1 Thu May 23 22:50:41 BST 2002 Nicholas Clark <nick@ccl4.org>
7 New regression tests integer.t
8 Add code to safely store large unsigned integers.
9 Change code not to attempt to store large integers (ie > 32 bits)
10 in network order as 32 bits.
11 *Never* underestimate the value of a pathological test suite carefully
12 crafted with maximum malice before writing a line of real code. It
13 prevents crafty bugs from stowing away in your released code.
14 It's much less embarrassing to find them before you ship.
15 (Well, never underestimate it if you ever want to work for me)
17 Fri May 17 22:48:59 BST 2002 Nicholas Clark <nick@ccl4.org>
21 Version 2.0, binary format 2.5 (but writes format 2.4 on pre 5.7.3)
23 The perl5 porters have decided to make sure that Storable still
24 builds on pre-5.8 perls, and make the 5.8 version available on CPAN.
25 The VERSION is now 2.0, and it passes all tests on 5.005_03, 5.6.1
26 and 5.6.1 with threads. On 5.6.0 t/downgrade.t fails tests 34 and 37,
27 due to a bug in 5.6.0 - upgrade to 5.6.1.
29 Jarkko and I have collated the list of changes the perl5 porters have
30 from the perl5 Changes file:
32 - data features of upcoming perl 5.8.0 are supported: Unicode hash
33 keys (Unicode hash values have been supported since Storable 1.0.1)
34 and "restricted hashes" (readonly hashes and hash entries)
35 - a newer version of perl can now be used to serialize data which is
36 not supported in earlier perls: Storable will attempt to do the
37 right thing for as long as possible, croaking only when safe data
38 conversion simply isn't possible. Alternatively earlier perls can
39 opt to have a lossy downgrade data instead of croaking
40 - when built with perls pre 5.7.3 this Storable writes out files
41 with binary format 2.4, the same format as Storable 1.0.8 onwards.
42 This should mean that this Storable will inter-operate seamlessly
43 with any Storable 1.0.8 or newer on perls pre 5.7.3
44 - dclone() now works with empty string scalar objects
45 - retrieving of large hashes is now more efficient
46 - more routines autosplit out of the main module, so Storable should
47 load slightly more quickly
48 - better documentation
49 - the internal context objects are now freed explicitly, rather than
50 relying on thread or process exit
51 - bugs fixed in debugging trace code affecting builds made with 64 bit
53 - code tidy-ups to allow clean compiles with more warning options
54 turned on avoid problems with $@ getting corrupted on 5.005_03 if
55 Carp wasn't already loaded
56 - added &show_file_magic, so you can add to /etc/magic and teach
57 Unix's file command about Storable files
59 We plan to keep Storable on CPAN in sync with the Perl core, so
60 if you encounter bugs or other problems building or using Storable,
61 please let us know at perl5-porters@perl.org
64 Sat Dec 1 14:37:54 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
68 This is the LAST maintenance release of the Storable module.
69 Indeed, Storable is now part of perl 5.8, and will be maintained
70 as part of Perl. The CPAN module will remain available there
71 for people running pre-5.8 perls.
73 Avoid requiring Fcntl upfront, useful to embedded runtimes.
74 Use an eval {} for testing, instead of making Storable.pm
75 simply fail its compilation in the BEGIN block.
77 store_fd() will now correctly autoflush file if needed.
79 Tue Aug 28 23:53:20 MEST 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
83 Fixed truncation race with lock_retrieve() in lock_store().
84 The file has to be truncated only once the exclusive lock is held.
86 Removed spurious debugging messages in .xs file.
88 Sun Jul 1 13:27:32 MEST 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
92 Systematically use "=over 4" for POD linters.
93 Apparently, POD linters are much stricter than would
94 otherwise be needed, but that's OK.
96 Fixed memory corruption on croaks during thaw(). Thanks
97 to Claudio Garcia for reproducing this bug and providing the
98 code to exercise it. Added test cases for this bug, adapted
101 Made code compile cleanly with -Wall (from Jarkko Hietaniemi).
103 Changed tagnum and classnum from I32 to IV in context. Also
106 Thu Mar 15 01:22:32 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
110 Last version was wrongly compiling with assertions on, due
111 to an edit glitch. That did not cause any problem (apart from
112 a slight performance loss) excepted on Win* platforms, where the
113 assertion code does not compile.
115 Sat Feb 17 13:37:37 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
121 Forgot to increase version number at previous patch (there were
122 two of them, which is why we jump from 1.0.8 to 1.0.10).
124 Sat Feb 17 13:35:00 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
128 Version 1.0.8, binary format 2.4.
130 Fixed incorrect error message.
132 Now bless objects ASAP at retrieve time, which is meant to fix
135 * Indirect references to overloaded object were not able to
136 restore overloading if the object was not blessed yet,
137 which was possible since blessing occurred only after the
140 * Storable hooks asking for serialization of blessed ref could
141 get un-blessed refs at retrieval time, for the very same
144 The fix implemented here was suggested by Nick Ing-Simmons.
146 Added support for blessed ref to tied structures. This is the
147 cause for the binary format change.
149 Added EBCDIC version of the compatibility test with 0.6.11,
152 Added tests for the new features, and to make sure the bugs they
153 are meant to fix are indeed fixed.
155 Wed Jan 3 10:43:18 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
159 Removed spurious 'clean' entry in Makefile.PL.
161 Added CAN_FLOCK to determine whether we can flock() or not,
162 by inspecting Perl's configuration parameters, as determined
165 Trace offending package when overloading cannot be restored
168 Made context cleanup safer to avoid dup freeing, mostly in the
169 presence of repeated exceptions during store/retrieve (which can
170 cause memory leaks anyway, so it's just additional safety, not a
173 Sun Nov 5 18:23:48 MET 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
179 Fixed severe "object lost" bug for STORABLE_freeze returns,
180 when refs to lexicals, taken within the hook, were to be
181 serialized by Storable. Enhanced the t/recurse.t test to
182 stress hook a little more with refs to lexicals.
184 Thu Oct 26 19:14:38 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
190 Documented that store() and retrieve() can return undef.
191 That is, the error reporting is not always made via exceptions,
192 as the paragraph on error reporting was implying.
194 Auto requires module of blessed ref when STORABLE_thaw misses.
195 When the Storable engine looks for the STORABLE_thaw hook and
196 does not find it, it now tries to require the package into which
197 the blessed reference is.
199 Just check $^O, in t/lock.t: there's no need to pull the whole
200 Config module for that.
202 Mon Oct 23 20:03:49 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
208 Protected calls to flock() for DOS platform: apparently, the
209 flock/fcnlt emulation is reported to be broken on that
212 Added logcarp emulation if they don't have Log::Agent, since
213 we now use it to carp when lock_store/lock_retrieve is used
216 Fri Sep 29 21:52:29 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
222 Avoid using "tainted" and "dirty" since Perl remaps them via
223 cpp (i.e. #define). This is deeply harmful when threading
224 is enabled. This concerned both the context structure and
225 local variable and argument names. Brrr..., scary!
227 Thu Sep 28 23:46:39 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
233 Fixed spelling in README.
235 Added lock_store, lock_nstore, and lock_retrieve (advisory locking)
236 after a proposal from Erik Haugan <erik@solbors.no>.
238 Perls before 5.004_04 lack newSVpvn, added remapping in XS.
240 Fixed stupid typo in the t/utf8.t test.
242 Sun Sep 17 18:51:10 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
246 Version 1.0.1, binary format 2.3.
248 Documented that doubles are stored stringified by nstore().
250 Added Salvador Ortiz Garcia in CREDITS section, He identified
251 a bug in the store hooks and proposed the right fix: the class
252 id was allocated too soon. His bug case was also added to
253 the regression test suite.
255 Now only taint retrieved data when source was tainted. A bug
256 discovered by Marc Lehmann.
258 Added support for UTF-8 strings, a contribution of Marc Lehmann.
259 This is normally only activated in post-5.6 perls.
261 Thu Aug 31 23:06:06 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
263 First official release Storable 1.0, for inclusion in perl 5.7.0.
264 The license scheme is now compatible with Perl's.
266 Thu Aug 24 01:02:02 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
271 ANSI-fied most of the code, preparing for Perl core integration.
272 The next version of Storable will be 0.8, and will be integrated
273 into the Perl core (development branch).
275 Dispatch tables were moved upfront to relieve some compilers,
276 especially on AIX and Windows platforms.
278 Merged 64-bit fixes from perl5-porters.
280 Mon Aug 14 09:22:04 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
284 Added a refcnt dec in retrieve_tied_key(): sv_magic() increases
285 the refcnt on the mg_ptr as well.
287 Removed spurious dependency to Devel::Peek, which was used for
288 testing only in t/tied_items.t. Thanks to Conrad Heiney
289 <conrad@fringehead.org> for spotting it first.
291 Sun Aug 13 22:12:59 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
295 Marc Lehmann kindly contributed code to add overloading support
296 and to handle references to tied variables.
298 Rewrote leading blurb about compatibility to make it clearer what
299 "backward compatibility" is about: when I say 0.7 is backward
300 compatible with 0.6, it means the revision 0.7 can read files
303 Mention new Clone(3) extension in SEE ALSO.
305 Was wrongly optimizing for "undef" values in hashes by not
306 fully recursing: as a result, tied "undef" values were incorrectly
309 Sun Jul 30 12:59:17 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
311 First revision of Storable 0.7.
313 The serializing format is new, known as version 2.0. It is fully
314 backward compatible with 0.6. Earlier formats are deprecated and
315 have not even been tested: next version will drop pre-0.6 format.
317 Changes since 0.6@11:
319 - Moved interface to the "beta" status. Some tiny parts are still
320 subject to change, but nothing important enough to warrant an "alpha"
323 - Slightly reduced the size of the Storable image by factorizing
324 object class names and removing final object storage notification due
325 to a redesign of the blessed object storing.
327 - Classes can now redefine how they wish their instances to be serialized
328 and/or deep cloned. Serializing hooks are written in Perl code.
330 - The engine is now fully re-entrant.
332 Sun Apr 2 23:47:50 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
336 Added provision to detect more recent binary formats, since
337 the new upcoming Storable-0.7 will use a different format.
338 In order to prevent attempting the de-serialization of newer
339 formats by older versions, I'm adding this now to the 0.6 series.
341 I'm expecting this revision to be the last of the 0.6 series.
342 Unless it does not work with perl 5.6, which I don't use yet,
343 and therefore against which I cannot test.
345 Wed Mar 29 19:55:21 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
349 Added note about format incompatibilities with old versions
350 (i.e. pre 0.5@9 formats, which cannot be understood as there
351 was no versionning information in the file by then).
353 Protect all $@ variables when eval {} used, to avoid corrupting
354 it when store/retrieve is called within an exception handler.
356 Mistakenly included "patchlevel.h" instead of <patchlevel.h>,
357 preventing Perl's patchlevel from being included, which is
358 needed starting from 5.6.
360 Tue May 12 09:15:15 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
364 Fixed shared "undef" bug in hashes, which did not remain shared
365 through store/retrieve.
367 Thu Feb 10 19:48:16 MET 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
371 added last_op_in_netorder() predicate
372 documented last_op_in_netorder()
373 added tests for the new last_op_in_netorder() predicate
375 Wed Oct 20 19:07:36 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
379 Forgot to update VERSION
381 Tue Oct 19 21:25:02 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
385 Added mention of japanese translation for the manual page.
387 Fixed typo in macro that made threaded code not compilable,
388 especially on Win32 platforms.
390 Changed detection of older perls (pre-5.005) by testing PATCHLEVEL
391 directly instead of relying on internal symbols.
393 Tue Sep 14 22:13:28 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
397 Integrated "thread-safe" patch from Murray Nesbitt.
398 Note that this may not be very efficient for threaded code,
399 see comment in the code.
401 Try to avoid compilation warning on 64-bit CPUs. Can't test it,
402 since I don't have access to such machines.
404 Mon Jul 12 14:37:19 METDST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
408 changed my e-mail to pobox.
410 mentionned it is not thread-safe.
412 updated version number.
414 uses new internal PL_* naming convention.
416 Fri Jul 3 13:38:16 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
420 Updated benchmark figures due to recent optimizations done in
421 store(): tagnums are now stored as-is in the hash table, so
422 no surrounding SV is created. And the "shared keys" mode for
423 hash table was turned off.
425 Fixed backward compatibility (wrt 0.5@9) for retrieval of
426 blessed refs. That old version did something wrong, but the
427 bugfix prevented correct retrieval of the old format.
429 Mon Jun 22 11:00:48 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
433 Changed benchmark figures.
435 Adjust refcnt of tied objects after calling sv_magic() to avoid
436 memory leaks. Contributed by Jeff Gresham.
438 Fri Jun 12 11:50:04 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
442 Added workaround for persistent LVALUE-ness in perl5.004. All
443 scalars tagged as being an lvalue are handled as if they were
444 not an lvalue at all. Added test for that LVALUE bug workaround.
446 Now handles Perl immortal scalars explicitely, by storing &sv_yes
447 as such, explicitely.
449 Retrieval of non-immortal undef cannot be shared. Previous
450 version was over-optimizing by not creating a separate SV for
451 all undefined scalars seen.
453 Thu Jun 4 17:21:51 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
457 Baseline for Storable-0.6@0.
459 This version introduces a binary incompatibility in the generated
460 binary image, which is more compact than older ones by approximatively
461 15%, depending on the exact degree of sharing in your structures.
463 The good news is that your older images can still be retrieved with
464 this version, i.e. backward compatibility is preserved. This version
465 of Storable can only generate new binaries however.
467 Another good news is that the retrieval of data structure is
468 significantly quicker than before, because a Perl array is used
469 instead of a hash table to keep track of retrieved objects, and
470 also because the image being smaller, less I/O function calls are
473 Tue May 12 09:15:15 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
477 Version number now got from Storable.pm directly.
479 Fixed overzealous sv_type() optimization, which would make
480 Storable fail when faced with an "upgraded" SV to the PVIV
481 or PVNV kind containing a reference.
483 Thu Apr 30 15:11:30 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
487 Extended the SYNOPSIS section to give quick overview of the
488 routines and their signature.
490 Optimized sv_type() to avoid flags checking when not needed, i.e.
491 when their type makes it impossible for them to be refs or tied.
492 This slightly increases throughput by a few percents when refs
493 and tied variables are marginal occurrences in your data.
495 Stubs for XS now use OutputStream and InputStream file types to
496 make it work when the given file is actually a socket. Perl
497 makes a distinction for sockets in its internal I/O structures
498 by having both a read and a write structure, whereas plain files
501 Tue Jun 3 09:41:33 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
505 Thanks to a contribution from Benjamin A. Holzman, Storable is now
506 able to correctly serialize tied SVs, i.e. tied arrays, hashes
509 Thu Apr 9 18:07:51 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
513 I said SvPOK() had changed to SvPOKp(), but that was a lie...
515 Wed Apr 8 13:14:29 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
519 Wrote sizeof(SV *) instead of sizeof(I32) when portable, which
520 in effect mangled the object tags and prevented portability
521 accross 32/64 bit architectures!
523 Wed Mar 25 14:57:02 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
527 Added code example for store_fd() and retrieve_fd() in the
528 man page, to emphasize that file descriptors must be passed as
529 globs, not as plain strings.
531 Cannot use SV addresses as tag when using nstore() on LP64. This
532 was the cause of problems when creating a storable image on an
533 LP64 machine and retrieving it on an ILP32 system, which is
534 exactly what nstore() is meant for...
536 However, we continue to use SV addresses as tags for plain store(),
537 because benchamarking shows that it saves up to 8% of the store
538 time, and store() is meant to be fast at the expense of lack
541 This means there will be approximately an 8% degradation of
542 performance for nstore(), but it's now working as expected.
543 That cost may vary on your machine of course, since it is
544 solely caused by the memory allocation overhead used to create
545 unique SV tags for each distinct stored SV.
547 Tue Jan 20 09:21:53 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
551 Don't use any '_' in version number.
553 Tue Jan 13 17:51:50 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
557 Updated version number.
559 added binmode() calls for systems where it matters.
561 Be sure to pass globs, not plain file strings, to C routines,
562 so that Storable can be used under the Perl debugger.
564 Wed Nov 5 10:53:22 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
568 Fix memory leaks on seen hash table and returned SV refs.
570 Storable did not work properly when tainting enabled.
572 Fixed "Allocation too large" messages in freeze/thaw and added.
573 proper regression test in t/freeze.t.
575 Tue Jun 3 09:41:33 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
579 Updated version number
581 Added freeze/thaw interface and dclone.
583 Fri May 16 10:45:47 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
587 Forgot that AutoLoader does not export its own AUTOLOAD.
590 use AutoLoader 'AUTOLOAD';
592 but that would not be backward compatible. So the export is
595 Tue Mar 25 11:21:32 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
599 Empty scalar strings are now "defined" at retrieval time.
601 New test to ensure an empty string is defined when retrieved.
603 Thu Feb 27 16:32:44 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
607 Updated version number
609 Declare VERSION as being used
611 Fixed a typo in the PerlIO_putc remapping.
612 PerlIO_read and perlIO_write inverted size/nb_items.
613 (only relevant for pre-perl5.004 versions)
615 Thu Feb 27 15:58:31 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
619 Updated version number
621 Added VERSION identification
623 Allow build with perl5.003, which is ante perlIO time
625 Mon Jan 13 17:53:18 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
631 Wed Jan 22 15:19:56 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
635 Updated version number in Makefile.PL.
637 Added "thanks to" section to README.
639 Documented new forgive_me variable.
643 Added forgive_me support to allow store() of data structures
644 containing non-storable items like CODE refs.