actions to another key.
195 + Fix : Exception objects thrown from callbacks at compile time were not
197 + Fix : free callbacks triggered during global destruction could lead to
198 segfaults because the wizard was already freed.
199 + Fix : Refcount of scalars with free callback.
200 + Fix : Segfaults when using get or uvar magic simultaneously with clear
202 + Doc : The documentation about what the callbacks take and should
203 return is now somewhat clearer.
205 0.27 2009-01-18 23:40 UTC
206 + Fix : Offset of 1 with len magic on scalars.
207 + Fix : Segfaults that seem to happen when one croak in a callback at
208 compile time (thanks Ash Berlin for the testcase).
209 + Upd : Resources in META.yml (Florian Ragwitz)
210 + Tst : Large cleanup resulting in a wider set of tested behaviours.
212 0.26 2008-11-29 22:05 UTC
213 + Fix : Leaks of cloned coderefs that access lexicals.
214 + Fix : Building with maint-5.8 and blead (thanks Andreas Koenig).
216 0.25 2008-11-12 19:40 UTC
217 + Fix : Compatibility with 5.8.9.
218 + Fix : Old Pod::Coverage don't ignore CLONE.
219 + Upd : META.yml spec updated to 1.4.
221 0.24 2008-10-12 14:55 UTC
222 + Fix : Really fix it.
224 0.23 2008-10-11 17:25 UTC
225 + Fix : Building on perls with multiplicity but not ithreads.
227 0.22 2008-09-29 19:40 UTC
228 + Doc : Explain how magic differs from overloading.
229 + Fix : Disable thread safety for perl 5.8 on Windows. If you want it
230 enabled, reliable feedback is welcome.
232 0.21_02 2008-09-28 23:10 UTC
233 + Fix : Hide previously deleted magic tokens in vmg_svt_free (Yuval
236 0.21_01 2008-09-22 13:25 UTC
237 + Chg : Shrink the context data.
238 + Fix : Enable thread safety features for perls that just lack
240 + Tst : Output threads and threads::shared versions in threads tests.
242 0.20 2008-09-20 23:35 UTC
243 + Fix : The module is now thread safe. You can test if it had been
244 built with thread safety features enabled through the
245 VMG_THREADSAFE constant.
246 + Fix : Correct backtrace when wizard() croaks.
247 + Fix : 'cast $a, undef' should throw an 'invalid wizard' error instead
248 of an 'invalid signature'.
250 0.19 2008-07-05 15:05 UTC
251 + Add : The patchlevel with which the module has been built is now
252 exported with the VMG_PERL_PATCHLEVEL constant.
253 + Chg : Version checking macros overhaul.
254 + Fix : 'uvar' magic should only be enabled in maint-5.10 and higher
255 since it's never going to be backported to 5.8.
256 + Fix : Likewise, the copy callback API change should only be enabled in
258 + Fix : Instead of those ackward fake refcount increments in the wizard
259 destructor, flag the wizard with SVf_BREAK and force temps
260 cleanup. This should fix the wizard destruction for all perls,
261 and solves the leak introduced by the previous solution.
263 0.18 2008-07-04 12:20 UTC
264 + Fix : Enable uvar magic when patchlevel >= 28419.
265 + Fix : Set VMG_COMPAT_ARRAY_PUSH_NOLEN for Windows ActiveState Perl
268 0.17 2008-05-11 09:05 UTC
269 + Fix : 5.8.4 and lower don't have mPUSHi. Gah. Yeah, I know, I should
272 0.16 2008-05-10 22:05 UTC
273 + Add : The samples/copy.pl script.
274 + Chg : The sv_magical() + vmg_mg_magical() combo was simplified into
276 + Tst : t/33-code.t was testing scalars, not code.
277 + Tst : is() and like() are better than ok().
278 + Tst : 100% coverage reached.
280 0.15 2008-04-11 18:25 UTC
281 + Chg : Factor vmg_cb_call{1,2,3}() into one function.
282 + Fix : len magic is no longer called when taking the length() of a
283 magical scalar since p32969. The VMG_COMPAT_SCALAR_LENGTH_NOLEN
284 constant was added to cover this.
285 + Tst : More tests for t/22-len.t.
287 0.14 2008-03-24 12:35 UTC
288 + Fix : t/16-huf.t failures on Solaris and FreeBSD caused by not
289 updating mg->mg_ptr after Renew-ing it on dispell.
290 + Fix : Undefining MGf_DUP caused the wizard's magic to be wrongly
291 initialized. MGf_DUP now appears as true from the user point of
292 view, but the dup callback isn't actually set.
293 + Fix : Warnings with blead due to copy callbacks now taking an I32 (API
295 + Fix : vmg_svt_val() and vmg_uvar_del() aren't required if no uvar
297 + Tst : Tests now output the version of optional modules loaded.
299 0.13 2008-03-19 14:35 UTC
300 + Doc : Link to coverage report.
301 + Fix : Correct dependencies listing in META.yml.
302 + Tst : Improved test coverage.
303 + Tst : Print the patchlevel as a comment.
304 + Tst : Use is() where it's relevant.
305 + Tst : t/16-huf.t now really tests interaction with H::U::FH.
307 0.12 2008-02-07 18:15 UTC
308 + Fix : POD error. Thanks to Chris Williams (BinGOs) for the quick
311 0.11 2008-02-07 17:55 UTC
312 + Add : Copy callbacks now receive the current key/index in $_[2].
313 The current value/element is now in $_[3].
314 + Chg : The trigger of copy magic on list assignment with perls greater
315 than 5.10.0 was caused by uvar magic. Hence,
316 VMG_COMPAT_HASH_LISTASSIGN_COPY was removed. Use VMG_UVAR
318 + Fix : Build failures on patched perls.
319 + Tst : Added missing exported symbols to 01-import.t.
321 0.10 2008-02-04 11:30 UTC
322 + Add : New script : samples/vm_vs_tie.pl, that benchmarks our uvar
323 magic versus tied hashes.
324 + Add : The VMG_COMPAT_* constants can be used from userspace to check
325 perl magic abilities.
326 + Fix : Callbacks that returned undef made us croak, breaking the
327 variable behaviour (regression test in 14-callbacks.t).
328 + Fix : uvar callbacks weren't tested for non-NULL-ity before being
329 called (regression test in 28-uvar.t).
330 + Tst : Fix typo in 25-copy.t that prevented Tie::Hash tests to be ran.
332 0.09 2008-02-02 11:30 UTC
333 + Doc : Explicitely say that uvar callbacks are safely ignored for
335 + Doc : Document caveats and fix the usual set of typos.
336 + Fix : vmg_dispell() didn't check if the ext magic were ours when
337 counting wizards that have uvar callbacks, resulting in a
338 possible memory misread.
339 + Fix : getdata() now returns directly the data object, and no longer a
340 copy. This caused a leak.
341 + Tst : Prefix author tests by 9*-.
342 + Tst : New optional author test : 95-portability-files.t, that uses
343 Test::Portability::Files when it's present.
344 + Tst : New test : 14-self.t, that tests application of magic on the
346 + Tst : Move Hash::Util::FieldHash tests out of 11-multiple.t to
349 0.08 2008-02-01 16:55 UTC
350 + Add : copy magic for tied arrays/hashes.
352 + Add : uvar magics : fetch, store, exists, delete for hashes.
355 Internal development versions.
357 0.06 2007-11-20 10:10 UTC
358 + Chg : 5.7.3 is now officially required.
359 + Fix : "data" test failures on 5.8.{0,2}.
360 + Fix : Drand01() vs rand().
362 0.05 2007-11-19 09:10 UTC
363 + Fix : 5.10.0_RC1 compatibility fix.
365 0.04 2007-08-28 12:25 UTC
366 + Chg : s/require (XSLoader)/use $1/.
367 + Fix : Tests are now strict.
368 + Fix : Complete dependencies.
370 0.03 2007-08-01 17:20 UTC
371 + Add : Passing the signature of an already defined magic to wizard()
372 now returns the corresponding magic object.
373 + Add : You can pass the numeric signature as the wizard argument of
374 cast(), dispell() and getdata().
375 + Add : Any argument specified after the wizard (or the signature) in a
376 call to cast() is now passed to the private data constructor in
378 + Chg : $_[0] is now always a reference to the magic variable in all
379 callbacks. The reason for changing from the previous behaviour
380 is that one may want to apply the same magic to a plain scalar
381 and to a scalar reference, and needs a way to distinguish
382 between them in the callback (say, ref()).
383 + Fix : Wizard object destruction used not to free the signature.
385 0.02 2007-07-27 13:50 UTC
386 + Fix : In response to test report 548152 :
387 Newx() and SvMAGIC_set() not present on older perls.
388 + Fix : In response to test report 548275 :
389 Since perl 5.9.5, 'clear' magic is invoked when an array is
390 undefined (bug #43357). Moreover, 'len' magic is no longer
391 called by pushing an element since perl 5.9.3.
392 + Fix : Missing glob test in MANIFEST.
394 0.01 2007-07-25 16:15 UTC
395 First version, released on an unsuspecting world.