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