Commit | Line | Data |
fa523c3a |
1 | Thu Jun 17 12:26:43 BST 2004 Nicholas Clark <nick@ccl4.org> |
2 | |
3 | Version 2.13 |
4 | |
5 | 1. Don't change the type of top level overloaded references to RV - |
6 | they are perfectly correct as PVMG |
7 | 2. Storable needs to cope with incoming frozen data that happens to be |
8 | utf8 encoded. |
9 | |
754c00ca |
10 | Wed Mar 17 15:40:29 GMT 2004 Nicholas Clark <nick@ccl4.org> |
11 | |
3f575d8d |
12 | Version 2.12 |
13 | |
754c00ca |
14 | 1. Add regression tests for the auto-require of STORABLE_thaw |
165cc789 |
15 | 2. Add auto-require of modules to restore overloading (and tests) |
3f575d8d |
16 | 3. Change to no context (should give speedup with ithreads) |
754c00ca |
17 | |
dfd91409 |
18 | Sat Mar 13 20:11:03 GMT 2004 Nicholas Clark <nick@ccl4.org> |
19 | |
20 | Version 2.11 |
21 | |
22 | 1. Storing restricted hashes in canonical order would SEGV. Fixed. |
23 | 2. It was impossible to retrieve references to PL_sv_no and and |
24 | PL_sv_undef from STORABLE_thaw hooks. |
25 | 3. restrict.t was failing on 5.8.0, due to 5.8.0's unique |
26 | implementation of restricted hashes using PL_sv_undef |
27 | 4. These changes allow a space optimisation for restricted hashes. |
28 | |
a8b7ef86 |
29 | Sat Jan 24 16:22:32 IST 2004 Abhijit Menon-Sen <ams@wiw.org> |
30 | |
31 | Version 2.10 |
32 | |
33 | 1. Thread safety: Storable::CLONE/init_perlinterp() now create |
34 | a new Perl context for each new ithread. |
35 | (From Stas Bekman and Jan Dubois.) |
36 | 2. Fix a tag count mismatch with $Storable::Deparse that caused |
37 | all back-references after a stored sub to be off-by-N (where |
38 | N was the number of code references in between). |
39 | (From Sam Vilain.) |
40 | 3. Prevent CODE references from turning into SCALAR references. |
41 | (From Slaven Rezic.) |
42 | |
dcf2e277 |
43 | Sat Jan 3 18:49:18 GMT 2004 Nicholas Clark <nick@ccl4.org> |
44 | |
45 | Version 2.09 |
46 | |
47 | Fix minor problems with the CPAN release |
7224c650 |
48 | 1: Make Storable.xs work on 5.8.2 and later (already in the core) |
dcf2e277 |
49 | 2: Ship the linux hints file |
50 | 3: Ship Test::More for the benefit of Perls pre 5.6.2 |
7224c650 |
51 | 4: Correct Makefile.PL to only install in core for 5.8.0 and later |
dcf2e277 |
52 | |
fcaa57e7 |
53 | Sat Sep 6 01:08:20 IST 2003 Abhijit Menon-Sen <ams@wiw.org> |
54 | |
55 | Version 2.08 |
56 | |
57 | This release works around a 5.8.0 bug which caused hashes to not |
58 | be marked as having key flags even though an HEK had HEK_WASUTF8 |
59 | set. (Note that the only reasonable solution is to silently drop |
60 | the flag from the affected key.) |
61 | |
62 | Users of RT 3 who were seeing assertion failures should upgrade. |
63 | (Perl 5.8.1 will have the bug fixed.) |
64 | |
a5271820 |
65 | Mon May 5 10:24:16 IST 2003 Abhijit Menon-Sen <ams@wiw.org> |
66 | |
67 | Version 2.07 |
68 | |
69 | Minor bugfixes (self-tied objects are now correctly stored, as |
70 | are the results of additions larger than INT_MAX). |
71 | |
a2307be4 |
72 | Mon Oct 7 21:56:38 BST 2002 Nicholas Clark <nick@ccl4.org> |
73 | |
74 | Version 2.06 |
75 | |
76 | Remove qr// from t/downgrade.t so that it will run on 5.004 |
77 | Mention $File::Spec::VERSION a second time in t/forgive.t so that it |
78 | runs without warnings in 5.004 (this may be a 5.00405 bug I'm working |
79 | round) |
80 | Fix t/integer.t initialisation to actually generate 64 bits of 9c |
81 | Fix comparison tests to use eval to get around 64 bit IV conversion |
82 | issues on 5.6.x, following my t/integer.t ^ precedence bug found by |
83 | Rafael Garcia-Suarez |
84 | Alter t/malice.t to work with Test/More.pm in t/, and skip individual |
85 | subtests that use $Config{ptrsize}, so that the rest of the test can |
86 | now be run with 5.004 |
87 | Change t/malice.t and the error message in check_magic in Storable.xs |
88 | from "Pointer integer size" to "Pointer size" |
89 | Remove prerequisite of Test::More from Makefile.PL |
90 | Ship Test::Builder, Test::Simple and Test::More in t |
91 | |
9cfdba97 |
92 | Thu Oct 3 08:57:22 IST 2002 Abhijit Menon-Sen <ams@wiw.org> |
93 | |
94 | Version 2.05 |
95 | |
96 | Adds support for CODE references from Slaven Rezic |
97 | <slaven.rezic@berlin.de>. |
98 | |
17374ab3 |
99 | Fri Jun 7 23:55:41 BST 2002 Nicholas Clark |
100 | |
101 | Version 2.04 |
102 | |
103 | Bug fix from Radu Greab <radu@netsoft.ro> (plus regression test) |
104 | to fix a recently introduced bug detected by Dave Rolsky. |
105 | Bug was that for a non threaded build, the class information was |
106 | being lost at freeze time on the first object with a STORABLE_freeze |
107 | hook. Consequentially the object was not blessed at all when thawed. |
108 | (The presence (or lack) of STORABLE_thaw was irrelevant; this was |
109 | a store-time data lost bug, caused by failure to initialize internal |
110 | context) |
111 | The bug was introduced as development perl change 16442 (on |
112 | 2002/05/07), so has been present since 2.00. |
113 | Patches to introduce more regression tests to reduce the chance of |
114 | a reoccurance of this sort of goof are always welcome. |
115 | |
ee0f7aac |
116 | Thu May 30 20:31:08 BST 2002 Nicholas Clark <nick@ccl4.org> |
117 | |
8932d55c |
118 | Version 2.03 Header changes on 5.6.x on Unix where IV is long long |
119 | |
120 | 5.6.x introduced the ability to have IVs as long long. However, |
121 | Configure still defined BYTEORDER based on the size of a long. |
122 | Storable uses the BYTEORDER value as part of the header, but |
123 | doesn't explicity store sizeof(IV) anywhere in the header. |
124 | Hence on 5.6.x built with IV as long long on a platform that |
125 | uses Configure (ie most things except VMS and Windows) headers |
126 | are identical for the different IV sizes, despite the files |
127 | containing some fields based on sizeof(IV) |
128 | |
129 | 5.8.0 is consistent; all platforms have BYTEORDER in config.h |
130 | based on sizeof(IV) rather than sizeof(long). This means that |
131 | the value of BYTEORDER will change from (say) 4321 to 87654321 |
132 | between 5.6.1 and 5.8.0 built with the same options to Configure |
133 | on the same machine. This means that the Storable header will |
134 | differ, and the two versions will wrongly thing that they are |
135 | incompatible. |
136 | |
137 | For the benefit of long term consistency, Storable now |
138 | implements the 5.8.0 BYTEORDER policy on 5.6.x. This means that |
139 | 2.03 onwards default to be incompatible with 2.02 and earlier |
140 | (ie the large 1.0.x installed base) on the same 5.6.x perl. |
141 | |
142 | To allow interworking, a new variable |
143 | $Storable::interwork_56_64bit is introduced. It defaults to |
144 | false. Set it to true to read and write old format files. Don't |
145 | use it unless you have existing stored data written with 5.6.x |
146 | that you couldn't otherwise read, or you need to interwork with |
147 | a machine running older Storable on a 5.6.x with long long IVs |
148 | (i.e., you probably don't need to use it). |
149 | |
2aeb6432 |
150 | Sat May 25 22:38:39 BST 2002 Nicholas Clark <nick@ccl4.org> |
151 | |
25f64a11 |
152 | Version 2.02 |
153 | |
154 | Rewrite Storable.xs so that the file header structure for write_magic |
155 | is built at compile time, and check_magic attempts to the header in |
156 | blocks rather than byte per byte. These changes make the compiled |
157 | extension 2.25% smaller, but are not significant enough to give a |
158 | noticeable speed up. |
2aeb6432 |
159 | |
db670f21 |
160 | Thu May 23 22:50:41 BST 2002 Nicholas Clark <nick@ccl4.org> |
161 | |
25f64a11 |
162 | Version 2.01 |
163 | |
164 | - New regression tests integer.t |
165 | - Add code to safely store large unsigned integers. |
166 | - Change code not to attempt to store large integers (ie > 32 bits) |
167 | in network order as 32 bits. |
168 | |
169 | *Never* underestimate the value of a pathological test suite carefully |
170 | crafted with maximum malice before writing a line of real code. It |
171 | prevents crafty bugs from stowing away in your released code. |
172 | It's much less embarrassing to find them before you ship. |
173 | (Well, never underestimate it if you ever want to work for me) |
174 | |
8989364b |
175 | Fri May 17 22:48:59 BST 2002 Nicholas Clark <nick@ccl4.org> |
176 | |
25f64a11 |
177 | Version 2.0, binary format 2.5 (but writes format 2.4 on pre 5.7.3) |
178 | |
179 | The perl5 porters have decided to make sure that Storable still |
180 | builds on pre-5.8 perls, and make the 5.8 version available on CPAN. |
181 | The VERSION is now 2.0, and it passes all tests on 5.005_03, 5.6.1 |
182 | and 5.6.1 with threads. On 5.6.0 t/downgrade.t fails tests 34 and 37, |
183 | due to a bug in 5.6.0 - upgrade to 5.6.1. |
184 | |
185 | Jarkko and I have collated the list of changes the perl5 porters have |
186 | from the perl5 Changes file: |
187 | |
188 | - data features of upcoming perl 5.8.0 are supported: Unicode hash |
189 | keys (Unicode hash values have been supported since Storable 1.0.1) |
190 | and "restricted hashes" (readonly hashes and hash entries) |
191 | - a newer version of perl can now be used to serialize data which is |
192 | not supported in earlier perls: Storable will attempt to do the |
193 | right thing for as long as possible, croaking only when safe data |
194 | conversion simply isn't possible. Alternatively earlier perls can |
195 | opt to have a lossy downgrade data instead of croaking |
196 | - when built with perls pre 5.7.3 this Storable writes out files |
197 | with binary format 2.4, the same format as Storable 1.0.8 onwards. |
198 | This should mean that this Storable will inter-operate seamlessly |
199 | with any Storable 1.0.8 or newer on perls pre 5.7.3 |
200 | - dclone() now works with empty string scalar objects |
201 | - retrieving of large hashes is now more efficient |
202 | - more routines autosplit out of the main module, so Storable should |
203 | load slightly more quickly |
204 | - better documentation |
205 | - the internal context objects are now freed explicitly, rather than |
206 | relying on thread or process exit |
207 | - bugs fixed in debugging trace code affecting builds made with 64 bit |
208 | IVs |
209 | - code tidy-ups to allow clean compiles with more warning options |
210 | turned on avoid problems with $@ getting corrupted on 5.005_03 if |
211 | Carp wasn't already loaded |
212 | - added &show_file_magic, so you can add to /etc/magic and teach |
213 | Unix's file command about Storable files |
214 | |
215 | We plan to keep Storable on CPAN in sync with the Perl core, so |
216 | if you encounter bugs or other problems building or using Storable, |
217 | please let us know at perl5-porters@perl.org |
218 | Patches welcome! |
219 | |
596596d5 |
220 | Sat Dec 1 14:37:54 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
221 | |
25f64a11 |
222 | This is the LAST maintenance release of the Storable module. |
223 | Indeed, Storable is now part of perl 5.8, and will be maintained |
224 | as part of Perl. The CPAN module will remain available there |
225 | for people running pre-5.8 perls. |
596596d5 |
226 | |
25f64a11 |
227 | Avoid requiring Fcntl upfront, useful to embedded runtimes. |
228 | Use an eval {} for testing, instead of making Storable.pm |
229 | simply fail its compilation in the BEGIN block. |
596596d5 |
230 | |
25f64a11 |
231 | store_fd() will now correctly autoflush file if needed. |
596596d5 |
232 | |
6e0ac6f5 |
233 | Tue Aug 28 23:53:20 MEST 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
234 | |
25f64a11 |
235 | Fixed truncation race with lock_retrieve() in lock_store(). |
236 | The file has to be truncated only once the exclusive lock is held. |
6e0ac6f5 |
237 | |
25f64a11 |
238 | Removed spurious debugging messages in .xs file. |
6e0ac6f5 |
239 | |
e993d95c |
240 | Sun Jul 1 13:27:32 MEST 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
241 | |
25f64a11 |
242 | Systematically use "=over 4" for POD linters. |
243 | Apparently, POD linters are much stricter than would |
244 | otherwise be needed, but that's OK. |
e993d95c |
245 | |
25f64a11 |
246 | Fixed memory corruption on croaks during thaw(). Thanks |
247 | to Claudio Garcia for reproducing this bug and providing the |
248 | code to exercise it. Added test cases for this bug, adapted |
249 | from Claudio's code. |
e993d95c |
250 | |
25f64a11 |
251 | Made code compile cleanly with -Wall (from Jarkko Hietaniemi). |
e993d95c |
252 | |
25f64a11 |
253 | Changed tagnum and classnum from I32 to IV in context. Also |
254 | from Jarkko. |
e993d95c |
255 | |
8be2b38b |
256 | Thu Mar 15 01:22:32 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
257 | |
25f64a11 |
258 | Last version was wrongly compiling with assertions on, due |
259 | to an edit glitch. That did not cause any problem (apart from |
260 | a slight performance loss) excepted on Win* platforms, where the |
261 | assertion code does not compile. |
8be2b38b |
262 | |
263 | Sat Feb 17 13:37:37 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
264 | |
25f64a11 |
265 | Version 1.0.10. |
8be2b38b |
266 | |
25f64a11 |
267 | Forgot to increase version number at previous patch (there were |
268 | two of them, which is why we jump from 1.0.8 to 1.0.10). |
8be2b38b |
269 | |
b12202d0 |
270 | Sat Feb 17 13:35:00 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
271 | |
25f64a11 |
272 | Version 1.0.8, binary format 2.4. |
b12202d0 |
273 | |
25f64a11 |
274 | Fixed incorrect error message. |
b12202d0 |
275 | |
25f64a11 |
276 | Now bless objects ASAP at retrieve time, which is meant to fix |
277 | two bugs: |
b12202d0 |
278 | |
25f64a11 |
279 | * Indirect references to overloaded object were not able to |
280 | restore overloading if the object was not blessed yet, |
281 | which was possible since blessing occurred only after the |
282 | recursive retrieval. |
b12202d0 |
283 | |
25f64a11 |
284 | * Storable hooks asking for serialization of blessed ref could |
285 | get un-blessed refs at retrieval time, for the very same |
286 | reason. |
b12202d0 |
287 | |
25f64a11 |
288 | The fix implemented here was suggested by Nick Ing-Simmons. |
b12202d0 |
289 | |
25f64a11 |
290 | Added support for blessed ref to tied structures. This is the |
291 | cause for the binary format change. |
b12202d0 |
292 | |
25f64a11 |
293 | Added EBCDIC version of the compatibility test with 0.6.11, |
294 | from Peter Prymmer |
b12202d0 |
295 | |
25f64a11 |
296 | Added tests for the new features, and to make sure the bugs they |
297 | are meant to fix are indeed fixed. |
b12202d0 |
298 | |
862382c7 |
299 | Wed Jan 3 10:43:18 MET 2001 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
300 | |
25f64a11 |
301 | Removed spurious 'clean' entry in Makefile.PL. |
862382c7 |
302 | |
25f64a11 |
303 | Added CAN_FLOCK to determine whether we can flock() or not, |
304 | by inspecting Perl's configuration parameters, as determined |
305 | by Configure. |
862382c7 |
306 | |
25f64a11 |
307 | Trace offending package when overloading cannot be restored |
308 | on a scalar. |
862382c7 |
309 | |
25f64a11 |
310 | Made context cleanup safer to avoid dup freeing, mostly in the |
311 | presence of repeated exceptions during store/retrieve (which can |
312 | cause memory leaks anyway, so it's just additional safety, not a |
313 | definite fix). |
862382c7 |
314 | |
90826881 |
315 | Sun Nov 5 18:23:48 MET 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
316 | |
25f64a11 |
317 | Version 1.0.6. |
90826881 |
318 | |
25f64a11 |
319 | Fixed severe "object lost" bug for STORABLE_freeze returns, |
320 | when refs to lexicals, taken within the hook, were to be |
321 | serialized by Storable. Enhanced the t/recurse.t test to |
322 | stress hook a little more with refs to lexicals. |
90826881 |
323 | |
212e9bde |
324 | Thu Oct 26 19:14:38 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
325 | |
25f64a11 |
326 | Version 1.0.5. |
212e9bde |
327 | |
25f64a11 |
328 | Documented that store() and retrieve() can return undef. |
329 | That is, the error reporting is not always made via exceptions, |
330 | as the paragraph on error reporting was implying. |
212e9bde |
331 | |
25f64a11 |
332 | Auto requires module of blessed ref when STORABLE_thaw misses. |
333 | When the Storable engine looks for the STORABLE_thaw hook and |
334 | does not find it, it now tries to require the package into which |
335 | the blessed reference is. |
212e9bde |
336 | |
25f64a11 |
337 | Just check $^O, in t/lock.t: there's no need to pull the whole |
338 | Config module for that. |
212e9bde |
339 | |
8be2b38b |
340 | Mon Oct 23 20:03:49 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
341 | |
25f64a11 |
342 | Version 1.0.4. |
8be2b38b |
343 | |
25f64a11 |
344 | Protected calls to flock() for DOS platform: apparently, the |
345 | flock/fcnlt emulation is reported to be broken on that |
346 | platform. |
8be2b38b |
347 | |
25f64a11 |
348 | Added logcarp emulation if they don't have Log::Agent, since |
349 | we now use it to carp when lock_store/lock_retrieve is used |
350 | on DOS. |
8be2b38b |
351 | |
dd19458b |
352 | Fri Sep 29 21:52:29 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
353 | |
25f64a11 |
354 | Version 1.0.3. |
dd19458b |
355 | |
25f64a11 |
356 | Avoid using "tainted" and "dirty" since Perl remaps them via |
357 | cpp (i.e. #define). This is deeply harmful when threading |
358 | is enabled. This concerned both the context structure and |
359 | local variable and argument names. Brrr..., scary! |
dd19458b |
360 | |
361 | Thu Sep 28 23:46:39 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
362 | |
25f64a11 |
363 | Version 1.0.2. |
dd19458b |
364 | |
25f64a11 |
365 | Fixed spelling in README. |
dd19458b |
366 | |
25f64a11 |
367 | Added lock_store, lock_nstore, and lock_retrieve (advisory locking) |
368 | after a proposal from Erik Haugan <erik@solbors.no>. |
dd19458b |
369 | |
25f64a11 |
370 | Perls before 5.004_04 lack newSVpvn, added remapping in XS. |
dd19458b |
371 | |
25f64a11 |
372 | Fixed stupid typo in the t/utf8.t test. |
dd19458b |
373 | |
374 | Sun Sep 17 18:51:10 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
375 | |
25f64a11 |
376 | Version 1.0.1, binary format 2.3. |
dd19458b |
377 | |
25f64a11 |
378 | Documented that doubles are stored stringified by nstore(). |
dd19458b |
379 | |
25f64a11 |
380 | Added Salvador Ortiz Garcia in CREDITS section, He identified |
381 | a bug in the store hooks and proposed the right fix: the class |
382 | id was allocated too soon. His bug case was also added to |
383 | the regression test suite. |
dd19458b |
384 | |
25f64a11 |
385 | Now only taint retrieved data when source was tainted. A bug |
386 | discovered by Marc Lehmann. |
dd19458b |
387 | |
25f64a11 |
388 | Added support for UTF-8 strings, a contribution of Marc Lehmann. |
389 | This is normally only activated in post-5.6 perls. |
dd19458b |
390 | |
9e21b3d0 |
391 | Thu Aug 31 23:06:06 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
392 | |
25f64a11 |
393 | First official release Storable 1.0, for inclusion in perl 5.7.0. |
394 | The license scheme is now compatible with Perl's. |
9e21b3d0 |
395 | |
f0ffaed8 |
396 | Thu Aug 24 01:02:02 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
397 | |
25f64a11 |
398 | ANSI-fied most of the code, preparing for Perl core integration. |
399 | The next version of Storable will be 0.8, and will be integrated |
400 | into the Perl core (development branch). |
f0ffaed8 |
401 | |
25f64a11 |
402 | Dispatch tables were moved upfront to relieve some compilers, |
403 | especially on AIX and Windows platforms. |
f0ffaed8 |
404 | |
25f64a11 |
405 | Merged 64-bit fixes from perl5-porters. |
f0ffaed8 |
406 | |
7a6a85bf |
407 | Mon Aug 14 09:22:04 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
408 | |
25f64a11 |
409 | Added a refcnt dec in retrieve_tied_key(): sv_magic() increases |
410 | the refcnt on the mg_ptr as well. |
7a6a85bf |
411 | |
25f64a11 |
412 | Removed spurious dependency to Devel::Peek, which was used for |
413 | testing only in t/tied_items.t. Thanks to Conrad Heiney |
414 | <conrad@fringehead.org> for spotting it first. |
7a6a85bf |
415 | |
416 | Sun Aug 13 22:12:59 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
417 | |
25f64a11 |
418 | Marc Lehmann kindly contributed code to add overloading support |
419 | and to handle references to tied variables. |
7a6a85bf |
420 | |
25f64a11 |
421 | Rewrote leading blurb about compatibility to make it clearer what |
422 | "backward compatibility" is about: when I say 0.7 is backward |
423 | compatible with 0.6, it means the revision 0.7 can read files |
424 | produced by 0.6. |
7a6a85bf |
425 | |
25f64a11 |
426 | Mention new Clone(3) extension in SEE ALSO. |
7a6a85bf |
427 | |
25f64a11 |
428 | Was wrongly optimizing for "undef" values in hashes by not |
429 | fully recursing: as a result, tied "undef" values were incorrectly |
430 | serialized. |
7a6a85bf |
431 | |
432 | Sun Jul 30 12:59:17 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
433 | |
25f64a11 |
434 | First revision of Storable 0.7. |
7a6a85bf |
435 | |
25f64a11 |
436 | The serializing format is new, known as version 2.0. It is fully |
437 | backward compatible with 0.6. Earlier formats are deprecated and |
438 | have not even been tested: next version will drop pre-0.6 format. |
7a6a85bf |
439 | |
25f64a11 |
440 | Changes since 0.6@11: |
7a6a85bf |
441 | |
25f64a11 |
442 | - Moved interface to the "beta" status. Some tiny parts are still |
443 | subject to change, but nothing important enough to warrant an "alpha" |
444 | status any longer. |
7a6a85bf |
445 | |
25f64a11 |
446 | - Slightly reduced the size of the Storable image by factorizing |
447 | object class names and removing final object storage notification due |
448 | to a redesign of the blessed object storing. |
7a6a85bf |
449 | |
25f64a11 |
450 | - Classes can now redefine how they wish their instances to be serialized |
451 | and/or deep cloned. Serializing hooks are written in Perl code. |
7a6a85bf |
452 | |
25f64a11 |
453 | - The engine is now fully re-entrant. |
7a6a85bf |
454 | |
455 | Sun Apr 2 23:47:50 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
456 | |
25f64a11 |
457 | Added provision to detect more recent binary formats, since |
458 | the new upcoming Storable-0.7 will use a different format. |
459 | In order to prevent attempting the de-serialization of newer |
460 | formats by older versions, I'm adding this now to the 0.6 series. |
7a6a85bf |
461 | |
25f64a11 |
462 | I'm expecting this revision to be the last of the 0.6 series. |
463 | Unless it does not work with perl 5.6, which I don't use yet, |
464 | and therefore against which I cannot test. |
7a6a85bf |
465 | |
466 | Wed Mar 29 19:55:21 MEST 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
467 | |
25f64a11 |
468 | Added note about format incompatibilities with old versions |
469 | (i.e. pre 0.5@9 formats, which cannot be understood as there |
470 | was no versionning information in the file by then). |
7a6a85bf |
471 | |
25f64a11 |
472 | Protect all $@ variables when eval {} used, to avoid corrupting |
473 | it when store/retrieve is called within an exception handler. |
7a6a85bf |
474 | |
25f64a11 |
475 | Mistakenly included "patchlevel.h" instead of <patchlevel.h>, |
476 | preventing Perl's patchlevel from being included, which is |
477 | needed starting from 5.6. |
7a6a85bf |
478 | |
479 | Tue May 12 09:15:15 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
480 | |
25f64a11 |
481 | Fixed shared "undef" bug in hashes, which did not remain shared |
482 | through store/retrieve. |
7a6a85bf |
483 | |
484 | Thu Feb 10 19:48:16 MET 2000 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
485 | |
25f64a11 |
486 | added last_op_in_netorder() predicate |
487 | documented last_op_in_netorder() |
488 | added tests for the new last_op_in_netorder() predicate |
7a6a85bf |
489 | |
490 | Wed Oct 20 19:07:36 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
491 | |
25f64a11 |
492 | Forgot to update VERSION |
7a6a85bf |
493 | |
494 | Tue Oct 19 21:25:02 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
495 | |
25f64a11 |
496 | Added mention of japanese translation for the manual page. |
7a6a85bf |
497 | |
25f64a11 |
498 | Fixed typo in macro that made threaded code not compilable, |
499 | especially on Win32 platforms. |
7a6a85bf |
500 | |
25f64a11 |
501 | Changed detection of older perls (pre-5.005) by testing PATCHLEVEL |
502 | directly instead of relying on internal symbols. |
7a6a85bf |
503 | |
504 | Tue Sep 14 22:13:28 MEST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
505 | |
25f64a11 |
506 | Integrated "thread-safe" patch from Murray Nesbitt. |
507 | Note that this may not be very efficient for threaded code, |
508 | see comment in the code. |
7a6a85bf |
509 | |
25f64a11 |
510 | Try to avoid compilation warning on 64-bit CPUs. Can't test it, |
511 | since I don't have access to such machines. |
7a6a85bf |
512 | |
513 | Mon Jul 12 14:37:19 METDST 1999 Raphael Manfredi <Raphael_Manfredi@pobox.com> |
514 | |
25f64a11 |
515 | changed my e-mail to pobox. |
7a6a85bf |
516 | |
25f64a11 |
517 | mentionned it is not thread-safe. |
7a6a85bf |
518 | |
25f64a11 |
519 | updated version number. |
7a6a85bf |
520 | |
25f64a11 |
521 | uses new internal PL_* naming convention. |
7a6a85bf |
522 | |
523 | Fri Jul 3 13:38:16 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
524 | |
25f64a11 |
525 | Updated benchmark figures due to recent optimizations done in |
526 | store(): tagnums are now stored as-is in the hash table, so |
527 | no surrounding SV is created. And the "shared keys" mode for |
528 | hash table was turned off. |
7a6a85bf |
529 | |
25f64a11 |
530 | Fixed backward compatibility (wrt 0.5@9) for retrieval of |
531 | blessed refs. That old version did something wrong, but the |
532 | bugfix prevented correct retrieval of the old format. |
7a6a85bf |
533 | |
534 | Mon Jun 22 11:00:48 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
535 | |
25f64a11 |
536 | Changed benchmark figures. |
7a6a85bf |
537 | |
25f64a11 |
538 | Adjust refcnt of tied objects after calling sv_magic() to avoid |
539 | memory leaks. Contributed by Jeff Gresham. |
7a6a85bf |
540 | |
541 | Fri Jun 12 11:50:04 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
542 | |
25f64a11 |
543 | Added workaround for persistent LVALUE-ness in perl5.004. All |
544 | scalars tagged as being an lvalue are handled as if they were |
545 | not an lvalue at all. Added test for that LVALUE bug workaround. |
7a6a85bf |
546 | |
25f64a11 |
547 | Now handles Perl immortal scalars explicitely, by storing &sv_yes |
548 | as such, explicitely. |
7a6a85bf |
549 | |
25f64a11 |
550 | Retrieval of non-immortal undef cannot be shared. Previous |
551 | version was over-optimizing by not creating a separate SV for |
552 | all undefined scalars seen. |
7a6a85bf |
553 | |
554 | Thu Jun 4 17:21:51 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
555 | |
25f64a11 |
556 | Baseline for Storable-0.6@0. |
7a6a85bf |
557 | |
25f64a11 |
558 | This version introduces a binary incompatibility in the generated |
559 | binary image, which is more compact than older ones by approximatively |
560 | 15%, depending on the exact degree of sharing in your structures. |
7a6a85bf |
561 | |
25f64a11 |
562 | The good news is that your older images can still be retrieved with |
563 | this version, i.e. backward compatibility is preserved. This version |
564 | of Storable can only generate new binaries however. |
7a6a85bf |
565 | |
25f64a11 |
566 | Another good news is that the retrieval of data structure is |
567 | significantly quicker than before, because a Perl array is used |
568 | instead of a hash table to keep track of retrieved objects, and |
569 | also because the image being smaller, less I/O function calls are |
570 | made. |
7a6a85bf |
571 | |
572 | Tue May 12 09:15:15 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
573 | |
25f64a11 |
574 | Version number now got from Storable.pm directly. |
7a6a85bf |
575 | |
25f64a11 |
576 | Fixed overzealous sv_type() optimization, which would make |
577 | Storable fail when faced with an "upgraded" SV to the PVIV |
578 | or PVNV kind containing a reference. |
7a6a85bf |
579 | |
580 | Thu Apr 30 15:11:30 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
581 | |
25f64a11 |
582 | Extended the SYNOPSIS section to give quick overview of the |
583 | routines and their signature. |
7a6a85bf |
584 | |
25f64a11 |
585 | Optimized sv_type() to avoid flags checking when not needed, i.e. |
586 | when their type makes it impossible for them to be refs or tied. |
587 | This slightly increases throughput by a few percents when refs |
588 | and tied variables are marginal occurrences in your data. |
7a6a85bf |
589 | |
25f64a11 |
590 | Stubs for XS now use OutputStream and InputStream file types to |
591 | make it work when the given file is actually a socket. Perl |
592 | makes a distinction for sockets in its internal I/O structures |
593 | by having both a read and a write structure, whereas plain files |
594 | share the same one. |
7a6a85bf |
595 | |
596 | Tue Jun 3 09:41:33 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
597 | |
25f64a11 |
598 | Thanks to a contribution from Benjamin A. Holzman, Storable is now |
599 | able to correctly serialize tied SVs, i.e. tied arrays, hashes |
600 | and scalars. |
7a6a85bf |
601 | |
602 | Thu Apr 9 18:07:51 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
603 | |
25f64a11 |
604 | I said SvPOK() had changed to SvPOKp(), but that was a lie... |
7a6a85bf |
605 | |
606 | Wed Apr 8 13:14:29 METDST 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
607 | |
25f64a11 |
608 | Wrote sizeof(SV *) instead of sizeof(I32) when portable, which |
609 | in effect mangled the object tags and prevented portability |
610 | accross 32/64 bit architectures! |
7a6a85bf |
611 | |
612 | Wed Mar 25 14:57:02 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
613 | |
25f64a11 |
614 | Added code example for store_fd() and retrieve_fd() in the |
615 | man page, to emphasize that file descriptors must be passed as |
616 | globs, not as plain strings. |
7a6a85bf |
617 | |
25f64a11 |
618 | Cannot use SV addresses as tag when using nstore() on LP64. This |
619 | was the cause of problems when creating a storable image on an |
620 | LP64 machine and retrieving it on an ILP32 system, which is |
621 | exactly what nstore() is meant for... |
7a6a85bf |
622 | |
25f64a11 |
623 | However, we continue to use SV addresses as tags for plain store(), |
624 | because benchamarking shows that it saves up to 8% of the store |
625 | time, and store() is meant to be fast at the expense of lack |
626 | of portability. |
7a6a85bf |
627 | |
25f64a11 |
628 | This means there will be approximately an 8% degradation of |
629 | performance for nstore(), but it's now working as expected. |
630 | That cost may vary on your machine of course, since it is |
631 | solely caused by the memory allocation overhead used to create |
632 | unique SV tags for each distinct stored SV. |
7a6a85bf |
633 | |
634 | Tue Jan 20 09:21:53 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
635 | |
25f64a11 |
636 | Don't use any '_' in version number. |
7a6a85bf |
637 | |
638 | Tue Jan 13 17:51:50 MET 1998 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
639 | |
25f64a11 |
640 | Updated version number. |
7a6a85bf |
641 | |
25f64a11 |
642 | added binmode() calls for systems where it matters. |
7a6a85bf |
643 | |
25f64a11 |
644 | Be sure to pass globs, not plain file strings, to C routines, |
645 | so that Storable can be used under the Perl debugger. |
7a6a85bf |
646 | |
647 | Wed Nov 5 10:53:22 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
648 | |
25f64a11 |
649 | Fix memory leaks on seen hash table and returned SV refs. |
7a6a85bf |
650 | |
25f64a11 |
651 | Storable did not work properly when tainting enabled. |
7a6a85bf |
652 | |
25f64a11 |
653 | Fixed "Allocation too large" messages in freeze/thaw and added. |
654 | proper regression test in t/freeze.t. |
7a6a85bf |
655 | |
656 | Tue Jun 3 09:41:33 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
657 | |
25f64a11 |
658 | Updated version number |
7a6a85bf |
659 | |
25f64a11 |
660 | Added freeze/thaw interface and dclone. |
7a6a85bf |
661 | |
662 | Fri May 16 10:45:47 METDST 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
663 | |
25f64a11 |
664 | Forgot that AutoLoader does not export its own AUTOLOAD. |
665 | I could use |
7a6a85bf |
666 | |
25f64a11 |
667 | use AutoLoader 'AUTOLOAD'; |
668 | |
669 | but that would not be backward compatible. So the export is |
670 | done by hand... |
7a6a85bf |
671 | |
672 | Tue Mar 25 11:21:32 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
673 | |
25f64a11 |
674 | Empty scalar strings are now "defined" at retrieval time. |
7a6a85bf |
675 | |
25f64a11 |
676 | New test to ensure an empty string is defined when retrieved. |
7a6a85bf |
677 | |
678 | Thu Feb 27 16:32:44 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
679 | |
25f64a11 |
680 | Updated version number |
7a6a85bf |
681 | |
25f64a11 |
682 | Declare VERSION as being used |
7a6a85bf |
683 | |
25f64a11 |
684 | Fixed a typo in the PerlIO_putc remapping. |
685 | PerlIO_read and perlIO_write inverted size/nb_items. |
686 | (only relevant for pre-perl5.004 versions) |
7a6a85bf |
687 | |
688 | Thu Feb 27 15:58:31 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
689 | |
25f64a11 |
690 | Updated version number |
7a6a85bf |
691 | |
25f64a11 |
692 | Added VERSION identification |
7a6a85bf |
693 | |
25f64a11 |
694 | Allow build with perl5.003, which is ante perlIO time |
7a6a85bf |
695 | |
696 | Mon Jan 13 17:53:18 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
697 | |
25f64a11 |
698 | Random code fixes. |
7a6a85bf |
699 | |
700 | Wed Jan 22 15:19:56 MET 1997 Raphael Manfredi <Raphael_Manfredi@grenoble.hp.com> |
701 | |
25f64a11 |
702 | Updated version number in Makefile.PL. |
7a6a85bf |
703 | |
25f64a11 |
704 | Added "thanks to" section to README. |
7a6a85bf |
705 | |
25f64a11 |
706 | Documented new forgive_me variable. |
7a6a85bf |
707 | |
25f64a11 |
708 | Made 64-bit clean. |
7a6a85bf |
709 | |
25f64a11 |
710 | Added forgive_me support to allow store() of data structures |
711 | containing non-storable items like CODE refs. |