1 Sat Dec 1 14:37:54 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
5 This is the LAST maintenance release of the Storable module.
6 Indeed, Storable is now part of perl 5.8, and will be maintained
7 as part of Perl. The CPAN module will remain available there
8 for people running pre-5.8 perls.
10 Avoid requiring Fcntl upfront, useful to embedded runtimes.
11 Use an eval {} for testing, instead of making Storable.pm
12 simply fail its compilation in the BEGIN block.
14 store_fd() will now correctly autoflush file if needed.
16 Tue Aug 28 23:53:20 MEST 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
20 Fixed truncation race with lock_retrieve() in lock_store().
21 The file has to be truncated only once the exclusive lock is held.
23 Removed spurious debugging messages in .xs file.
25 Sun Jul 1 13:27:32 MEST 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
29 Systematically use "=over 4" for POD linters.
30 Apparently, POD linters are much stricter than would
31 otherwise be needed, but that's OK.
33 Fixed memory corruption on croaks during thaw(). Thanks
34 to Claudio Garcia for reproducing this bug and providing the
35 code to exercise it. Added test cases for this bug, adapted
38 Made code compile cleanly with -Wall (from Jarkko Hietaniemi).
40 Changed tagnum and classnum from I32 to IV in context. Also
43 Thu Mar 15 01:22:32 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
47 Last version was wrongly compiling with assertions on, due
48 to an edit glitch. That did not cause any problem (apart from
49 a slight performance loss) excepted on Win* platforms, where the
50 assertion code does not compile.
52 Sat Feb 17 13:37:37 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
58 Forgot to increase version number at previous patch (there were
59 two of them, which is why we jump from 1.0.8 to 1.0.10).
61 Sat Feb 17 13:35:00 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
65 Version 1.0.8, binary format 2.4.
67 Fixed incorrect error message.
69 Now bless objects ASAP at retrieve time, which is meant to fix
72 * Indirect references to overloaded object were not able to
73 restore overloading if the object was not blessed yet,
74 which was possible since blessing occurred only after the
77 * Storable hooks asking for serialization of blessed ref could
78 get un-blessed refs at retrieval time, for the very same
81 The fix implemented here was suggested by Nick Ing-Simmons.
83 Added support for blessed ref to tied structures. This is the
84 cause for the binary format change.
86 Added EBCDIC version of the compatibility test with 0.6.11,
89 Added tests for the new features, and to make sure the bugs they
90 are meant to fix are indeed fixed.
92 Wed Jan 3 10:43:18 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
96 Removed spurious 'clean' entry in Makefile.PL.
98 Added CAN_FLOCK to determine whether we can flock() or not,
99 by inspecting Perl's configuration parameters, as determined
102 Trace offending package when overloading cannot be restored
105 Made context cleanup safer to avoid dup freeing, mostly in the
106 presence of repeated exceptions during store/retrieve (which can
107 cause memory leaks anyway, so it's just additional safety, not a
110 Sun Nov 5 18:23:48 MET 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
116 Fixed severe "object lost" bug for STORABLE_freeze returns,
117 when refs to lexicals, taken within the hook, were to be
118 serialized by Storable. Enhanced the t/recurse.t test to
119 stress hook a little more with refs to lexicals.
121 Thu Oct 26 19:14:38 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
127 Documented that store() and retrieve() can return undef.
128 That is, the error reporting is not always made via exceptions,
129 as the paragraph on error reporting was implying.
131 Auto requires module of blessed ref when STORABLE_thaw misses.
132 When the Storable engine looks for the STORABLE_thaw hook and
133 does not find it, it now tries to require the package into which
134 the blessed reference is.
136 Just check $^O, in t/lock.t: there's no need to pull the whole
137 Config module for that.
139 Mon Oct 23 20:03:49 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
145 Protected calls to flock() for DOS platform: apparently, the
146 flock/fcnlt emulation is reported to be broken on that
149 Added logcarp emulation if they don't have Log::Agent, since
150 we now use it to carp when lock_store/lock_retrieve is used
153 Fri Sep 29 21:52:29 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
159 Avoid using "tainted" and "dirty" since Perl remaps them via
160 cpp (i.e. #define). This is deeply harmful when threading
161 is enabled. This concerned both the context structure and
162 local variable and argument names. Brrr..., scary!
164 Thu Sep 28 23:46:39 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
170 Fixed spelling in README.
172 Added lock_store, lock_nstore, and lock_retrieve (advisory locking)
173 after a proposal from Erik Haugan <erik@solbors.no>.
175 Perls before 5.004_04 lack newSVpvn, added remapping in XS.
177 Fixed stupid typo in the t/utf8.t test.
179 Sun Sep 17 18:51:10 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
183 Version 1.0.1, binary format 2.3.
185 Documented that doubles are stored stringified by nstore().
187 Added Salvador Ortiz Garcia in CREDITS section, He identified
188 a bug in the store hooks and proposed the right fix: the class
189 id was allocated too soon. His bug case was also added to
190 the regression test suite.
192 Now only taint retrieved data when source was tainted. A bug
193 discovered by Marc Lehmann.
195 Added support for UTF-8 strings, a contribution of Marc Lehmann.
196 This is normally only activated in post-5.6 perls.
198 Thu Aug 31 23:06:06 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
200 First official release Storable 1.0, for inclusion in perl 5.7.0.
201 The license scheme is now compatible with Perl's.
203 Thu Aug 24 01:02:02 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
208 ANSI-fied most of the code, preparing for Perl core integration.
209 The next version of Storable will be 0.8, and will be integrated
210 into the Perl core (development branch).
212 Dispatch tables were moved upfront to relieve some compilers,
213 especially on AIX and Windows platforms.
215 Merged 64-bit fixes from perl5-porters.
217 Mon Aug 14 09:22:04 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
221 Added a refcnt dec in retrieve_tied_key(): sv_magic() increases
222 the refcnt on the mg_ptr as well.
224 Removed spurious dependency to Devel::Peek, which was used for
225 testing only in t/tied_items.t. Thanks to Conrad Heiney
226 <conrad@fringehead.org> for spotting it first.
228 Sun Aug 13 22:12:59 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
232 Marc Lehmann kindly contributed code to add overloading support
233 and to handle references to tied variables.
235 Rewrote leading blurb about compatibility to make it clearer what
236 "backward compatibility" is about: when I say 0.7 is backward
237 compatible with 0.6, it means the revision 0.7 can read files
240 Mention new Clone(3) extension in SEE ALSO.
242 Was wrongly optimizing for "undef" values in hashes by not
243 fully recursing: as a result, tied "undef" values were incorrectly
246 Sun Jul 30 12:59:17 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
248 First revision of Storable 0.7.
250 The serializing format is new, known as version 2.0. It is fully
251 backward compatible with 0.6. Earlier formats are deprecated and
252 have not even been tested: next version will drop pre-0.6 format.
254 Changes since 0.6@11:
256 - Moved interface to the "beta" status. Some tiny parts are still
257 subject to change, but nothing important enough to warrant an "alpha"
260 - Slightly reduced the size of the Storable image by factorizing
261 object class names and removing final object storage notification due
262 to a redesign of the blessed object storing.
264 - Classes can now redefine how they wish their instances to be serialized
265 and/or deep cloned. Serializing hooks are written in Perl code.
267 - The engine is now fully re-entrant.
269 Sun Apr 2 23:47:50 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
273 Added provision to detect more recent binary formats, since
274 the new upcoming Storable-0.7 will use a different format.
275 In order to prevent attempting the de-serialization of newer
276 formats by older versions, I'm adding this now to the 0.6 series.
278 I'm expecting this revision to be the last of the 0.6 series.
279 Unless it does not work with perl 5.6, which I don't use yet,
280 and therefore against which I cannot test.
282 Wed Mar 29 19:55:21 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
286 Added note about format incompatibilities with old versions
287 (i.e. pre 0.5@9 formats, which cannot be understood as there
288 was no versionning information in the file by then).
290 Protect all $@ variables when eval {} used, to avoid corrupting
291 it when store/retrieve is called within an exception handler.
293 Mistakenly included "patchlevel.h" instead of <patchlevel.h>,
294 preventing Perl's patchlevel from being included, which is
295 needed starting from 5.6.
297 Tue May 12 09:15:15 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
301 Fixed shared "undef" bug in hashes, which did not remain shared
302 through store/retrieve.
304 Thu Feb 10 19:48:16 MET 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
308 added last_op_in_netorder() predicate
309 documented last_op_in_netorder()
310 added tests for the new last_op_in_netorder() predicate
312 Wed Oct 20 19:07:36 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
316 Forgot to update VERSION
318 Tue Oct 19 21:25:02 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
322 Added mention of japanese translation for the manual page.
324 Fixed typo in macro that made threaded code not compilable,
325 especially on Win32 platforms.
327 Changed detection of older perls (pre-5.005) by testing PATCHLEVEL
328 directly instead of relying on internal symbols.
330 Tue Sep 14 22:13:28 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
334 Integrated "thread-safe" patch from Murray Nesbitt.
335 Note that this may not be very efficient for threaded code,
336 see comment in the code.
338 Try to avoid compilation warning on 64-bit CPUs. Can't test it,
339 since I don't have access to such machines.
341 Mon Jul 12 14:37:19 METDST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
345 changed my e-mail to pobox.
347 mentionned it is not thread-safe.
349 updated version number.
351 uses new internal PL_* naming convention.
353 Fri Jul 3 13:38:16 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
357 Updated benchmark figures due to recent optimizations done in
358 store(): tagnums are now stored as-is in the hash table, so
359 no surrounding SV is created. And the "shared keys" mode for
360 hash table was turned off.
362 Fixed backward compatibility (wrt 0.5@9) for retrieval of
363 blessed refs. That old version did something wrong, but the
364 bugfix prevented correct retrieval of the old format.
366 Mon Jun 22 11:00:48 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
370 Changed benchmark figures.
372 Adjust refcnt of tied objects after calling sv_magic() to avoid
373 memory leaks. Contributed by Jeff Gresham.
375 Fri Jun 12 11:50:04 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
379 Added workaround for persistent LVALUE-ness in perl5.004. All
380 scalars tagged as being an lvalue are handled as if they were
381 not an lvalue at all. Added test for that LVALUE bug workaround.
383 Now handles Perl immortal scalars explicitely, by storing &sv_yes
384 as such, explicitely.
386 Retrieval of non-immortal undef cannot be shared. Previous
387 version was over-optimizing by not creating a separate SV for
388 all undefined scalars seen.
390 Thu Jun 4 17:21:51 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
394 Baseline for Storable-0.6@0.
396 This version introduces a binary incompatibility in the generated
397 binary image, which is more compact than older ones by approximatively
398 15%, depending on the exact degree of sharing in your structures.
400 The good news is that your older images can still be retrieved with
401 this version, i.e. backward compatibility is preserved. This version
402 of Storable can only generate new binaries however.
404 Another good news is that the retrieval of data structure is
405 significantly quicker than before, because a Perl array is used
406 instead of a hash table to keep track of retrieved objects, and
407 also because the image being smaller, less I/O function calls are
410 Tue May 12 09:15:15 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
414 Version number now got from Storable.pm directly.
416 Fixed overzealous sv_type() optimization, which would make
417 Storable fail when faced with an "upgraded" SV to the PVIV
418 or PVNV kind containing a reference.
420 Thu Apr 30 15:11:30 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
424 Extended the SYNOPSIS section to give quick overview of the
425 routines and their signature.
427 Optimized sv_type() to avoid flags checking when not needed, i.e.
428 when their type makes it impossible for them to be refs or tied.
429 This slightly increases throughput by a few percents when refs
430 and tied variables are marginal occurrences in your data.
432 Stubs for XS now use OutputStream and InputStream file types to
433 make it work when the given file is actually a socket. Perl
434 makes a distinction for sockets in its internal I/O structures
435 by having both a read and a write structure, whereas plain files
438 Tue Jun 3 09:41:33 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
442 Thanks to a contribution from Benjamin A. Holzman, Storable is now
443 able to correctly serialize tied SVs, i.e. tied arrays, hashes
446 Thu Apr 9 18:07:51 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
450 I said SvPOK() had changed to SvPOKp(), but that was a lie...
452 Wed Apr 8 13:14:29 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
456 Wrote sizeof(SV *) instead of sizeof(I32) when portable, which
457 in effect mangled the object tags and prevented portability
458 accross 32/64 bit architectures!
460 Wed Mar 25 14:57:02 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
464 Added code example for store_fd() and retrieve_fd() in the
465 man page, to emphasize that file descriptors must be passed as
466 globs, not as plain strings.
468 Cannot use SV addresses as tag when using nstore() on LP64. This
469 was the cause of problems when creating a storable image on an
470 LP64 machine and retrieving it on an ILP32 system, which is
471 exactly what nstore() is meant for...
473 However, we continue to use SV addresses as tags for plain store(),
474 because benchamarking shows that it saves up to 8% of the store
475 time, and store() is meant to be fast at the expense of lack
478 This means there will be approximately an 8% degradation of
479 performance for nstore(), but it's now working as expected.
480 That cost may vary on your machine of course, since it is
481 solely caused by the memory allocation overhead used to create
482 unique SV tags for each distinct stored SV.
484 Tue Jan 20 09:21:53 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
488 Don't use any '_' in version number.
490 Tue Jan 13 17:51:50 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
494 Updated version number.
496 added binmode() calls for systems where it matters.
498 Be sure to pass globs, not plain file strings, to C routines,
499 so that Storable can be used under the Perl debugger.
501 Wed Nov 5 10:53:22 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
505 Fix memory leaks on seen hash table and returned SV refs.
507 Storable did not work properly when tainting enabled.
509 Fixed "Allocation too large" messages in freeze/thaw and added.
510 proper regression test in t/freeze.t.
512 Tue Jun 3 09:41:33 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
516 Updated version number
518 Added freeze/thaw interface and dclone.
520 Fri May 16 10:45:47 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
524 Forgot that AutoLoader does not export its own AUTOLOAD.
527 use AutoLoader 'AUTOLOAD';
529 but that would not be backward compatible. So the export is
532 Tue Mar 25 11:21:32 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
536 Empty scalar strings are now "defined" at retrieval time.
538 New test to ensure an empty string is defined when retrieved.
540 Thu Feb 27 16:32:44 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
544 Updated version number
546 Declare VERSION as being used
548 Fixed a typo in the PerlIO_putc remapping.
549 PerlIO_read and perlIO_write inverted size/nb_items.
550 (only relevant for pre-perl5.004 versions)
552 Thu Feb 27 15:58:31 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
556 Updated version number
558 Added VERSION identification
560 Allow build with perl5.003, which is ante perlIO time
562 Mon Jan 13 17:53:18 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
568 Wed Jan 22 15:19:56 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
572 Updated version number in Makefile.PL.
574 Added "thanks to" section to README.
576 Documented new forgive_me variable.
580 Added forgive_me support to allow store() of data structures
581 containing non-storable items like CODE refs.