1 Revision history for autovivification
3 0.17 2017-07-31 17:15 UTC
4 + Chg : A large chunk of boilerplate XS code, which is also used in
5 other XS modules, has been factored out of the main .xs file
6 to a collection of .h files in the xsh subdirectory.
7 + Fix : The new optimization in perl 5.27.3 for scalar(keys(%$hashref))
8 is now correcty supported.
10 0.16 2015-07-01 14:30 UTC
11 + Fix : Broken linkage on Windows.
13 0.15 2015-05-14 20:35 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 : The module has been taught about the new multideref
18 optimization, and will function properly on perl 5.21.11 and
20 + Fix : The module will no longer crash or behave erratically when it
21 is loaded by several threads (or Windows emulated processes)
23 + Fix : Update the Windows ActivePerl + gcc 3.4 workaround for
24 ExtUtils::MakeMaker 7.04. Thanks Christian Walde for reporting
25 and feedback on this issue.
26 + Fix : Be really compatible with the optional OP_PARENT feature.
27 + Tst : $ENV{$Config{ldlibpthname}} is now preserved on all platforms,
28 which will address failures of t/43-peep.t and
29 t/51-threads-teardown.t with unusual compilers (like icc) that
30 link all their compiled objects to their own libraries.
32 0.14 2014-11-01 22:25 UTC
33 + Fix : [RT #99458] : AnyEvent::Loop hangs if autovivification 0.13 is
35 [RT #99904] : 'no autovivification' breaks Lingua::EN::Inflect
37 This was a regression in 0.13 which caused the peephole
38 optimizer to loop infinitely on "... while 1" constructs.
39 However, the Lingua::EN::Inflect problem only occured in perl
41 Thanks Toshio Ito and Kent Fredric for reporting.
43 0.13 2014-10-04 16:55 UTC
44 This release contains a change that, while being very likely to be safe,
45 can potentially cause freezes during code compilation. Every release
46 should be carefully tested before being put in production, but this is
47 especially true for this one.
48 + Add : Support for the PERL_OP_PARENT optional feature introduced in
50 + Doc : The CAVEATS section now warns about the global slowdown during
51 compilation caused by this pragma.
52 + Fix : [RT #97703] : Android support
53 t/51-threads-teardown.t will no longer fail on Android.
54 Thanks Brian Fraser for reporting.
55 + Fix : Segfaults in eval in an END block of a Win32 pseudo-fork.
56 + Fix : Segfaults during global destruction of a thread or a
58 + Opt : The global slowdown caused by this module has been greatly
60 Thanks Ævar Arnfjörð Bjarmason for reporting and testing the
63 0.12 2013-09-05 17:20 UTC
64 + Fix : Check functions are now replaced and restored in a thread-safe
65 manner, either by using the wrap_op_checker() function from perl
66 when it is available (starting from perl 5.16) or by taking the
67 OP_REFCNT mutex on older perls.
68 + Tst : Author tests are no longer bundled with this distribution.
69 They are only made available to authors in the git repository.
71 0.11 2013-02-08 19:25 UTC
72 This is a maintenance release. The code contains no functional change.
73 Satisfied users of version 0.10 can skip this update.
74 + Doc : POD tweaks and clarifications.
75 + Tst : Threads tests will not fail anymore if resources constraints
76 prevent the system from creating all the required threads.
77 + Tst : Author tests overhaul.
79 0.10 2011-08-24 15:00 UTC
80 + Fix : The pragma no longer vivifies the "autovivification" entry
81 in the hints hash %^H on perl 5.8.
83 0.09 2011-01-05 18:40 UTC
84 + Fix : [RT #64435] : Hangs with File::Copy in Config.pm.
85 This was actually a regression introduced together with the new
86 peephole optimizer strategy, and that caused the pragma to hang
87 on constructs like "for (;;) { ... }".
88 Thanks Michael Schilli for reporting.
90 0.08 2011-01-03 21:00 UTC
91 + Fix : Building on Windows.
93 0.07 2010-12-31 16:20 UTC
94 + Chg : perl 5.8.3 is required.
95 + Doc : Complements and clarifications.
96 + Fix : Segmentation faults and misbehaviours in threaded applications.
97 + Fix : Compatibility with perl 5.13.7.
98 Thanks Andreas J. König for reporting and Andrew Main for
100 + Fix : Broken linkage on Windows with gcc 3.4, which appears in
101 particular when using ActivePerl's default compiler suite.
102 For those setups, the autovivification shared library will now
103 be linked against the perl dll directly (instead of the import
105 + Opt : The pragma takes slightly more time at compile-time, but is
106 slightly faster at run-time.
107 + Tst : Lengthy tests have been ported to Test::Leaner, making the
108 whole test suite about 50% faster.
109 + Tst : Threads tests are now only run on perl 5.13.4 and higher.
110 They could segfault randomly because of what seems to be an
111 internal bug of Perl, which has been addressed in 5.13.4.
112 There is also an environment variable that allows you to
113 forcefully run those tests, but it should be set only for
114 author testing and not for end users.
116 0.06 2010-04-24 17:40 UTC
117 + Add : The A_THREADSAFE and A_FORKSAFE constants.
118 + Fix : [RT #56870] : "no autovivification" vs Regexp::Common.
119 This was a bug in how tied arrays and hashes were handled.
120 Thanks Michael G. Schwern for reporting.
121 + Fix : Scope leaks under perl 5.8-5.10.0.
122 + Fix : Segfaults when first loading the pragma from inside a thread.
124 0.05 2010-03-05 23:15 UTC
125 + Fix : [RT #55154] : Crashes and assertion failures when deparsing and
126 re-eval-uating some code compiled while autovivification was in
128 Thanks Michael G. Schwern for reporting.
129 + Fix : [RT #53647] : "leys" typo in pod.
130 Thanks Hinrik Orn Sigurdsson for reporting.
132 0.04 2010-01-10 00:30 UTC
133 + Add : Array and hash slices are now handled by the pragma.
134 + Fix : Work around Kwalitee test misfailures.
136 0.03 2009-06-23 22:20 UTC
137 + Add : Handle old-fashion dereferencing (like $$hashref{key}).
138 + Chg : Aliasing constructs (for ($x{foo}) { ... }) are now covered by
139 the 'store' category (and no longer the 'fetch' one). This is
140 because there's no way to know at compile-time if the alias
142 + Fix : Quadratic complexity at compile-time.
143 + Fix : Segfaults when dereferencing globals.
144 + Fix : Segfaults on big-endian systems.
145 + Tst : Really test plain arrays and hashes.
146 + Tst : Improved coverage.
148 0.02 2009-06-17 18:05 UTC
149 + Add : 'fetch' also applies to aliasing ("for ($hashref->{key}) { }").
150 + Fix : Don't segfault on "keys/values %$hashref", and don't vivify if
152 + Fix : Plain dereferencing shouldn't have a different behaviour when
153 the pragma is in use.
154 + Tst : Improved coverage.
156 0.01 2009-06-14 20:10 UTC
157 First version, released on an unsuspecting world.