1 Thu May 30 20:31:08 BST 2002 Nicholas Clark <nick@ccl4.org>
3 Version 2.03 Header changes on 5.6.x on Unix where IV is long long
5 5.6.x introduced the ability to have IVs as long long. However,
6 Configure still defined BYTEORDER based on the size of a long.
7 Storable uses the BYTEORDER value as part of the header, but
8 doesn't explicity store sizeof(IV) anywhere in the header.
9 Hence on 5.6.x built with IV as long long on a platform that
10 uses Configure (ie most things except VMS and Windows) headers
11 are identical for the different IV sizes, despite the files
12 containing some fields based on sizeof(IV)
14 5.8.0 is consistent; all platforms have BYTEORDER in config.h
15 based on sizeof(IV) rather than sizeof(long). This means that
16 the value of BYTEORDER will change from (say) 4321 to 87654321
17 between 5.6.1 and 5.8.0 built with the same options to Configure
18 on the same machine. This means that the Storable header will
19 differ, and the two versions will wrongly thing that they are
22 For the benefit of long term consistency, Storable now
23 implements the 5.8.0 BYTEORDER policy on 5.6.x. This means that
24 2.03 onwards default to be incompatible with 2.02 and earlier
25 (ie the large 1.0.x installed base) on the same 5.6.x perl.
27 To allow interworking, a new variable
28 $Storable::interwork_56_64bit is introduced. It defaults to
29 false. Set it to true to read and write old format files. Don't
30 use it unless you have existing stored data written with 5.6.x
31 that you couldn't otherwise read, or you need to interwork with
32 a machine running older Storable on a 5.6.x with long long IVs
33 (i.e., you probably don't need to use it).
35 Sat May 25 22:38:39 BST 2002 Nicholas Clark <nick@ccl4.org>
39 Rewrite Storable.xs so that the file header structure for write_magic
40 is built at compile time, and check_magic attempts to the header in
41 blocks rather than byte per byte. These changes make the compiled
42 extension 2.25% smaller, but are not significant enough to give a
45 Thu May 23 22:50:41 BST 2002 Nicholas Clark <nick@ccl4.org>
49 - New regression tests integer.t
50 - Add code to safely store large unsigned integers.
51 - Change code not to attempt to store large integers (ie > 32 bits)
52 in network order as 32 bits.
54 *Never* underestimate the value of a pathological test suite carefully
55 crafted with maximum malice before writing a line of real code. It
56 prevents crafty bugs from stowing away in your released code.
57 It's much less embarrassing to find them before you ship.
58 (Well, never underestimate it if you ever want to work for me)
60 Fri May 17 22:48:59 BST 2002 Nicholas Clark <nick@ccl4.org>
62 Version 2.0, binary format 2.5 (but writes format 2.4 on pre 5.7.3)
64 The perl5 porters have decided to make sure that Storable still
65 builds on pre-5.8 perls, and make the 5.8 version available on CPAN.
66 The VERSION is now 2.0, and it passes all tests on 5.005_03, 5.6.1
67 and 5.6.1 with threads. On 5.6.0 t/downgrade.t fails tests 34 and 37,
68 due to a bug in 5.6.0 - upgrade to 5.6.1.
70 Jarkko and I have collated the list of changes the perl5 porters have
71 from the perl5 Changes file:
73 - data features of upcoming perl 5.8.0 are supported: Unicode hash
74 keys (Unicode hash values have been supported since Storable 1.0.1)
75 and "restricted hashes" (readonly hashes and hash entries)
76 - a newer version of perl can now be used to serialize data which is
77 not supported in earlier perls: Storable will attempt to do the
78 right thing for as long as possible, croaking only when safe data
79 conversion simply isn't possible. Alternatively earlier perls can
80 opt to have a lossy downgrade data instead of croaking
81 - when built with perls pre 5.7.3 this Storable writes out files
82 with binary format 2.4, the same format as Storable 1.0.8 onwards.
83 This should mean that this Storable will inter-operate seamlessly
84 with any Storable 1.0.8 or newer on perls pre 5.7.3
85 - dclone() now works with empty string scalar objects
86 - retrieving of large hashes is now more efficient
87 - more routines autosplit out of the main module, so Storable should
88 load slightly more quickly
89 - better documentation
90 - the internal context objects are now freed explicitly, rather than
91 relying on thread or process exit
92 - bugs fixed in debugging trace code affecting builds made with 64 bit
94 - code tidy-ups to allow clean compiles with more warning options
95 turned on avoid problems with $@ getting corrupted on 5.005_03 if
96 Carp wasn't already loaded
97 - added &show_file_magic, so you can add to /etc/magic and teach
98 Unix's file command about Storable files
100 We plan to keep Storable on CPAN in sync with the Perl core, so
101 if you encounter bugs or other problems building or using Storable,
102 please let us know at perl5-porters@perl.org
105 Sat Dec 1 14:37:54 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
107 This is the LAST maintenance release of the Storable module.
108 Indeed, Storable is now part of perl 5.8, and will be maintained
109 as part of Perl. The CPAN module will remain available there
110 for people running pre-5.8 perls.
112 Avoid requiring Fcntl upfront, useful to embedded runtimes.
113 Use an eval {} for testing, instead of making Storable.pm
114 simply fail its compilation in the BEGIN block.
116 store_fd() will now correctly autoflush file if needed.
118 Tue Aug 28 23:53:20 MEST 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
120 Fixed truncation race with lock_retrieve() in lock_store().
121 The file has to be truncated only once the exclusive lock is held.
123 Removed spurious debugging messages in .xs file.
125 Sun Jul 1 13:27:32 MEST 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
127 Systematically use "=over 4" for POD linters.
128 Apparently, POD linters are much stricter than would
129 otherwise be needed, but that's OK.
131 Fixed memory corruption on croaks during thaw(). Thanks
132 to Claudio Garcia for reproducing this bug and providing the
133 code to exercise it. Added test cases for this bug, adapted
136 Made code compile cleanly with -Wall (from Jarkko Hietaniemi).
138 Changed tagnum and classnum from I32 to IV in context. Also
141 Thu Mar 15 01:22:32 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
143 Last version was wrongly compiling with assertions on, due
144 to an edit glitch. That did not cause any problem (apart from
145 a slight performance loss) excepted on Win* platforms, where the
146 assertion code does not compile.
148 Sat Feb 17 13:37:37 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
152 Forgot to increase version number at previous patch (there were
153 two of them, which is why we jump from 1.0.8 to 1.0.10).
155 Sat Feb 17 13:35:00 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
157 Version 1.0.8, binary format 2.4.
159 Fixed incorrect error message.
161 Now bless objects ASAP at retrieve time, which is meant to fix
164 * Indirect references to overloaded object were not able to
165 restore overloading if the object was not blessed yet,
166 which was possible since blessing occurred only after the
169 * Storable hooks asking for serialization of blessed ref could
170 get un-blessed refs at retrieval time, for the very same
173 The fix implemented here was suggested by Nick Ing-Simmons.
175 Added support for blessed ref to tied structures. This is the
176 cause for the binary format change.
178 Added EBCDIC version of the compatibility test with 0.6.11,
181 Added tests for the new features, and to make sure the bugs they
182 are meant to fix are indeed fixed.
184 Wed Jan 3 10:43:18 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
186 Removed spurious 'clean' entry in Makefile.PL.
188 Added CAN_FLOCK to determine whether we can flock() or not,
189 by inspecting Perl's configuration parameters, as determined
192 Trace offending package when overloading cannot be restored
195 Made context cleanup safer to avoid dup freeing, mostly in the
196 presence of repeated exceptions during store/retrieve (which can
197 cause memory leaks anyway, so it's just additional safety, not a
200 Sun Nov 5 18:23:48 MET 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
204 Fixed severe "object lost" bug for STORABLE_freeze returns,
205 when refs to lexicals, taken within the hook, were to be
206 serialized by Storable. Enhanced the t/recurse.t test to
207 stress hook a little more with refs to lexicals.
209 Thu Oct 26 19:14:38 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
213 Documented that store() and retrieve() can return undef.
214 That is, the error reporting is not always made via exceptions,
215 as the paragraph on error reporting was implying.
217 Auto requires module of blessed ref when STORABLE_thaw misses.
218 When the Storable engine looks for the STORABLE_thaw hook and
219 does not find it, it now tries to require the package into which
220 the blessed reference is.
222 Just check $^O, in t/lock.t: there's no need to pull the whole
223 Config module for that.
225 Mon Oct 23 20:03:49 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
229 Protected calls to flock() for DOS platform: apparently, the
230 flock/fcnlt emulation is reported to be broken on that
233 Added logcarp emulation if they don't have Log::Agent, since
234 we now use it to carp when lock_store/lock_retrieve is used
237 Fri Sep 29 21:52:29 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
241 Avoid using "tainted" and "dirty" since Perl remaps them via
242 cpp (i.e. #define). This is deeply harmful when threading
243 is enabled. This concerned both the context structure and
244 local variable and argument names. Brrr..., scary!
246 Thu Sep 28 23:46:39 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
250 Fixed spelling in README.
252 Added lock_store, lock_nstore, and lock_retrieve (advisory locking)
253 after a proposal from Erik Haugan <erik@solbors.no>.
255 Perls before 5.004_04 lack newSVpvn, added remapping in XS.
257 Fixed stupid typo in the t/utf8.t test.
259 Sun Sep 17 18:51:10 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
261 Version 1.0.1, binary format 2.3.
263 Documented that doubles are stored stringified by nstore().
265 Added Salvador Ortiz Garcia in CREDITS section, He identified
266 a bug in the store hooks and proposed the right fix: the class
267 id was allocated too soon. His bug case was also added to
268 the regression test suite.
270 Now only taint retrieved data when source was tainted. A bug
271 discovered by Marc Lehmann.
273 Added support for UTF-8 strings, a contribution of Marc Lehmann.
274 This is normally only activated in post-5.6 perls.
276 Thu Aug 31 23:06:06 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
278 First official release Storable 1.0, for inclusion in perl 5.7.0.
279 The license scheme is now compatible with Perl's.
281 Thu Aug 24 01:02:02 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
283 ANSI-fied most of the code, preparing for Perl core integration.
284 The next version of Storable will be 0.8, and will be integrated
285 into the Perl core (development branch).
287 Dispatch tables were moved upfront to relieve some compilers,
288 especially on AIX and Windows platforms.
290 Merged 64-bit fixes from perl5-porters.
292 Mon Aug 14 09:22:04 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
294 Added a refcnt dec in retrieve_tied_key(): sv_magic() increases
295 the refcnt on the mg_ptr as well.
297 Removed spurious dependency to Devel::Peek, which was used for
298 testing only in t/tied_items.t. Thanks to Conrad Heiney
299 <conrad@fringehead.org> for spotting it first.
301 Sun Aug 13 22:12:59 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
303 Marc Lehmann kindly contributed code to add overloading support
304 and to handle references to tied variables.
306 Rewrote leading blurb about compatibility to make it clearer what
307 "backward compatibility" is about: when I say 0.7 is backward
308 compatible with 0.6, it means the revision 0.7 can read files
311 Mention new Clone(3) extension in SEE ALSO.
313 Was wrongly optimizing for "undef" values in hashes by not
314 fully recursing: as a result, tied "undef" values were incorrectly
317 Sun Jul 30 12:59:17 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
319 First revision of Storable 0.7.
321 The serializing format is new, known as version 2.0. It is fully
322 backward compatible with 0.6. Earlier formats are deprecated and
323 have not even been tested: next version will drop pre-0.6 format.
325 Changes since 0.6@11:
327 - Moved interface to the "beta" status. Some tiny parts are still
328 subject to change, but nothing important enough to warrant an "alpha"
331 - Slightly reduced the size of the Storable image by factorizing
332 object class names and removing final object storage notification due
333 to a redesign of the blessed object storing.
335 - Classes can now redefine how they wish their instances to be serialized
336 and/or deep cloned. Serializing hooks are written in Perl code.
338 - The engine is now fully re-entrant.
340 Sun Apr 2 23:47:50 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
342 Added provision to detect more recent binary formats, since
343 the new upcoming Storable-0.7 will use a different format.
344 In order to prevent attempting the de-serialization of newer
345 formats by older versions, I'm adding this now to the 0.6 series.
347 I'm expecting this revision to be the last of the 0.6 series.
348 Unless it does not work with perl 5.6, which I don't use yet,
349 and therefore against which I cannot test.
351 Wed Mar 29 19:55:21 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
353 Added note about format incompatibilities with old versions
354 (i.e. pre 0.5@9 formats, which cannot be understood as there
355 was no versionning information in the file by then).
357 Protect all $@ variables when eval {} used, to avoid corrupting
358 it when store/retrieve is called within an exception handler.
360 Mistakenly included "patchlevel.h" instead of <patchlevel.h>,
361 preventing Perl's patchlevel from being included, which is
362 needed starting from 5.6.
364 Tue May 12 09:15:15 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
366 Fixed shared "undef" bug in hashes, which did not remain shared
367 through store/retrieve.
369 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>
377 Forgot to update VERSION
379 Tue Oct 19 21:25:02 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
381 Added mention of japanese translation for the manual page.
383 Fixed typo in macro that made threaded code not compilable,
384 especially on Win32 platforms.
386 Changed detection of older perls (pre-5.005) by testing PATCHLEVEL
387 directly instead of relying on internal symbols.
389 Tue Sep 14 22:13:28 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
391 Integrated "thread-safe" patch from Murray Nesbitt.
392 Note that this may not be very efficient for threaded code,
393 see comment in the code.
395 Try to avoid compilation warning on 64-bit CPUs. Can't test it,
396 since I don't have access to such machines.
398 Mon Jul 12 14:37:19 METDST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
400 changed my e-mail to pobox.
402 mentionned it is not thread-safe.
404 updated version number.
406 uses new internal PL_* naming convention.
408 Fri Jul 3 13:38:16 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
410 Updated benchmark figures due to recent optimizations done in
411 store(): tagnums are now stored as-is in the hash table, so
412 no surrounding SV is created. And the "shared keys" mode for
413 hash table was turned off.
415 Fixed backward compatibility (wrt 0.5@9) for retrieval of
416 blessed refs. That old version did something wrong, but the
417 bugfix prevented correct retrieval of the old format.
419 Mon Jun 22 11:00:48 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
421 Changed benchmark figures.
423 Adjust refcnt of tied objects after calling sv_magic() to avoid
424 memory leaks. Contributed by Jeff Gresham.
426 Fri Jun 12 11:50:04 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
428 Added workaround for persistent LVALUE-ness in perl5.004. All
429 scalars tagged as being an lvalue are handled as if they were
430 not an lvalue at all. Added test for that LVALUE bug workaround.
432 Now handles Perl immortal scalars explicitely, by storing &sv_yes
433 as such, explicitely.
435 Retrieval of non-immortal undef cannot be shared. Previous
436 version was over-optimizing by not creating a separate SV for
437 all undefined scalars seen.
439 Thu Jun 4 17:21:51 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
441 Baseline for Storable-0.6@0.
443 This version introduces a binary incompatibility in the generated
444 binary image, which is more compact than older ones by approximatively
445 15%, depending on the exact degree of sharing in your structures.
447 The good news is that your older images can still be retrieved with
448 this version, i.e. backward compatibility is preserved. This version
449 of Storable can only generate new binaries however.
451 Another good news is that the retrieval of data structure is
452 significantly quicker than before, because a Perl array is used
453 instead of a hash table to keep track of retrieved objects, and
454 also because the image being smaller, less I/O function calls are
457 Tue May 12 09:15:15 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
459 Version number now got from Storable.pm directly.
461 Fixed overzealous sv_type() optimization, which would make
462 Storable fail when faced with an "upgraded" SV to the PVIV
463 or PVNV kind containing a reference.
465 Thu Apr 30 15:11:30 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
467 Extended the SYNOPSIS section to give quick overview of the
468 routines and their signature.
470 Optimized sv_type() to avoid flags checking when not needed, i.e.
471 when their type makes it impossible for them to be refs or tied.
472 This slightly increases throughput by a few percents when refs
473 and tied variables are marginal occurrences in your data.
475 Stubs for XS now use OutputStream and InputStream file types to
476 make it work when the given file is actually a socket. Perl
477 makes a distinction for sockets in its internal I/O structures
478 by having both a read and a write structure, whereas plain files
481 Tue Jun 3 09:41:33 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
483 Thanks to a contribution from Benjamin A. Holzman, Storable is now
484 able to correctly serialize tied SVs, i.e. tied arrays, hashes
487 Thu Apr 9 18:07:51 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
489 I said SvPOK() had changed to SvPOKp(), but that was a lie...
491 Wed Apr 8 13:14:29 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
493 Wrote sizeof(SV *) instead of sizeof(I32) when portable, which
494 in effect mangled the object tags and prevented portability
495 accross 32/64 bit architectures!
497 Wed Mar 25 14:57:02 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
499 Added code example for store_fd() and retrieve_fd() in the
500 man page, to emphasize that file descriptors must be passed as
501 globs, not as plain strings.
503 Cannot use SV addresses as tag when using nstore() on LP64. This
504 was the cause of problems when creating a storable image on an
505 LP64 machine and retrieving it on an ILP32 system, which is
506 exactly what nstore() is meant for...
508 However, we continue to use SV addresses as tags for plain store(),
509 because benchamarking shows that it saves up to 8% of the store
510 time, and store() is meant to be fast at the expense of lack
513 This means there will be approximately an 8% degradation of
514 performance for nstore(), but it's now working as expected.
515 That cost may vary on your machine of course, since it is
516 solely caused by the memory allocation overhead used to create
517 unique SV tags for each distinct stored SV.
519 Tue Jan 20 09:21:53 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
521 Don't use any '_' in version number.
523 Tue Jan 13 17:51:50 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
525 Updated version number.
527 added binmode() calls for systems where it matters.
529 Be sure to pass globs, not plain file strings, to C routines,
530 so that Storable can be used under the Perl debugger.
532 Wed Nov 5 10:53:22 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
534 Fix memory leaks on seen hash table and returned SV refs.
536 Storable did not work properly when tainting enabled.
538 Fixed "Allocation too large" messages in freeze/thaw and added.
539 proper regression test in t/freeze.t.
541 Tue Jun 3 09:41:33 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
543 Updated version number
545 Added freeze/thaw interface and dclone.
547 Fri May 16 10:45:47 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
549 Forgot that AutoLoader does not export its own AUTOLOAD.
552 use AutoLoader 'AUTOLOAD';
554 but that would not be backward compatible. So the export is
557 Tue Mar 25 11:21:32 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
559 Empty scalar strings are now "defined" at retrieval time.
561 New test to ensure an empty string is defined when retrieved.
563 Thu Feb 27 16:32:44 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
565 Updated version number
567 Declare VERSION as being used
569 Fixed a typo in the PerlIO_putc remapping.
570 PerlIO_read and perlIO_write inverted size/nb_items.
571 (only relevant for pre-perl5.004 versions)
573 Thu Feb 27 15:58:31 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
575 Updated version number
577 Added VERSION identification
579 Allow build with perl5.003, which is ante perlIO time
581 Mon Jan 13 17:53:18 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
585 Wed Jan 22 15:19:56 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
587 Updated version number in Makefile.PL.
589 Added "thanks to" section to README.
591 Documented new forgive_me variable.
595 Added forgive_me support to allow store() of data structures
596 containing non-storable items like CODE refs.