Once again syncing after too long an absence
[p5sagit/p5-mst-13.2.git] / ext / Storable / ChangeLog
CommitLineData
0e06870b 1Wed Jan 3 10:43:18 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com>
2
3. Description:
4
5 Removed spurious 'clean' entry in Makefile.PL.
6
7 Added CAN_FLOCK to determine whether we can flock() or not,
8 by inspecting Perl's configuration parameters, as determined
9 by Configure.
10
11 Trace offending package when overloading cannot be restored
12 on a scalar.
13
14 Made context cleanup safer to avoid dup freeing, mostly in the
15 presence of repeated exceptions during store/retrieve (which can
16 cause memory leaks anyway, so it's just additional safety, not a
17 definite fix).
18
19Sun Nov 5 18:23:48 MET 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
20
21. Description:
22
23 Version 1.0.6.
24
25 Fixed severe "object lost" bug for STORABLE_freeze returns,
26 when refs to lexicals, taken within the hook, were to be
27 serialized by Storable. Enhanced the t/recurse.t test to
28 stress hook a little more with refs to lexicals.
29
30Thu Oct 26 19:14:38 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
31
32. Description:
33
34 Version 1.0.5.
35
36 Documented that store() and retrieve() can return undef.
37 That is, the error reporting is not always made via exceptions,
38 as the paragraph on error reporting was implying.
39
40 Auto requires module of blessed ref when STORABLE_thaw misses.
41 When the Storable engine looks for the STORABLE_thaw hook and
42 does not find it, it now tries to require the package into which
43 the blessed reference is.
44
45 Just check $^O, in t/lock.t: there's no need to pull the whole
46 Config module for that.
47
22d4bb9c 48Fri Sep 29 21:52:29 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
49
50. Description:
51
52 Version 1.0.3.
53
54 Avoid using "tainted" and "dirty" since Perl remaps them via
55 cpp (i.e. #define). This is deeply harmful when threading
56 is enabled. This concerned both the context structure and
57 local variable and argument names. Brrr..., scary!
58
59Thu Sep 28 23:46:39 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
60
61. Description:
62
63 Version 1.0.2.
64
65 Fixed spelling in README.
66
67 Added lock_store, lock_nstore, and lock_retrieve (advisory locking)
68 after a proposal from Erik Haugan <erik@solbors.no>.
69
70 Perls before 5.004_04 lack newSVpvn, added remapping in XS.
71
72 Fixed stupid typo in the t/utf8.t test.
73
74Sun Sep 17 18:51:10 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
75
76. Description:
77
78 Version 1.0.1, binary format 2.3.
79
80 Documented that doubles are stored stringified by nstore().
81
82 Added Salvador Ortiz Garcia in CREDITS section, He identified
83 a bug in the store hooks and proposed the right fix: the class
84 id was allocated too soon. His bug case was also added to
85 the regression test suite.
86
87 Now only taint retrieved data when source was tainted. A bug
88 discovered by Marc Lehmann.
89
90 Added support for UTF-8 strings, a contribution of Marc Lehmann.
91 This is normally only activated in post-5.6 perls.
92
93Thu Aug 31 23:06:06 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
94
95 First official release Storable 1.0, for inclusion in perl 5.7.0.
96 The license scheme is now compatible with Perl's.
97
98Thu Aug 24 01:02:02 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
99
100. Description:
101
102
103 ANSI-fied most of the code, preparing for Perl core integration.
104 The next version of Storable will be 0.8, and will be integrated
105 into the Perl core (development branch).
106
107 Dispatch tables were moved upfront to relieve some compilers,
108 especially on AIX and Windows platforms.
109
110 Merged 64-bit fixes from perl5-porters.
111
112Mon Aug 14 09:22:04 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
113
114. Description:
115
116 Added a refcnt dec in retrieve_tied_key(): sv_magic() increases
117 the refcnt on the mg_ptr as well.
118
119 Removed spurious dependency to Devel::Peek, which was used for
120 testing only in t/tied_items.t. Thanks to Conrad Heiney
121 <conrad@fringehead.org> for spotting it first.
122
123Sun Aug 13 22:12:59 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
124
125. Description:
126
127 Marc Lehmann kindly contributed code to add overloading support
128 and to handle references to tied variables.
129
130 Rewrote leading blurb about compatibility to make it clearer what
131 "backward compatibility" is about: when I say 0.7 is backward
132 compatible with 0.6, it means the revision 0.7 can read files
133 produced by 0.6.
134
135 Mention new Clone(3) extension in SEE ALSO.
136
137 Was wrongly optimizing for "undef" values in hashes by not
138 fully recursing: as a result, tied "undef" values were incorrectly
139 serialized.
140
141Sun Jul 30 12:59:17 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
142
143 First revision of Storable 0.7.
144
145 The serializing format is new, known as version 2.0. It is fully
146 backward compatible with 0.6. Earlier formats are deprecated and
147 have not even been tested: next version will drop pre-0.6 format.
148
149 Changes since 0.6@11:
150
151 - Moved interface to the "beta" status. Some tiny parts are still
152 subject to change, but nothing important enough to warrant an "alpha"
153 status any longer.
154
155 - Slightly reduced the size of the Storable image by factorizing
156 object class names and removing final object storage notification due
157 to a redesign of the blessed object storing.
158
159 - Classes can now redefine how they wish their instances to be serialized
160 and/or deep cloned. Serializing hooks are written in Perl code.
161
162 - The engine is now fully re-entrant.
163
164Sun Apr 2 23:47:50 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
165
166. Description:
167
168 Added provision to detect more recent binary formats, since
169 the new upcoming Storable-0.7 will use a different format.
170 In order to prevent attempting the de-serialization of newer
171 formats by older versions, I'm adding this now to the 0.6 series.
172
173 I'm expecting this revision to be the last of the 0.6 series.
174 Unless it does not work with perl 5.6, which I don't use yet,
175 and therefore against which I cannot test.
176
177Wed Mar 29 19:55:21 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
178
179. Description:
180
181 Added note about format incompatibilities with old versions
182 (i.e. pre 0.5@9 formats, which cannot be understood as there
183 was no versionning information in the file by then).
184
185 Protect all $@ variables when eval {} used, to avoid corrupting
186 it when store/retrieve is called within an exception handler.
187
188 Mistakenly included "patchlevel.h" instead of <patchlevel.h>,
189 preventing Perl's patchlevel from being included, which is
190 needed starting from 5.6.
191
192Tue May 12 09:15:15 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
193
194. Description:
195
196 Fixed shared "undef" bug in hashes, which did not remain shared
197 through store/retrieve.
198
199Thu Feb 10 19:48:16 MET 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com>
200
201. Description:
202
203 added last_op_in_netorder() predicate
204 documented last_op_in_netorder()
205 added tests for the new last_op_in_netorder() predicate
206
207Wed Oct 20 19:07:36 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
208
209. Description:
210
211 Forgot to update VERSION
212
213Tue Oct 19 21:25:02 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
214
215. Description:
216
217 Added mention of japanese translation for the manual page.
218
219 Fixed typo in macro that made threaded code not compilable,
220 especially on Win32 platforms.
221
222 Changed detection of older perls (pre-5.005) by testing PATCHLEVEL
223 directly instead of relying on internal symbols.
224
225Tue Sep 14 22:13:28 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
226
227. Description:
228
229 Integrated "thread-safe" patch from Murray Nesbitt.
230 Note that this may not be very efficient for threaded code,
231 see comment in the code.
232
233 Try to avoid compilation warning on 64-bit CPUs. Can't test it,
234 since I don't have access to such machines.
235
236Mon Jul 12 14:37:19 METDST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com>
237
238. Description:
239
240 changed my e-mail to pobox.
241
242 mentionned it is not thread-safe.
243
244 updated version number.
245
246 uses new internal PL_* naming convention.
247
248Fri Jul 3 13:38:16 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
249
250. Description:
251
252 Updated benchmark figures due to recent optimizations done in
253 store(): tagnums are now stored as-is in the hash table, so
254 no surrounding SV is created. And the "shared keys" mode for
255 hash table was turned off.
256
257 Fixed backward compatibility (wrt 0.5@9) for retrieval of
258 blessed refs. That old version did something wrong, but the
259 bugfix prevented correct retrieval of the old format.
260
261Mon Jun 22 11:00:48 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
262
263. Description:
264
265 Changed benchmark figures.
266
267 Adjust refcnt of tied objects after calling sv_magic() to avoid
268 memory leaks. Contributed by Jeff Gresham.
269
270Fri Jun 12 11:50:04 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
271
272. Description:
273
274 Added workaround for persistent LVALUE-ness in perl5.004. All
275 scalars tagged as being an lvalue are handled as if they were
276 not an lvalue at all. Added test for that LVALUE bug workaround.
277
278 Now handles Perl immortal scalars explicitely, by storing &sv_yes
279 as such, explicitely.
280
281 Retrieval of non-immortal undef cannot be shared. Previous
282 version was over-optimizing by not creating a separate SV for
283 all undefined scalars seen.
284
285Thu Jun 4 17:21:51 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
286
287. Description:
288
289 Baseline for Storable-0.6@0.
290
291 This version introduces a binary incompatibility in the generated
292 binary image, which is more compact than older ones by approximatively
293 15%, depending on the exact degree of sharing in your structures.
294
295 The good news is that your older images can still be retrieved with
296 this version, i.e. backward compatibility is preserved. This version
297 of Storable can only generate new binaries however.
298
299 Another good news is that the retrieval of data structure is
300 significantly quicker than before, because a Perl array is used
301 instead of a hash table to keep track of retrieved objects, and
302 also because the image being smaller, less I/O function calls are
303 made.
304
305Tue May 12 09:15:15 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
306
307. Description:
308
309 Version number now got from Storable.pm directly.
310
311 Fixed overzealous sv_type() optimization, which would make
312 Storable fail when faced with an "upgraded" SV to the PVIV
313 or PVNV kind containing a reference.
314
315Thu Apr 30 15:11:30 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
316
317. Description:
318
319 Extended the SYNOPSIS section to give quick overview of the
320 routines and their signature.
321
322 Optimized sv_type() to avoid flags checking when not needed, i.e.
323 when their type makes it impossible for them to be refs or tied.
324 This slightly increases throughput by a few percents when refs
325 and tied variables are marginal occurrences in your data.
326
327 Stubs for XS now use OutputStream and InputStream file types to
328 make it work when the given file is actually a socket. Perl
329 makes a distinction for sockets in its internal I/O structures
330 by having both a read and a write structure, whereas plain files
331 share the same one.
332
333Tue Jun 3 09:41:33 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
334
335. Description:
336
337 Thanks to a contribution from Benjamin A. Holzman, Storable is now
338 able to correctly serialize tied SVs, i.e. tied arrays, hashes
339 and scalars.
340
341Thu Apr 9 18:07:51 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
342
343. Description:
344
345 I said SvPOK() had changed to SvPOKp(), but that was a lie...
346
347Wed Apr 8 13:14:29 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
348
349. Description:
350
351 Wrote sizeof(SV *) instead of sizeof(I32) when portable, which
352 in effect mangled the object tags and prevented portability
353 accross 32/64 bit architectures!
354
355Wed Mar 25 14:57:02 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
356
357. Description:
358
359 Added code example for store_fd() and retrieve_fd() in the
360 man page, to emphasize that file descriptors must be passed as
361 globs, not as plain strings.
362
363 Cannot use SV addresses as tag when using nstore() on LP64. This
364 was the cause of problems when creating a storable image on an
365 LP64 machine and retrieving it on an ILP32 system, which is
366 exactly what nstore() is meant for...
367
368 However, we continue to use SV addresses as tags for plain store(),
369 because benchamarking shows that it saves up to 8% of the store
370 time, and store() is meant to be fast at the expense of lack
371 of portability.
372
373 This means there will be approximately an 8% degradation of
374 performance for nstore(), but it's now working as expected.
375 That cost may vary on your machine of course, since it is
376 solely caused by the memory allocation overhead used to create
377 unique SV tags for each distinct stored SV.
378
379Tue Jan 20 09:21:53 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
380
381. Description:
382
383 Don't use any '_' in version number.
384
385Tue Jan 13 17:51:50 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
386
387. Description:
388
389 Updated version number.
390
391 added binmode() calls for systems where it matters.
392
393 Be sure to pass globs, not plain file strings, to C routines,
394 so that Storable can be used under the Perl debugger.
395
396Wed Nov 5 10:53:22 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
397
398. Description:
399
400 Fix memory leaks on seen hash table and returned SV refs.
401
402 Storable did not work properly when tainting enabled.
403
404 Fixed "Allocation too large" messages in freeze/thaw and added.
405 proper regression test in t/freeze.t.
406
407Tue Jun 3 09:41:33 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
408
409. Description:
410
411 Updated version number
412
413 Added freeze/thaw interface and dclone.
414
415Fri May 16 10:45:47 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
416
417. Description:
418
419 Forgot that AutoLoader does not export its own AUTOLOAD.
420 I could use
421
422 use AutoLoader 'AUTOLOAD';
423
424 but that would not be backward compatible. So the export is
425 done by hand...
426
427Tue Mar 25 11:21:32 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
428
429. Description:
430
431 Empty scalar strings are now "defined" at retrieval time.
432
433 New test to ensure an empty string is defined when retrieved.
434
435Thu Feb 27 16:32:44 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
436
437. Description:
438
439 Updated version number
440
441 Declare VERSION as being used
442
443 Fixed a typo in the PerlIO_putc remapping.
444 PerlIO_read and perlIO_write inverted size/nb_items.
445 (only relevant for pre-perl5.004 versions)
446
447Thu Feb 27 15:58:31 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
448
449. Description:
450
451 Updated version number
452
453 Added VERSION identification
454
455 Allow build with perl5.003, which is ante perlIO time
456
457Mon Jan 13 17:53:18 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
458
459. Description:
460
461 Random code fixes.
462
463Wed Jan 22 15:19:56 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com>
464
465. Description:
466
467 Updated version number in Makefile.PL.
468
469 Added "thanks to" section to README.
470
471 Documented new forgive_me variable.
472
473 Made 64-bit clean.
474
475 Added forgive_me support to allow store() of data structures
476 containing non-storable items like CODE refs.
477