1 2010-01-27 10:27:00 TonyC
3 * fix a format string and arguments for some debugging text
4 * linkify references to alternatives to Storable
6 2020-01-27 11:01:00 TonyC
8 * add casts to match some I32 parameters to "%d" formats (#17339)
9 * fix dependencies in Makefile.PL -> META (#17422)
10 * make use of note() optional, this requires a newer version of
11 Test::More and there's a circular dependency between later
12 versions of Test::More and Storable (#17422)
14 2019-11-19 07:59:39 TonyC
16 * update bug tracker to point at github (#17298)
17 * disallow vstring magic strings over 2GB-1 (#17306)
18 * mark some ASCII dependent tests as ASCII platform only
20 2019-08-08 11:48:00 TonyC
22 * correct a data type to ensure the check for too large results from
23 STORABLE_freeze() are detected correctly (detected by Coverity)
24 * removed remains of stack size detection from the build process.
25 * moved CAN_FLOCK detection into XS to simplify the build process.
27 2019-06-11 10:43:00 TonyC
29 * (perl #134179) fix self-referencing structures that include regexps
30 * bless regexps to preserve bless qr//, "Foo"
32 2019-04-23 16:00:00 xsawyerx
38 * (perl #133708) don't build-time probe for stack limits at all
42 * (perl #133411) don't probe for stack limits with -Dusecrosscompile
44 2018-04-27 20:40:00 xsawyerx
46 * Fix Strawberry Perl build failures.
48 2018-04-21 22:00:00 xsawyerx
50 * Fix binary artifacts from distribution.
52 2018-04-21 16:49:00 xsawyerx
54 * Fix "provides" in metadata (META.yml/META.json) to use the Storable
55 template instead of a small other file (which also didn't exist).
57 2018-04-21 11:23:00 xsawyerx
59 * (perl #132849) try to disable core files when deliberatly segfaulting.
60 * (perl #127743) don't probe Storable limits so much.
61 * (perl #132893) don't probe for Storable recursion limits on old Win32.
62 * (perl #132870) workaround VC2017 compiler bug.
63 * (perl #127743) re-work for debugging builds with MSVC.
64 * (perl #133039) dont build a Storable.so/.dll with a static perl build.
66 2018-02-07 15:08:00 tonyc
69 * support large object ids. The code in theory supported arrays
70 with more than 2**32 elements, but references to the elements
71 emitted at the end of the array with be retrieved as references to
73 * 32-bit object ids over 2**31-1 weren't correctly handled.
74 * hook object id generation now supports 64-bit ids where needed
75 * writing 64-bit lengths in network order now works
76 * reading 64-bit lengths in network order now reads the components
78 * retrieving large object tags are now only handled on 64-bit
79 platforms, large object tags should only be emitted for objects
80 that are too large for the 32-bit address space, so it was only
82 * reading 32-bit lengths for LSCALAR and LUTF8STR as unsigned
84 * reading flagged large object hashes didn't read the flags
85 * treat the 32-bit size of hook data as unsigned, values over 2GB
86 were treated as large (close to 2**64) parameters to NEWSV().
88 * added support for hook data over 4GB in size
89 * zero length data receievd from STORABLE_freeze() no longer
90 results in an invalid SV being passed to STORABLE_thaw/_attach()
92 * where practical, padding is now cleared when emitting a long
94 * cache the value of $Storable::DEBUGME (since cperl enabled
95 Storable TRACEME builds for all -DDEBUGGING builds)
96 * no longer discard exceptions thrown by
97 STORABLE_freeze/_thaw/attach() (perl #25933)
98 * fix dependencies used to build Storable.pm from __Storable__.pm
99 * add experimental support for freezing/thawing regular
100 expressions (perl #50608)
101 * re-work recursion limiting to store the limit in a perl variable
102 instead of baked into Storable.$so. This allows static Storable
103 builds to work, and avoids the kind of circular reference on
106 2017-07-24 13:57:13 rurban
109 * mingw fix: use safe defaults, not segfaulting defaults.
110 mingw fails on the stacksize binary search, leaving it empty.
112 Wed Apr 19 09:11:07 2017 +0200 Reini Urban <rurban@cpan.org>
115 * enhance stack reserve from 8 to 16
116 * fix LD_LIBRARY_PATH usage for CORE
117 * fixed some coverity warnings and leaks
118 * added a release make target
120 Wed Mar 29 21:04:28 2017 +0200 Reini Urban <rurban@cpan.org>
123 * croak on sizes read > I32_MAX
124 * simplify last_op_in_netorder
125 * protect from empty retrieve_vstring
126 * protect store_other error buf, potential static
129 Tue Mar 14 09:52:20 2017 +0100 Reini Urban <rurban@cpan.org>
132 * CORE-only improvements to stacksize
134 Thu Mar 9 19:20:19 2017 +0100 Reini Urban <rurban@cpan.org>
137 * compute the stacksizes, improve cleanup within croak
138 from stack exhaustion.
139 * added stack_depth and stack_depth_hash getters.
141 Wed Mar 8 21:03:43 CET 2017 Reini Urban <rurban@cpan.org>
144 * finetune the max stack limit, for C++, DEBUGGING and 32bit.
145 * fix t/blessed.t for cperl5.22
147 Sun Mar 5 13:36:47 2017 +0100 Reini Urban <rurban@cpan.org>
150 * Fixed a podchecker issue
152 Sun Mar 5 11:42:04 2017 +0100 Reini Urban <rurban@cpan.org>
155 * Fixed wrong recursion depth error with large arrays containing
157 L<[cperl #257]|https://github.com/perl11/cperl/issues/257>
159 Thu Feb 2 12:40:44 2017 +0100 Reini Urban <rurban@cpanel.net>
162 * Add leak tests for [cpan #97316], [perl #121928]
163 * Limit the max recursion depth to 1200 on 32bit systems.
164 We have no max_depth option yet, as in JSON::XS.
166 Thu Feb 2 11:59:21 2017 +0100 Reini Urban <rurban@cpanel.net>
169 * Fix retrieve_tied_array which fails since 5.16
171 * Improve t/blessed.t in the creation of sv_yes/sv_no
174 Tue Jan 31 02:55:30 2017 +0100 Reini Urban <rurban@cpanel.net>
177 * Tune t/recurse.t stack-overflow limit more.
179 Mon Jan 30 19:50:29 2017 +0100 Reini Urban <rurban@cpanel.net>
182 * Tune t/recurse.t stack-overflow limit. Small 64bit systems overflow
183 even with depth 3000, where 32bit are ok.
185 Mon Jan 30 15:13:38 2017 +0100 Reini Urban <rurban@cpanel.net>
188 * Protect against stack overflows with nested arrays and hashes
189 [cpan #97526]. This imposes a new limit to your nested structures,
190 but JSON::XS has a limit of 512. We use a max_depth of 3000 for the
191 typical stack limit of 8k.
194 Sun Jan 29 11:36:43 2017 +0100 Reini Urban <rurban@cpanel.net>
197 * Protect against classname len overflow on the stack
198 and 2x on the heap with retrieve_bless and retrieve_hook.
199 A serious security issue with malcrafted storable files or buffers,
200 but p5p accepts no CVE on Storable attacks. See RT #130635
202 * Fix NULL ptr SEGVs with retrieve_code and retrieve_other.
203 See RT #130098 (reported and fixed by JD)
204 * Fix wrong huge LOBJECT support, broken since 3.00c.
205 Repro with `export PERL_TEST_MEMORY=8`
206 * Fix the few remaining 2-arg open calls.
207 * Portability and backport fixes back to 5.6.2
209 Sat Jan 7 09:01:29 2017 +0100 Reini Urban <rurban@cpanel.net>
212 * fix printf types and warnings, esp. for 32bit use64bitint
213 * Change sv_setpvn(…, "…", …) to sv_setpvs(…, "…")
215 Tue Jul 26 11:49:33 2016 +1000 Tony Cook <tony@develop-help.com>
218 * remove . from @INC when loading optional modules
220 Sun Nov 20 18:06:45 2016 +0100 Reini Urban <rurban@cpanel.net>
223 * Fix -Wc++11-compat warnings, fix -Wchar-subscripts
225 Fri Sep 16 01:32:59 2016 +0200 Reini Urban <rurban@cpanel.net>
228 * Added warn_security("Movable-Type CVE-2015-1592 Storable metasploit attack")
229 when detecting the third destructive metasploit vector,
230 thawing bless \"mt-config.cgi", "CGITempFile".
232 Thu Mar 31 17:10:27 2016 +0200 Reini Urban <rurban@cpanel.net>
235 * Added support for u64 strings, arrays and hashes >2G
236 via a new LOBJECT tag. This is for 32bit systems and lengths
237 between 2GB and 4GB (I32-U32), and 64bit (>I32).
238 * Bumped STORABLE_BIN_MINOR and STORABLE_BIN_WRITE_MINOR from 10 to 11
239 * fix parallel tests, use unique filenames.
240 * fixed 2 instances of 2arg open,
241 * added optional flag arguments to skip tie and bless on retrieve/thaw,
242 * added SECURITY WARNING and Large data support to docs
243 * compute CAN_FLOCK at compile-time
244 * reformat everything consistently
245 * enable DEBUGME tracing and asserts with -DDEBUGGING
246 * fix all 64 bit compiler warnings
247 * added some abstraction methods to avoid code duplication
249 ?????? p5p <perl5-porters@perl.org>
252 * Replace multiple 'use vars' by 'our'
253 * remove Config dependency
255 Wed Jul 2 16:25:25 IST 2014 Abhijit Menon-Sen <ams@toroid.org>
258 * [perl #121928] Fix memory leak for dclone inside freeze hook
260 * Do not call DESTROY for empty objects
264 Sat Jul 13 18:34:27 IST 2013 Abhijit Menon-Sen <ams@toroid.org>
267 * [perl #118829] Memory leaks in STORABLE_attach
269 * [perl #118139] Don't SEGV during global destruction
270 (Nicholas Clark, report/test from Reini Urban)
271 * Added security warnings section (Steffen Mueller)
272 * Update INSTALLDIRS to favour installation in 'site'
275 Tue 11 Sep 06:51:11 IST 2012 Abhijit Menon-Sen <ams@toroid.org>
278 Various bugfixes, including compatibility fixes for older
279 versions of Perl and vstring handling.
281 Sun 3 Jul 09:10:11 IST 2011 Abhijit Menon-Sen <ams@toroid.org>
284 Various bugfixes, notably including preventing nfreeze from
285 incorrectly stringifying integers.
287 Fri 3 Dec 14:12:32 GMT 2010 David Leadbeater <dgl@dgl.cx>
290 Support for serializing coderefs containing UTF-8.
292 Fri Nov 12 10:52:19 IST 2010 Abhijit Menon-Sen <ams@toroid.org>
296 Performance improvement for overloaded classes from Benjamin
299 Fri Nov 12 10:36:22 IST 2010 Abhijit Menon-Sen <ams@toroid.org>
303 Release the latest version from the Perl repository.
305 Thu Aug 6 10:55:50 IST 2009 Abhijit Menon-Sen <ams@toroid.org>
309 Includes hints/hpux.pl that was inadvertently left out of 2.20.
311 Mon May 18 09:38:20 IST 2009 Abhijit Menon-Sen <ams@toroid.org>
315 Fix bug handling blessed references to overloaded objects, plus
316 other miscellaneous fixes.
318 (Version 2.19 was released with 5.8.9.)
320 Thu Nov 22 13:24:18 IST 2007 Abhijit Menon-Sen <ams@toroid.org>
324 Compile fixes for older Perls. (No functional changes.)
326 Sat Nov 17 02:12:12 IST 2007 Abhijit Menon-Sen <ams@toroid.org>
330 Various broken tests fixed. (No functional changes.)
332 Sat Mar 31 06:11:06 IST 2007 Abhijit Menon-Sen <ams@toroid.org>
336 1. Fixes to Storable::dclone, read_magic, retrieve_lscalar
337 2. Storable 0.1 compatibility
338 3. Miscellaneous compile/leak/test/portability fixes
340 Mon May 23 22:48:49 IST 2005 Abhijit Menon-Sen <ams@wiw.org>
344 Minor changes to address a couple of compile problems.
346 Mon Apr 25 07:29:14 IST 2005 Abhijit Menon-Sen <ams@wiw.org>
350 1. Store weak references
351 2. Add STORABLE_attach hook.
353 Thu Jun 17 12:26:43 BST 2004 Nicholas Clark <nick@ccl4.org>
357 1. Don't change the type of top level overloaded references to RV -
358 they are perfectly correct as PVMG
359 2. Storable needs to cope with incoming frozen data that happens to be
362 Wed Mar 17 15:40:29 GMT 2004 Nicholas Clark <nick@ccl4.org>
366 1. Add regression tests for the auto-require of STORABLE_thaw
367 2. Add auto-require of modules to restore overloading (and tests)
368 3. Change to no context (should give speedup with ithreads)
370 Sat Mar 13 20:11:03 GMT 2004 Nicholas Clark <nick@ccl4.org>
374 1. Storing restricted hashes in canonical order would SEGV. Fixed.
375 2. It was impossible to retrieve references to PL_sv_no and
376 PL_sv_undef from STORABLE_thaw hooks.
377 3. restrict.t was failing on 5.8.0, due to 5.8.0's unique
378 implementation of restricted hashes using PL_sv_undef
379 4. These changes allow a space optimisation for restricted hashes.
381 Sat Jan 24 16:22:32 IST 2004 Abhijit Menon-Sen <ams@wiw.org>
385 1. Thread safety: Storable::CLONE/init_perlinterp() now create
386 a new Perl context for each new ithread.
387 (From Stas Bekman and Jan Dubois.)
388 2. Fix a tag count mismatch with $Storable::Deparse that caused
389 all back-references after a stored sub to be off-by-N (where
390 N was the number of code references in between).
392 3. Prevent CODE references from turning into SCALAR references.
395 Sat Jan 3 18:49:18 GMT 2004 Nicholas Clark <nick@ccl4.org>
399 Fix minor problems with the CPAN release
400 1: Make Storable.xs work on 5.8.2 and later (already in the core)
401 2: Ship the linux hints file
402 3: Ship Test::More for the benefit of Perls pre 5.6.2
403 4: Correct Makefile.PL to only install in core for 5.8.0 and later
405 Sat Sep 6 01:08:20 IST 2003 Abhijit Menon-Sen <ams@wiw.org>
409 This release works around a 5.8.0 bug which caused hashes to not
410 be marked as having key flags even though an HEK had HEK_WASUTF8
411 set. (Note that the only reasonable solution is to silently drop
412 the flag from the affected key.)
414 Users of RT 3 who were seeing assertion failures should upgrade.
415 (Perl 5.8.1 will have the bug fixed.)
417 Mon May 5 10:24:16 IST 2003 Abhijit Menon-Sen <ams@wiw.org>
421 Minor bugfixes (self-tied objects are now correctly stored, as
422 are the results of additions larger than INT_MAX).
424 Mon Oct 7 21:56:38 BST 2002 Nicholas Clark <nick@ccl4.org>
428 Remove qr// from t/downgrade.t so that it will run on 5.004
429 Mention $File::Spec::VERSION a second time in t/forgive.t so that it
430 runs without warnings in 5.004 (this may be a 5.00405 bug I'm working
432 Fix t/integer.t initialisation to actually generate 64 bits of 9c
433 Fix comparison tests to use eval to get around 64 bit IV conversion
434 issues on 5.6.x, following my t/integer.t ^ precedence bug found by
436 Alter t/malice.t to work with Test/More.pm in t/, and skip individual
437 subtests that use $Config{ptrsize}, so that the rest of the test can
438 now be run with 5.004
439 Change t/malice.t and the error message in check_magic in Storable.xs
440 from "Pointer integer size" to "Pointer size"
441 Remove prerequisite of Test::More from Makefile.PL
442 Ship Test::Builder, Test::Simple and Test::More in t
444 Thu Oct 3 08:57:22 IST 2002 Abhijit Menon-Sen <ams@wiw.org>
448 Adds support for CODE references from Slaven Rezic
449 <slaven.rezic@berlin.de>.
451 Fri Jun 7 23:55:41 BST 2002 Nicholas Clark
455 Bug fix from Radu Greab <radu@netsoft.ro> (plus regression test)
456 to fix a recently introduced bug detected by Dave Rolsky.
457 Bug was that for a non threaded build, the class information was
458 being lost at freeze time on the first object with a STORABLE_freeze
459 hook. Consequentially the object was not blessed at all when thawed.
460 (The presence (or lack) of STORABLE_thaw was irrelevant; this was
461 a store-time data lost bug, caused by failure to initialize internal
463 The bug was introduced as development perl change 16442 (on
464 2002/05/07), so has been present since 2.00.
465 Patches to introduce more regression tests to reduce the chance of
466 a reoccurrence of this sort of goof are always welcome.
468 Thu May 30 20:31:08 BST 2002 Nicholas Clark <nick@ccl4.org>
470 Version 2.03 Header changes on 5.6.x on Unix where IV is long long
472 5.6.x introduced the ability to have IVs as long long. However,
473 Configure still defined BYTEORDER based on the size of a long.
474 Storable uses the BYTEORDER value as part of the header, but
475 doesn't explicitly store sizeof(IV) anywhere in the header.
476 Hence on 5.6.x built with IV as long long on a platform that
477 uses Configure (ie most things except VMS and Windows) headers
478 are identical for the different IV sizes, despite the files
479 containing some fields based on sizeof(IV)
481 5.8.0 is consistent; all platforms have BYTEORDER in config.h
482 based on sizeof(IV) rather than sizeof(long). This means that
483 the value of BYTEORDER will change from (say) 4321 to 87654321
484 between 5.6.1 and 5.8.0 built with the same options to Configure
485 on the same machine. This means that the Storable header will
486 differ, and the two versions will wrongly thing that they are
489 For the benefit of long term consistency, Storable now
490 implements the 5.8.0 BYTEORDER policy on 5.6.x. This means that
491 2.03 onwards default to be incompatible with 2.02 and earlier
492 (ie the large 1.0.x installed base) on the same 5.6.x perl.
494 To allow interworking, a new variable
495 $Storable::interwork_56_64bit is introduced. It defaults to
496 false. Set it to true to read and write old format files. Don't
497 use it unless you have existing stored data written with 5.6.x
498 that you couldn't otherwise read, or you need to interwork with
499 a machine running older Storable on a 5.6.x with long long IVs
500 (i.e., you probably don't need to use it).
502 Sat May 25 22:38:39 BST 2002 Nicholas Clark <nick@ccl4.org>
506 Rewrite Storable.xs so that the file header structure for write_magic
507 is built at compile time, and check_magic attempts to the header in
508 blocks rather than byte per byte. These changes make the compiled
509 extension 2.25% smaller, but are not significant enough to give a
512 Thu May 23 22:50:41 BST 2002 Nicholas Clark <nick@ccl4.org>
516 - New regression tests integer.t
517 - Add code to safely store large unsigned integers.
518 - Change code not to attempt to store large integers (ie > 32 bits)
519 in network order as 32 bits.
521 *Never* underestimate the value of a pathological test suite carefully
522 crafted with maximum malice before writing a line of real code. It
523 prevents crafty bugs from stowing away in your released code.
524 It's much less embarrassing to find them before you ship.
525 (Well, never underestimate it if you ever want to work for me)
527 Fri May 17 22:48:59 BST 2002 Nicholas Clark <nick@ccl4.org>
529 Version 2.0, binary format 2.5 (but writes format 2.4 on pre 5.7.3)
531 The perl5 porters have decided to make sure that Storable still
532 builds on pre-5.8 perls, and make the 5.8 version available on CPAN.
533 The VERSION is now 2.0, and it passes all tests on 5.005_03, 5.6.1
534 and 5.6.1 with threads. On 5.6.0 t/downgrade.t fails tests 34 and 37,
535 due to a bug in 5.6.0 - upgrade to 5.6.1.
537 Jarkko and I have collated the list of changes the perl5 porters have
538 from the perl5 Changes file:
540 - data features of upcoming perl 5.8.0 are supported: Unicode hash
541 keys (Unicode hash values have been supported since Storable 1.0.1)
542 and "restricted hashes" (readonly hashes and hash entries)
543 - a newer version of perl can now be used to serialize data which is
544 not supported in earlier perls: Storable will attempt to do the
545 right thing for as long as possible, croaking only when safe data
546 conversion simply isn't possible. Alternatively earlier perls can
547 opt to have a lossy downgrade data instead of croaking
548 - when built with perls pre 5.7.3 this Storable writes out files
549 with binary format 2.4, the same format as Storable 1.0.8 onwards.
550 This should mean that this Storable will inter-operate seamlessly
551 with any Storable 1.0.8 or newer on perls pre 5.7.3
552 - dclone() now works with empty string scalar objects
553 - retrieving of large hashes is now more efficient
554 - more routines autosplit out of the main module, so Storable should
555 load slightly more quickly
556 - better documentation
557 - the internal context objects are now freed explicitly, rather than
558 relying on thread or process exit
559 - bugs fixed in debugging trace code affecting builds made with 64 bit
561 - code tidy-ups to allow clean compiles with more warning options
562 turned on avoid problems with $@ getting corrupted on 5.005_03 if
563 Carp wasn't already loaded
564 - added &show_file_magic, so you can add to /etc/magic and teach
565 Unix's file command about Storable files
567 We plan to keep Storable on CPAN in sync with the Perl core, so
568 if you encounter bugs or other problems building or using Storable,
569 please let us know at perl5-porters@perl.org
572 Sat Dec 1 14:37:54 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
574 This is the LAST maintenance release of the Storable module.
575 Indeed, Storable is now part of perl 5.8, and will be maintained
576 as part of Perl. The CPAN module will remain available there
577 for people running pre-5.8 perls.
579 Avoid requiring Fcntl upfront, useful to embedded runtimes.
580 Use an eval {} for testing, instead of making Storable.pm
581 simply fail its compilation in the BEGIN block.
583 store_fd() will now correctly autoflush file if needed.
585 Tue Aug 28 23:53:20 MEST 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
587 Fixed truncation race with lock_retrieve() in lock_store().
588 The file has to be truncated only once the exclusive lock is held.
590 Removed spurious debugging messages in .xs file.
592 Sun Jul 1 13:27:32 MEST 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
594 Systematically use "=over 4" for POD linters.
595 Apparently, POD linters are much stricter than would
596 otherwise be needed, but that's OK.
598 Fixed memory corruption on croaks during thaw(). Thanks
599 to Claudio Garcia for reproducing this bug and providing the
600 code to exercise it. Added test cases for this bug, adapted
603 Made code compile cleanly with -Wall (from Jarkko Hietaniemi).
605 Changed tagnum and classnum from I32 to IV in context. Also
608 Thu Mar 15 01:22:32 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
610 Last version was wrongly compiling with assertions on, due
611 to an edit glitch. That did not cause any problem (apart from
612 a slight performance loss) excepted on Win* platforms, where the
613 assertion code does not compile.
615 Sat Feb 17 13:37:37 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
619 Forgot to increase version number at previous patch (there were
620 two of them, which is why we jump from 1.0.8 to 1.0.10).
622 Sat Feb 17 13:35:00 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
624 Version 1.0.8, binary format 2.4.
626 Fixed incorrect error message.
628 Now bless objects ASAP at retrieve time, which is meant to fix
631 * Indirect references to overloaded object were not able to
632 restore overloading if the object was not blessed yet,
633 which was possible since blessing occurred only after the
636 * Storable hooks asking for serialization of blessed ref could
637 get un-blessed refs at retrieval time, for the very same
640 The fix implemented here was suggested by Nick Ing-Simmons.
642 Added support for blessed ref to tied structures. This is the
643 cause for the binary format change.
645 Added EBCDIC version of the compatibility test with 0.6.11,
648 Added tests for the new features, and to make sure the bugs they
649 are meant to fix are indeed fixed.
651 Wed Jan 3 10:43:18 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
653 Removed spurious 'clean' entry in Makefile.PL.
655 Added CAN_FLOCK to determine whether we can flock() or not,
656 by inspecting Perl's configuration parameters, as determined
659 Trace offending package when overloading cannot be restored
662 Made context cleanup safer to avoid dup freeing, mostly in the
663 presence of repeated exceptions during store/retrieve (which can
664 cause memory leaks anyway, so it's just additional safety, not a
667 Sun Nov 5 18:23:48 MET 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
671 Fixed severe "object lost" bug for STORABLE_freeze returns,
672 when refs to lexicals, taken within the hook, were to be
673 serialized by Storable. Enhanced the t/recurse.t test to
674 stress hook a little more with refs to lexicals.
676 Thu Oct 26 19:14:38 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
680 Documented that store() and retrieve() can return undef.
681 That is, the error reporting is not always made via exceptions,
682 as the paragraph on error reporting was implying.
684 Auto requires module of blessed ref when STORABLE_thaw misses.
685 When the Storable engine looks for the STORABLE_thaw hook and
686 does not find it, it now tries to require the package into which
687 the blessed reference is.
689 Just check $^O, in t/lock.t: there's no need to pull the whole
690 Config module for that.
692 Mon Oct 23 20:03:49 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
696 Protected calls to flock() for DOS platform: apparently, the
697 flock/fcnlt emulation is reported to be broken on that
700 Added logcarp emulation if they don't have Log::Agent, since
701 we now use it to carp when lock_store/lock_retrieve is used
704 Fri Sep 29 21:52:29 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
708 Avoid using "tainted" and "dirty" since Perl remaps them via
709 cpp (i.e. #define). This is deeply harmful when threading
710 is enabled. This concerned both the context structure and
711 local variable and argument names. Brrr..., scary!
713 Thu Sep 28 23:46:39 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
717 Fixed spelling in README.
719 Added lock_store, lock_nstore, and lock_retrieve (advisory locking)
720 after a proposal from Erik Haugan <erik@solbors.no>.
722 Perls before 5.004_04 lack newSVpvn, added remapping in XS.
724 Fixed stupid typo in the t/utf8.t test.
726 Sun Sep 17 18:51:10 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
728 Version 1.0.1, binary format 2.3.
730 Documented that doubles are stored stringified by nstore().
732 Added Salvador Ortiz Garcia in CREDITS section, He identified
733 a bug in the store hooks and proposed the right fix: the class
734 id was allocated too soon. His bug case was also added to
735 the regression test suite.
737 Now only taint retrieved data when source was tainted. A bug
738 discovered by Marc Lehmann.
740 Added support for UTF-8 strings, a contribution of Marc Lehmann.
741 This is normally only activated in post-5.6 perls.
743 Thu Aug 31 23:06:06 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
745 First official release Storable 1.0, for inclusion in perl 5.7.0.
746 The license scheme is now compatible with Perl's.
748 Thu Aug 24 01:02:02 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
750 ANSI-fied most of the code, preparing for Perl core integration.
751 The next version of Storable will be 0.8, and will be integrated
752 into the Perl core (development branch).
754 Dispatch tables were moved upfront to relieve some compilers,
755 especially on AIX and Windows platforms.
757 Merged 64-bit fixes from perl5-porters.
759 Mon Aug 14 09:22:04 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
761 Added a refcnt dec in retrieve_tied_key(): sv_magic() increases
762 the refcnt on the mg_ptr as well.
764 Removed spurious dependency to Devel::Peek, which was used for
765 testing only in t/tied_items.t. Thanks to Conrad Heiney
766 <conrad@fringehead.org> for spotting it first.
768 Sun Aug 13 22:12:59 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
770 Marc Lehmann kindly contributed code to add overloading support
771 and to handle references to tied variables.
773 Rewrote leading blurb about compatibility to make it clearer what
774 "backward compatibility" is about: when I say 0.7 is backward
775 compatible with 0.6, it means the revision 0.7 can read files
778 Mention new Clone(3) extension in SEE ALSO.
780 Was wrongly optimizing for "undef" values in hashes by not
781 fully recursing: as a result, tied "undef" values were incorrectly
784 Sun Jul 30 12:59:17 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
786 First revision of Storable 0.7.
788 The serializing format is new, known as version 2.0. It is fully
789 backward compatible with 0.6. Earlier formats are deprecated and
790 have not even been tested: next version will drop pre-0.6 format.
792 Changes since 0.6@11:
794 - Moved interface to the "beta" status. Some tiny parts are still
795 subject to change, but nothing important enough to warrant an "alpha"
798 - Slightly reduced the size of the Storable image by factorizing
799 object class names and removing final object storage notification due
800 to a redesign of the blessed object storing.
802 - Classes can now redefine how they wish their instances to be serialized
803 and/or deep cloned. Serializing hooks are written in Perl code.
805 - The engine is now fully re-entrant.
807 Sun Apr 2 23:47:50 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
809 Added provision to detect more recent binary formats, since
810 the new upcoming Storable-0.7 will use a different format.
811 In order to prevent attempting the de-serialization of newer
812 formats by older versions, I'm adding this now to the 0.6 series.
814 I'm expecting this revision to be the last of the 0.6 series.
815 Unless it does not work with perl 5.6, which I don't use yet,
816 and therefore against which I cannot test.
818 Wed Mar 29 19:55:21 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
820 Added note about format incompatibilities with old versions
821 (i.e. pre 0.5@9 formats, which cannot be understood as there
822 was no versionning information in the file by then).
824 Protect all $@ variables when eval {} used, to avoid corrupting
825 it when store/retrieve is called within an exception handler.
827 Mistakenly included "patchlevel.h" instead of <patchlevel.h>,
828 preventing Perl's patchlevel from being included, which is
829 needed starting from 5.6.
831 Tue May 12 09:15:15 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
833 Fixed shared "undef" bug in hashes, which did not remain shared
834 through store/retrieve.
836 Thu Feb 10 19:48:16 MET 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
838 added last_op_in_netorder() predicate
839 documented last_op_in_netorder()
840 added tests for the new last_op_in_netorder() predicate
842 Wed Oct 20 19:07:36 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
844 Forgot to update VERSION
846 Tue Oct 19 21:25:02 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
848 Added mention of japanese translation for the manual page.
850 Fixed typo in macro that made threaded code not compilable,
851 especially on Win32 platforms.
853 Changed detection of older perls (pre-5.005) by testing PATCHLEVEL
854 directly instead of relying on internal symbols.
856 Tue Sep 14 22:13:28 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
858 Integrated "thread-safe" patch from Murray Nesbitt.
859 Note that this may not be very efficient for threaded code,
860 see comment in the code.
862 Try to avoid compilation warning on 64-bit CPUs. Can't test it,
863 since I don't have access to such machines.
865 Mon Jul 12 14:37:19 METDST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
867 changed my e-mail to pobox.
869 mentioned it is not thread-safe.
871 updated version number.
873 uses new internal PL_* naming convention.
875 Fri Jul 3 13:38:16 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
877 Updated benchmark figures due to recent optimizations done in
878 store(): tagnums are now stored as-is in the hash table, so
879 no surrounding SV is created. And the "shared keys" mode for
880 hash table was turned off.
882 Fixed backward compatibility (wrt 0.5@9) for retrieval of
883 blessed refs. That old version did something wrong, but the
884 bugfix prevented correct retrieval of the old format.
886 Mon Jun 22 11:00:48 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
888 Changed benchmark figures.
890 Adjust refcnt of tied objects after calling sv_magic() to avoid
891 memory leaks. Contributed by Jeff Gresham.
893 Fri Jun 12 11:50:04 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
895 Added workaround for persistent LVALUE-ness in perl5.004. All
896 scalars tagged as being an lvalue are handled as if they were
897 not an lvalue at all. Added test for that LVALUE bug workaround.
899 Now handles Perl immortal scalars explicitly, by storing &sv_yes
902 Retrieval of non-immortal undef cannot be shared. Previous
903 version was over-optimizing by not creating a separate SV for
904 all undefined scalars seen.
906 Thu Jun 4 17:21:51 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
908 Baseline for Storable-0.6@0.
910 This version introduces a binary incompatibility in the generated
911 binary image, which is more compact than older ones by approximatively
912 15%, depending on the exact degree of sharing in your structures.
914 The good news is that your older images can still be retrieved with
915 this version, i.e. backward compatibility is preserved. This version
916 of Storable can only generate new binaries however.
918 Another good news is that the retrieval of data structure is
919 significantly quicker than before, because a Perl array is used
920 instead of a hash table to keep track of retrieved objects, and
921 also because the image being smaller, less I/O function calls are
924 Tue May 12 09:15:15 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
926 Version number now got from Storable.pm directly.
928 Fixed overzealous sv_type() optimization, which would make
929 Storable fail when faced with an "upgraded" SV to the PVIV
930 or PVNV kind containing a reference.
932 Thu Apr 30 15:11:30 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
934 Extended the SYNOPSIS section to give quick overview of the
935 routines and their signature.
937 Optimized sv_type() to avoid flags checking when not needed, i.e.
938 when their type makes it impossible for them to be refs or tied.
939 This slightly increases throughput by a few percents when refs
940 and tied variables are marginal occurrences in your data.
942 Stubs for XS now use OutputStream and InputStream file types to
943 make it work when the given file is actually a socket. Perl
944 makes a distinction for sockets in its internal I/O structures
945 by having both a read and a write structure, whereas plain files
948 Tue Jun 3 09:41:33 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
950 Thanks to a contribution from Benjamin A. Holzman, Storable is now
951 able to correctly serialize tied SVs, i.e. tied arrays, hashes
954 Thu Apr 9 18:07:51 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
956 I said SvPOK() had changed to SvPOKp(), but that was a lie...
958 Wed Apr 8 13:14:29 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
960 Wrote sizeof(SV *) instead of sizeof(I32) when portable, which
961 in effect mangled the object tags and prevented portability
962 across 32/64 bit architectures!
964 Wed Mar 25 14:57:02 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
966 Added code example for store_fd() and retrieve_fd() in the
967 man page, to emphasize that file descriptors must be passed as
968 globs, not as plain strings.
970 Cannot use SV addresses as tag when using nstore() on LP64. This
971 was the cause of problems when creating a storable image on an
972 LP64 machine and retrieving it on an ILP32 system, which is
973 exactly what nstore() is meant for...
975 However, we continue to use SV addresses as tags for plain store(),
976 because benchmarking shows that it saves up to 8% of the store
977 time, and store() is meant to be fast at the expense of lack
980 This means there will be approximately an 8% degradation of
981 performance for nstore(), but it's now working as expected.
982 That cost may vary on your machine of course, since it is
983 solely caused by the memory allocation overhead used to create
984 unique SV tags for each distinct stored SV.
986 Tue Jan 20 09:21:53 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
988 Don't use any '_' in version number.
990 Tue Jan 13 17:51:50 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
992 Updated version number.
994 added binmode() calls for systems where it matters.
996 Be sure to pass globs, not plain file strings, to C routines,
997 so that Storable can be used under the Perl debugger.
999 Wed Nov 5 10:53:22 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
1001 Fix memory leaks on seen hash table and returned SV refs.
1003 Storable did not work properly when tainting enabled.
1005 Fixed "Allocation too large" messages in freeze/thaw and added.
1006 proper regression test in t/freeze.t.
1008 Tue Jun 3 09:41:33 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
1010 Updated version number
1012 Added freeze/thaw interface and dclone.
1014 Fri May 16 10:45:47 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
1016 Forgot that AutoLoader does not export its own AUTOLOAD.
1019 use AutoLoader 'AUTOLOAD';
1021 but that would not be backward compatible. So the export is
1024 Tue Mar 25 11:21:32 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
1026 Empty scalar strings are now "defined" at retrieval time.
1028 New test to ensure an empty string is defined when retrieved.
1030 Thu Feb 27 16:32:44 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
1032 Updated version number
1034 Declare VERSION as being used
1036 Fixed a typo in the PerlIO_putc remapping.
1037 PerlIO_read and perlIO_write inverted size/nb_items.
1038 (only relevant for pre-perl5.004 versions)
1040 Thu Feb 27 15:58:31 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
1042 Updated version number
1044 Added VERSION identification
1046 Allow build with perl5.003, which is ante perlIO time
1048 Mon Jan 13 17:53:18 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
1052 Wed Jan 22 15:19:56 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
1054 Updated version number in Makefile.PL.
1056 Added "thanks to" section to README.
1058 Documented new forgive_me variable.
1062 Added forgive_me support to allow store() of data structures
1063 containing non-storable items like CODE refs.