1 Thu May 30 20:31:08 BST 2002 Nicholas Clark <nick@ccl4.org>
5 Version 2.03 Header changes on 5.6.x on Unix where IV is long long
7 5.6.x introduced the ability to have IVs as long long. However,
8 Configure still defined BYTEORDER based on the size of a long.
9 Storable uses the BYTEORDER value as part of the header, but doesn't
10 explicity store sizeof(IV) anywhere in the header. Hence on 5.6.x
11 built with IV as long long on a platform that uses Configure (ie most
12 things except VMS and Windows) headers are identical for the different
13 IV sizes, despite the files containing some fields based on sizeof(IV)
15 5.8.0 is consistent; all platforms have BYTEORDER in config.h based on
16 sizeof(IV) rather than sizeof(long). This means that the value of
17 BYTEORDER will change from (say) 4321 to 87654321 between 5.6.1 and
18 5.8.0 built with the same options to Configure on the same machine.
19 This means that the Storable header will differ, and the two versions
20 will wrongly thing that they are incompatible.
22 For the benefit of long term consistency, Storable now implements the
23 5.8.0 BYTEORDER policy on 5.6.x. This means that 2.03 onwards default
24 to be incompatible with 2.02 and earlier (ie the large 1.0.x installed
25 base) on the same 5.6.x perl.
27 To allow interworking, a new variable $Storable::interwork_56_64bit
28 is introduced. It defaults to false. Set it to true to read and
29 write old format files. Don't use it unless you have existing
30 stored data written with 5.6.x that you couldn't otherwise read,
31 or you need to interwork with a machine running older Storable on
32 a 5.6.x with long long IVs. ie you probably don't need to use it.
34 Sat May 25 22:38:39 BST 2002 Nicholas Clark <nick@ccl4.org>
38 Rewrite Storable.xs so that the file header structure for write_magic
39 is built at compile time, and check_magic attempts to the header in
40 blocks rather than byte per byte. These changes make the compiled
41 extension 2.25% smaller, but are not significant enough to give a
44 Thu May 23 22:50:41 BST 2002 Nicholas Clark <nick@ccl4.org>
48 - New regression tests integer.t
49 - Add code to safely store large unsigned integers.
50 - Change code not to attempt to store large integers (ie > 32 bits)
51 in network order as 32 bits.
53 *Never* underestimate the value of a pathological test suite carefully
54 crafted with maximum malice before writing a line of real code. It
55 prevents crafty bugs from stowing away in your released code.
56 It's much less embarrassing to find them before you ship.
57 (Well, never underestimate it if you ever want to work for me)
59 Fri May 17 22:48:59 BST 2002 Nicholas Clark <nick@ccl4.org>
61 Version 2.0, binary format 2.5 (but writes format 2.4 on pre 5.7.3)
63 The perl5 porters have decided to make sure that Storable still
64 builds on pre-5.8 perls, and make the 5.8 version available on CPAN.
65 The VERSION is now 2.0, and it passes all tests on 5.005_03, 5.6.1
66 and 5.6.1 with threads. On 5.6.0 t/downgrade.t fails tests 34 and 37,
67 due to a bug in 5.6.0 - upgrade to 5.6.1.
69 Jarkko and I have collated the list of changes the perl5 porters have
70 from the perl5 Changes file:
72 - data features of upcoming perl 5.8.0 are supported: Unicode hash
73 keys (Unicode hash values have been supported since Storable 1.0.1)
74 and "restricted hashes" (readonly hashes and hash entries)
75 - a newer version of perl can now be used to serialize data which is
76 not supported in earlier perls: Storable will attempt to do the
77 right thing for as long as possible, croaking only when safe data
78 conversion simply isn't possible. Alternatively earlier perls can
79 opt to have a lossy downgrade data instead of croaking
80 - when built with perls pre 5.7.3 this Storable writes out files
81 with binary format 2.4, the same format as Storable 1.0.8 onwards.
82 This should mean that this Storable will inter-operate seamlessly
83 with any Storable 1.0.8 or newer on perls pre 5.7.3
84 - dclone() now works with empty string scalar objects
85 - retrieving of large hashes is now more efficient
86 - more routines autosplit out of the main module, so Storable should
87 load slightly more quickly
88 - better documentation
89 - the internal context objects are now freed explicitly, rather than
90 relying on thread or process exit
91 - bugs fixed in debugging trace code affecting builds made with 64 bit
93 - code tidy-ups to allow clean compiles with more warning options
94 turned on avoid problems with $@ getting corrupted on 5.005_03 if
95 Carp wasn't already loaded
96 - added &show_file_magic, so you can add to /etc/magic and teach
97 Unix's file command about Storable files
99 We plan to keep Storable on CPAN in sync with the Perl core, so
100 if you encounter bugs or other problems building or using Storable,
101 please let us know at perl5-porters@perl.org
104 Sat Dec 1 14:37:54 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
106 This is the LAST maintenance release of the Storable module.
107 Indeed, Storable is now part of perl 5.8, and will be maintained
108 as part of Perl. The CPAN module will remain available there
109 for people running pre-5.8 perls.
111 Avoid requiring Fcntl upfront, useful to embedded runtimes.
112 Use an eval {} for testing, instead of making Storable.pm
113 simply fail its compilation in the BEGIN block.
115 store_fd() will now correctly autoflush file if needed.
117 Tue Aug 28 23:53:20 MEST 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
119 Fixed truncation race with lock_retrieve() in lock_store().
120 The file has to be truncated only once the exclusive lock is held.
122 Removed spurious debugging messages in .xs file.
124 Sun Jul 1 13:27:32 MEST 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
126 Systematically use "=over 4" for POD linters.
127 Apparently, POD linters are much stricter than would
128 otherwise be needed, but that's OK.
130 Fixed memory corruption on croaks during thaw(). Thanks
131 to Claudio Garcia for reproducing this bug and providing the
132 code to exercise it. Added test cases for this bug, adapted
135 Made code compile cleanly with -Wall (from Jarkko Hietaniemi).
137 Changed tagnum and classnum from I32 to IV in context. Also
140 Thu Mar 15 01:22:32 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
142 Last version was wrongly compiling with assertions on, due
143 to an edit glitch. That did not cause any problem (apart from
144 a slight performance loss) excepted on Win* platforms, where the
145 assertion code does not compile.
147 Sat Feb 17 13:37:37 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
151 Forgot to increase version number at previous patch (there were
152 two of them, which is why we jump from 1.0.8 to 1.0.10).
154 Sat Feb 17 13:35:00 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
156 Version 1.0.8, binary format 2.4.
158 Fixed incorrect error message.
160 Now bless objects ASAP at retrieve time, which is meant to fix
163 * Indirect references to overloaded object were not able to
164 restore overloading if the object was not blessed yet,
165 which was possible since blessing occurred only after the
168 * Storable hooks asking for serialization of blessed ref could
169 get un-blessed refs at retrieval time, for the very same
172 The fix implemented here was suggested by Nick Ing-Simmons.
174 Added support for blessed ref to tied structures. This is the
175 cause for the binary format change.
177 Added EBCDIC version of the compatibility test with 0.6.11,
180 Added tests for the new features, and to make sure the bugs they
181 are meant to fix are indeed fixed.
183 Wed Jan 3 10:43:18 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
185 Removed spurious 'clean' entry in Makefile.PL.
187 Added CAN_FLOCK to determine whether we can flock() or not,
188 by inspecting Perl's configuration parameters, as determined
191 Trace offending package when overloading cannot be restored
194 Made context cleanup safer to avoid dup freeing, mostly in the
195 presence of repeated exceptions during store/retrieve (which can
196 cause memory leaks anyway, so it's just additional safety, not a
199 Sun Nov 5 18:23:48 MET 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
203 Fixed severe "object lost" bug for STORABLE_freeze returns,
204 when refs to lexicals, taken within the hook, were to be
205 serialized by Storable. Enhanced the t/recurse.t test to
206 stress hook a little more with refs to lexicals.
208 Thu Oct 26 19:14:38 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
212 Documented that store() and retrieve() can return undef.
213 That is, the error reporting is not always made via exceptions,
214 as the paragraph on error reporting was implying.
216 Auto requires module of blessed ref when STORABLE_thaw misses.
217 When the Storable engine looks for the STORABLE_thaw hook and
218 does not find it, it now tries to require the package into which
219 the blessed reference is.
221 Just check $^O, in t/lock.t: there's no need to pull the whole
222 Config module for that.
224 Mon Oct 23 20:03:49 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
228 Protected calls to flock() for DOS platform: apparently, the
229 flock/fcnlt emulation is reported to be broken on that
232 Added logcarp emulation if they don't have Log::Agent, since
233 we now use it to carp when lock_store/lock_retrieve is used
236 Fri Sep 29 21:52:29 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
240 Avoid using "tainted" and "dirty" since Perl remaps them via
241 cpp (i.e. #define). This is deeply harmful when threading
242 is enabled. This concerned both the context structure and
243 local variable and argument names. Brrr..., scary!
245 Thu Sep 28 23:46:39 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
249 Fixed spelling in README.
251 Added lock_store, lock_nstore, and lock_retrieve (advisory locking)
252 after a proposal from Erik Haugan <erik@solbors.no>.
254 Perls before 5.004_04 lack newSVpvn, added remapping in XS.
256 Fixed stupid typo in the t/utf8.t test.
258 Sun Sep 17 18:51:10 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
260 Version 1.0.1, binary format 2.3.
262 Documented that doubles are stored stringified by nstore().
264 Added Salvador Ortiz Garcia in CREDITS section, He identified
265 a bug in the store hooks and proposed the right fix: the class
266 id was allocated too soon. His bug case was also added to
267 the regression test suite.
269 Now only taint retrieved data when source was tainted. A bug
270 discovered by Marc Lehmann.
272 Added support for UTF-8 strings, a contribution of Marc Lehmann.
273 This is normally only activated in post-5.6 perls.
275 Thu Aug 31 23:06:06 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
277 First official release Storable 1.0, for inclusion in perl 5.7.0.
278 The license scheme is now compatible with Perl's.
280 Thu Aug 24 01:02:02 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
282 ANSI-fied most of the code, preparing for Perl core integration.
283 The next version of Storable will be 0.8, and will be integrated
284 into the Perl core (development branch).
286 Dispatch tables were moved upfront to relieve some compilers,
287 especially on AIX and Windows platforms.
289 Merged 64-bit fixes from perl5-porters.
291 Mon Aug 14 09:22:04 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
293 Added a refcnt dec in retrieve_tied_key(): sv_magic() increases
294 the refcnt on the mg_ptr as well.
296 Removed spurious dependency to Devel::Peek, which was used for
297 testing only in t/tied_items.t. Thanks to Conrad Heiney
298 <conrad@fringehead.org> for spotting it first.
300 Sun Aug 13 22:12:59 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
302 Marc Lehmann kindly contributed code to add overloading support
303 and to handle references to tied variables.
305 Rewrote leading blurb about compatibility to make it clearer what
306 "backward compatibility" is about: when I say 0.7 is backward
307 compatible with 0.6, it means the revision 0.7 can read files
310 Mention new Clone(3) extension in SEE ALSO.
312 Was wrongly optimizing for "undef" values in hashes by not
313 fully recursing: as a result, tied "undef" values were incorrectly
316 Sun Jul 30 12:59:17 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
318 First revision of Storable 0.7.
320 The serializing format is new, known as version 2.0. It is fully
321 backward compatible with 0.6. Earlier formats are deprecated and
322 have not even been tested: next version will drop pre-0.6 format.
324 Changes since 0.6@11:
326 - Moved interface to the "beta" status. Some tiny parts are still
327 subject to change, but nothing important enough to warrant an "alpha"
330 - Slightly reduced the size of the Storable image by factorizing
331 object class names and removing final object storage notification due
332 to a redesign of the blessed object storing.
334 - Classes can now redefine how they wish their instances to be serialized
335 and/or deep cloned. Serializing hooks are written in Perl code.
337 - The engine is now fully re-entrant.
339 Sun Apr 2 23:47:50 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
341 Added provision to detect more recent binary formats, since
342 the new upcoming Storable-0.7 will use a different format.
343 In order to prevent attempting the de-serialization of newer
344 formats by older versions, I'm adding this now to the 0.6 series.
346 I'm expecting this revision to be the last of the 0.6 series.
347 Unless it does not work with perl 5.6, which I don't use yet,
348 and therefore against which I cannot test.
350 Wed Mar 29 19:55:21 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
352 Added note about format incompatibilities with old versions
353 (i.e. pre 0.5@9 formats, which cannot be understood as there
354 was no versionning information in the file by then).
356 Protect all $@ variables when eval {} used, to avoid corrupting
357 it when store/retrieve is called within an exception handler.
359 Mistakenly included "patchlevel.h" instead of <patchlevel.h>,
360 preventing Perl's patchlevel from being included, which is
361 needed starting from 5.6.
363 Tue May 12 09:15:15 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
365 Fixed shared "undef" bug in hashes, which did not remain shared
366 through store/retrieve.
368 Thu Feb 10 19:48:16 MET 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
370 added last_op_in_netorder() predicate
371 documented last_op_in_netorder()
372 added tests for the new last_op_in_netorder() predicate
374 Wed Oct 20 19:07:36 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
376 Forgot to update VERSION
378 Tue Oct 19 21:25:02 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
380 Added mention of japanese translation for the manual page.
382 Fixed typo in macro that made threaded code not compilable,
383 especially on Win32 platforms.
385 Changed detection of older perls (pre-5.005) by testing PATCHLEVEL
386 directly instead of relying on internal symbols.
388 Tue Sep 14 22:13:28 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
390 Integrated "thread-safe" patch from Murray Nesbitt.
391 Note that this may not be very efficient for threaded code,
392 see comment in the code.
394 Try to avoid compilation warning on 64-bit CPUs. Can't test it,
395 since I don't have access to such machines.
397 Mon Jul 12 14:37:19 METDST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
399 changed my e-mail to pobox.
401 mentionned it is not thread-safe.
403 updated version number.
405 uses new internal PL_* naming convention.
407 Fri Jul 3 13:38:16 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
409 Updated benchmark figures due to recent optimizations done in
410 store(): tagnums are now stored as-is in the hash table, so
411 no surrounding SV is created. And the "shared keys" mode for
412 hash table was turned off.
414 Fixed backward compatibility (wrt 0.5@9) for retrieval of
415 blessed refs. That old version did something wrong, but the
416 bugfix prevented correct retrieval of the old format.
418 Mon Jun 22 11:00:48 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
420 Changed benchmark figures.
422 Adjust refcnt of tied objects after calling sv_magic() to avoid
423 memory leaks. Contributed by Jeff Gresham.
425 Fri Jun 12 11:50:04 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
427 Added workaround for persistent LVALUE-ness in perl5.004. All
428 scalars tagged as being an lvalue are handled as if they were
429 not an lvalue at all. Added test for that LVALUE bug workaround.
431 Now handles Perl immortal scalars explicitely, by storing &sv_yes
432 as such, explicitely.
434 Retrieval of non-immortal undef cannot be shared. Previous
435 version was over-optimizing by not creating a separate SV for
436 all undefined scalars seen.
438 Thu Jun 4 17:21:51 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
440 Baseline for Storable-0.6@0.
442 This version introduces a binary incompatibility in the generated
443 binary image, which is more compact than older ones by approximatively
444 15%, depending on the exact degree of sharing in your structures.
446 The good news is that your older images can still be retrieved with
447 this version, i.e. backward compatibility is preserved. This version
448 of Storable can only generate new binaries however.
450 Another good news is that the retrieval of data structure is
451 significantly quicker than before, because a Perl array is used
452 instead of a hash table to keep track of retrieved objects, and
453 also because the image being smaller, less I/O function calls are
456 Tue May 12 09:15:15 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
458 Version number now got from Storable.pm directly.
460 Fixed overzealous sv_type() optimization, which would make
461 Storable fail when faced with an "upgraded" SV to the PVIV
462 or PVNV kind containing a reference.
464 Thu Apr 30 15:11:30 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
466 Extended the SYNOPSIS section to give quick overview of the
467 routines and their signature.
469 Optimized sv_type() to avoid flags checking when not needed, i.e.
470 when their type makes it impossible for them to be refs or tied.
471 This slightly increases throughput by a few percents when refs
472 and tied variables are marginal occurrences in your data.
474 Stubs for XS now use OutputStream and InputStream file types to
475 make it work when the given file is actually a socket. Perl
476 makes a distinction for sockets in its internal I/O structures
477 by having both a read and a write structure, whereas plain files
480 Tue Jun 3 09:41:33 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
482 Thanks to a contribution from Benjamin A. Holzman, Storable is now
483 able to correctly serialize tied SVs, i.e. tied arrays, hashes
486 Thu Apr 9 18:07:51 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
488 I said SvPOK() had changed to SvPOKp(), but that was a lie...
490 Wed Apr 8 13:14:29 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
492 Wrote sizeof(SV *) instead of sizeof(I32) when portable, which
493 in effect mangled the object tags and prevented portability
494 accross 32/64 bit architectures!
496 Wed Mar 25 14:57:02 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
498 Added code example for store_fd() and retrieve_fd() in the
499 man page, to emphasize that file descriptors must be passed as
500 globs, not as plain strings.
502 Cannot use SV addresses as tag when using nstore() on LP64. This
503 was the cause of problems when creating a storable image on an
504 LP64 machine and retrieving it on an ILP32 system, which is
505 exactly what nstore() is meant for...
507 However, we continue to use SV addresses as tags for plain store(),
508 because benchamarking shows that it saves up to 8% of the store
509 time, and store() is meant to be fast at the expense of lack
512 This means there will be approximately an 8% degradation of
513 performance for nstore(), but it's now working as expected.
514 That cost may vary on your machine of course, since it is
515 solely caused by the memory allocation overhead used to create
516 unique SV tags for each distinct stored SV.
518 Tue Jan 20 09:21:53 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
520 Don't use any '_' in version number.
522 Tue Jan 13 17:51:50 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
524 Updated version number.
526 added binmode() calls for systems where it matters.
528 Be sure to pass globs, not plain file strings, to C routines,
529 so that Storable can be used under the Perl debugger.
531 Wed Nov 5 10:53:22 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
533 Fix memory leaks on seen hash table and returned SV refs.
535 Storable did not work properly when tainting enabled.
537 Fixed "Allocation too large" messages in freeze/thaw and added.
538 proper regression test in t/freeze.t.
540 Tue Jun 3 09:41:33 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
542 Updated version number
544 Added freeze/thaw interface and dclone.
546 Fri May 16 10:45:47 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
548 Forgot that AutoLoader does not export its own AUTOLOAD.
551 use AutoLoader 'AUTOLOAD';
553 but that would not be backward compatible. So the export is
556 Tue Mar 25 11:21:32 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
558 Empty scalar strings are now "defined" at retrieval time.
560 New test to ensure an empty string is defined when retrieved.
562 Thu Feb 27 16:32:44 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
564 Updated version number
566 Declare VERSION as being used
568 Fixed a typo in the PerlIO_putc remapping.
569 PerlIO_read and perlIO_write inverted size/nb_items.
570 (only relevant for pre-perl5.004 versions)
572 Thu Feb 27 15:58:31 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
574 Updated version number
576 Added VERSION identification
578 Allow build with perl5.003, which is ante perlIO time
580 Mon Jan 13 17:53:18 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
584 Wed Jan 22 15:19:56 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
586 Updated version number in Makefile.PL.
588 Added "thanks to" section to README.
590 Documented new forgive_me variable.
594 Added forgive_me support to allow store() of data structures
595 containing non-storable items like CODE refs.