ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/JSON-XS/Changes
Revision: 1.188
Committed: Mon Nov 19 10:27:29 2018 UTC (5 years, 5 months ago) by root
Branch: MAIN
CVS Tags: rel-4_0
Changes since 1.187: +1 -2 lines
Log Message:
4.0

File Contents

# Content
1 Revision history for Perl extension JSON::XS
2
3 TODO: maybe detect and croak on more invalid inputs (e.g. +-inf/nan)
4 TODO: maybe avoid the reblessing and better support readonly objects.
5 TODO: how to cope with tagged values and standard json decoders
6 TODO: investigate magic (Eric Brine)
7 TODO: [PATCH] Types::Serialiser: Inline true(), false() and error() functions
8
9 4.0 Fri Nov 16 00:06:54 CET 2018
10 - 4.0 pre-release, do not use other than for testing.
11 - SECURITY IMPLICATION: this release enables allow_nonref by default
12 for compatibnility with RFC 7159 and newer. See "old" vs. "new"
13 JSON under SECURITY CONSIDERATIONS.
14 - reworked the "old" vs. "new" JSON section.
15 - add ->boolean_values to provide the values to which booleans
16 decode (requested by Aristotle Pagaltzis).
17 - decode would wrongly accept ASCII NUL characters instead of
18 reporting them as trailing garbage.
19 - work around what smells like a perl bug w.r.t. exceptions
20 thrown in callbacks.
21 - incremental parser now more or less respects allow_nonref.
22 - json_xs json-pretty now enables canonical mode.
23 - add documentation section about I-JSON.
24 - minor documentation fixes/updates.
25
26 3.04 Thu Aug 17 04:30:47 CEST 2017
27 - change exponential realloc algorithm on encoding and string decoding to be
28 really exponential (this helps slow/debugging allocators such as libumem)
29 (reported by Matthew Horsfall).
30 - string encoding would needlessly overallocate output space
31 (testcase by Matthew Horsfall).
32 - be very paranoid about extending buffer lengths and croak if buffers get too large,
33 which might (or might not) improve security.
34 - add cbor-packed type to json_xs.
35 - switch from YAML to YAML::XS in json_xs, as YAML is way too buggy and outdated.
36
37 3.03 Wed Nov 16 20:20:59 CET 2016
38 - fix a bug introduced by a perl bug workaround that would cause
39 incremental parsing to fail with a sv_chop panic.
40 - json_xs: toformat failure error message fixed.
41 - json_xs: allow cyclic data structures in CBOR.
42
43 3.02 Fri Feb 26 22:45:20 CET 2016
44 - allow_nonref now affects booleans (\1, $Types::Serialiser::Boolean)
45 as well (reported by Alex Efros).
46 - allow literal tabs in strings in relaxed mode (patch by
47 lubo.rintel@gooddata.com).
48 - support "cbor" format in json_xs tool.
49 - support (and fix) calling encode and decode in list context
50 (reported by Вадим Власов).
51 - work around a bug in older perls crashing when presented
52 with shared hash keys (Reini Urban).
53 - use stability canary.
54
55 3.01 Tue Oct 29 16:55:15 CET 2013
56 - backport to perls < 5.18 (reported by Paul Howarth).
57
58 3.0 Tue Oct 29 01:35:37 CET 2013
59 - implemented an object tagging extension (using the
60 Types::Serialiser serialisation protocol).
61 - reworked the documentation regarding object serialisation,
62 add a new OBJECT SERIALISATION section that explains th
63 whole process.
64 - new setting: allow_tags.
65 - switch to Types::Serialiser booleans.
66 - remove to_json/from_json.
67 - other minor improvements to the documentation.
68
69 2.34 Thu May 23 11:30:34 CEST 2013
70 - work around bugs in perl 5.18 breaking more than 100
71 widely used modules, without a fix in sight because
72 p5pers don't care about CPAN anymore.
73 - when canonicalising, only allocate up to 64 hash key
74 pointers on the stack. for larger hashes, use the heap,
75 to avoid using too much stackspace.
76 - discuss the problem with setlocale (reported by a few victims).
77
78 2.33 Wed Aug 1 21:03:52 CEST 2012
79 - internal encode/decode XS wrappers did not expect stack
80 moves caused by callbacks (analyzed and testcase by Jesse Luehrs).
81 - add bencode as to/from option in bin/json_xs.
82 - add -e option to json_xs, and none and string in/out formats.
83
84 2.32 Thu Aug 11 19:06:38 CEST 2011
85 - fix a bug in the initial whitespace accumulation.
86
87 2.31 Wed Jul 27 17:53:05 CEST 2011
88 - don't accumulate initial whitespace in the incremental buffer
89 (this can be useful to allow whitespace-keepalive on a tcp
90 connection without triggering the max_size limit).
91 - properly croak on some invalid inputs that are not strings
92 (e.g. undef) when trying to decode a json text (reported
93 and analyzed by Goro Fuji).
94
95 2.3 Wed Aug 18 01:26:47 CEST 2010
96 - make sure decoder doesn't change the decoding in the incremental
97 parser (testcase provided by Hendrik Schumacher).
98 - applied patch by DaTa for Data::Dumper support in json_xs.
99 - added -t dump support to json_xs, using Data::Dump.
100 - added -f eval support to json_xs.
101
102 2.29 Wed Mar 17 02:39:12 CET 2010
103 - fix a memory leak when callbacks set using filter_json_object
104 or filter_json_single_key_object were called (great testcase
105 by Eric Wilhelm).
106
107 2.28 Thu Mar 11 20:30:46 CET 2010
108 - implement our own atof function - perl's can be orders of
109 magnitudes slower than even the system one. on the positive
110 side, ours seems to be more exact in general than perl's.
111 (testcase provided by Tim Meadowcroft).
112 - clarify floating point conversion issues a bit.
113 - update jpsykes csrf article url.
114 - updated benchmark section - JSON::PP became much faster!
115
116 2.27 Thu Jan 7 07:35:08 CET 2010
117 - support relaxed option inside the incremental parser
118 (testcase provided by IKEGAMI via Makamaka).
119
120 2.26 Sat Oct 10 03:26:19 CEST 2009
121 - big integers could become truncated (based on patch
122 by Strobl Anton).
123 - output format change: indent now adds a final newline, which is
124 more expected and more true to the documentation.
125
126 2.25 Sat Aug 8 12:04:41 CEST 2009
127 - the perl debugger completely breaks lvalue subs - try to work
128 around the issue.
129 - ignore RMAGICAL hashes w.r.t. CANONICAL.
130 - try to work around a possible char signedness issue on aix.
131 - require common sense.
132
133 2.24 Sat May 30 08:25:45 CEST 2009
134 - the incremental parser did not update its parse offset
135 pointer correctly when parsing utf8-strings (nicely
136 debugged by Martin Evans).
137 - appending a non-utf8-string to the incremental parser
138 in utf8 mode failed to upgrade the string.
139 - wording of parse error messages has been improved.
140
141 2.232 Sun Feb 22 11:12:25 CET 2009
142 - use an exponential algorithm to extend strings, to
143 help platforms with bad or abysmal==windows memory
144 allocater performance, at the expense of some memory
145 wastage (use shrink to recover this extra memory).
146 (nicely analysed by Dmitry Karasik).
147
148 2.2311 Thu Feb 19 02:12:54 CET 2009
149 - add a section "JSON and ECMAscript" to explain some
150 incompatibilities between the two (problem was noted by
151 various people).
152 - add t/20_faihu.t.
153
154 2.231 Thu Nov 20 04:59:08 CET 2008
155 - work around 5.10.0 magic bugs where manipulating magic values
156 (such as $1) would permanently damage them as perl would
157 ignore the magicalness, by making a full copy of the string,
158 reported by Dmitry Karasik.
159 - work around spurious warnings under older perl 5.8's.
160
161 2.23 Mon Sep 29 05:08:29 CEST 2008
162 - fix a compilation problem when perl is not using char * as, well,
163 char *.
164 - use PL_hexdigit in favour of rolling our own.
165
166 2.2222 Sun Jul 20 18:49:00 CEST 2008
167 - same game again, broken 5.10 finds yet another assertion
168 failure, and the workaround causes additional runtime warnings.
169 Work around the next assertion AND the warning. 5.10 seriously
170 needs to adjust it's attitude against working code.
171
172 2.222 Sat Jul 19 06:15:34 CEST 2008
173 - you work around one -DDEBUGGING assertion bug in perl 5.10
174 just to hit the next one. work around this one, too.
175
176 2.22 Tue Jul 15 13:26:51 CEST 2008
177 - allow higher nesting levels in incremental parser.
178 - error out earlier in some cases in the incremental parser
179 (as suggested by Yuval Kogman).
180 - improve incr-parser test (Yuval Kogman).
181
182 2.21 Tue Jun 3 08:43:23 CEST 2008
183 - (hopefully) work around a perl 5.10 bug with -DDEBUGGING.
184 - remove the experimental status of the incremental parser interface.
185 - move =encoding around again, to avoid bugs with search.cpan.org.
186 when can we finally have utf-8 in pod???
187 - add ->incr_reset method.
188
189 2.2 Wed Apr 16 20:37:25 CEST 2008
190 - lifted the log2 rounding restriction of max_depth and max_size.
191 - make booleans mutable by creating a copy instead of handing out
192 the same scalar (reported by pasha sadri).
193 - added support for incremental json parsing (still EXPERIMENTAL).
194 - implemented and added a json_xs command line utility that can convert
195 from/to a number of serialisation formats - tell me if you need more.
196 - implement allow_unknown/get_allow_unknown methods.
197 - fixed documentation of max_depth w.r.t. higher and equal.
198 - moved down =encoding directive a bit, too much breaks if it's the first
199 pod directive :/.
200 - removed documentation section on other modules, it became somewhat
201 outdated and is nowadays mostly of historical interest.
202
203 2.1 Wed Mar 19 23:23:18 CET 2008
204 - update documentation here and there: add a large section
205 about utf8/latin1/ascii flags, add a security consideration
206 and extend and clarify the JSON and YAML section.
207 - medium speed enhancements when encoding/decoding non-ascii chars.
208 - minor speedup in number encoding case.
209 - extend and clarify the section on incompatibilities
210 between YAML and JSON.
211 - switch to static inline from just inline when using gcc.
212 - add =encoding utf-8 to the manpage, now that perl 5.10 supports it.
213 - fix some issues with UV to JSON conversion of unknown impact.
214 - published the yahoo locals search result used in benchmarks as the
215 original url changes so comparison is impossible.
216
217 2.01 Wed Dec 5 11:40:28 CET 2007
218 - INCOMPATIBLE API CHANGE: to_json and from_json have been
219 renamed to encode_json/decode_json for JSON.pm compatibility.
220 The old functions croak and might be replaced by JSON.pm
221 comaptible versions in some later release.
222
223 2.0 Tue Dec 4 11:30:46 CET 2007
224 - this is supposed to be the first version of JSON::XS
225 compatible with version 2.0+ of the JSON module.
226 Using the JSON module as frontend to JSON::XS should be
227 as fast as using JSON::XS directly, so consider using it
228 instead.
229 - added get_* methods for all "simple" options.
230 - make JSON::XS subclassable.
231
232 1.53 Tue Nov 13 23:58:33 CET 2007
233 - minor doc clarifications.
234 - fixed many doc typos (patch by Thomas L. Shinnick).
235
236 1.52 Mon Oct 15 03:22:06 CEST 2007
237 - remove =encoding pod directive again, it confuses too many pod
238 parsers :/.
239
240 1.51 Sat Oct 13 03:55:56 CEST 2007
241 - encode empty arrays/hashes in a compact way when pretty is enabled.
242 - apparently JSON::XS was used to find some bugs in the
243 JSON_checker testsuite, so add (the corrected) JSON_checker tests to
244 the testsuite.
245 - quite a bit of doc updates/extension.
246 - require 5.8.2, as this seems to be the first unicode-stable version.
247
248 1.5 Tue Aug 28 04:05:38 CEST 2007
249 - add support for tied hashes, based on ideas and testcase by
250 Marcus Holland-Moritz.
251 - implemented relaxed parsing mode where some extensions are being
252 accepted. generation is still JSON-only.
253
254 1.44 Wed Aug 22 01:02:44 CEST 2007
255 - very experimental process-emulation support, slowing everything down.
256 the horribly broken perl threads are still not supported - YMMV.
257
258 1.43 Thu Jul 26 13:26:37 CEST 2007
259 - convert big json numbers exclusively consisting of digits to NV
260 only when there is no loss of precision, otherwise to string.
261
262 1.42 Tue Jul 24 00:51:18 CEST 2007
263 - fix a crash caused by not handling missing array elements
264 (report and testcase by Jay Kuri).
265
266 1.41 Tue Jul 10 18:21:44 CEST 2007
267 - fix compilation with NDEBUG (assert side-effect),
268 affects convert_blessed only.
269 - fix a bug in decode filters calling ENTER; SAVETMPS;
270 one time too often.
271 - catch a typical error in TO_JSON methods.
272 - antique-ised XS.xs again to work with outdated
273 C compilers (windows...).
274
275 1.4 Mon Jul 2 10:06:30 CEST 2007
276 - add convert_blessed setting.
277 - encode did not catch all blessed objects, encoding their
278 contents in most cases. This has been fixed by introducing
279 the allow_blessed setting.
280 - added filter_json_object and filter_json_single_key_object
281 settings that specify a callback to be called when
282 all/specific json objects are encountered.
283 - assume that most object keys are simple ascii words and
284 optimise this case, penalising the general case. This can
285 speed up decoding by 30% in typical cases and gives
286 a smaller and faster perl hash.
287 - implemented simpleminded, optional resource size checking
288 in decode_json.
289 - remove objToJson/jsonToObj aliases, as the next version
290 of JSON will not have them either.
291 - bit the bullet and converted the very simple json object
292 into a more complex one.
293 - work around a bug where perl wrongly claims an integer
294 is not an integer.
295 - unbundle JSON::XS::Boolean into own pm file so Storable
296 and similar modules can resolve the overloading when thawing.
297
298 1.3 Sun Jun 24 01:55:02 CEST 2007
299 - make JSON::XS::true and false special overloaded objects
300 and return those instead of 1 and 0 for those json atoms
301 (JSON::PP compatibility is NOT achieved yet).
302 - add JSON::XS::is_bool predicate to test for those special
303 values.
304 - add a reference to
305 http://jpsykes.com/47/practical-csrf-and-json-security.
306 - removed require 5.8.8 again, it is just not very expert-friendly.
307 Also try to be more compatible with slightly older versions,
308 which are not recommended (because they are buggy).
309
310 1.24 Mon Jun 11 05:40:49 CEST 2007
311 - added informative section on JSON-as-YAML.
312 - get rid of some c99-isms again.
313 - localise dec->cur in decode_str, speeding up
314 string decoding considerably (>15% on my amd64 + gcc).
315 - increased SHORT_STRING_LEN to 16kb: stack space is
316 usually plenty, and this actually saves memory
317 when !shrinking as short strings will fit perfectly.
318
319 1.23 Wed Jun 6 20:13:06 CEST 2007
320 - greatly improved small integer encoding and decoding speed.
321 - implement a number of µ-optimisations.
322 - updated benchmarks.
323
324 1.22 Thu May 24 00:07:25 CEST 2007
325 - require 5.8.8 explicitly as older perls do not seem to offer
326 the required macros.
327 - possibly made it compile on so-called C compilers by microsoft.
328
329 1.21 Wed May 9 18:40:32 CEST 2007
330 - character offset reported for trailing garbage was random.
331
332 1.2 Wed May 9 18:35:01 CEST 2007
333 - decode did not work with magical scalars (doh!).
334 - added latin1 flag to produce JSON texts in the latin1 subset
335 of unicode.
336 - flag trailing garbage as error.
337 - new decode_prefix method that returns the number
338 of characters consumed by a decode.
339 - max octets/char in perls UTF-X is actually 13, not 11,
340 as pointed out by Glenn Linderman.
341 - fixed typoe reported by YAMASHINA Hio.
342
343 1.11 Mon Apr 9 07:05:49 CEST 2007
344 - properly 0-terminate sv's returned by encode to help
345 C libraries that expect that 0 to be there.
346 - partially "port" JSON from C to microsofts fucking broken
347 pseudo-C. They should be burned to the ground for pissing
348 on standards. And I should be stoned for even trying to
349 support this filthy excuse for a c compiler.
350
351 1.1 Wed Apr 4 01:45:00 CEST 2007
352 - clarify documentation (pointed out by Quinn Weaver).
353 - decode_utf8 sometimes did not correctly flag errors,
354 leading to segfaults.
355 - further reduced default nesting depth to 512 due to the test
356 failure by that anonymous "chris" whose e-mail address seems
357 to be impossible to get. Tests on other freebsd systems indicate
358 that this is likely a problem in his/her configuration and not this
359 module.
360 - renamed json => JSON in error messages.
361 - corrected the character offset in some error messages.
362
363 1.01 Sat Mar 31 16:15:40 CEST 2007
364 - do not segfault when from_json/decode gets passed
365 a non-string object (reported by Florian Ragwitz).
366 This has no effect on normal operation.
367
368 1.0 Thu Mar 29 04:43:34 CEST 2007
369 - the long awaited (by me) 1.0 version.
370 - add \0 (JSON::XS::false) and \1 (JSON::XS::true) mappings to JSON
371 true and false.
372 - add some more notes to shrink, as suggested by Alex Efros.
373 - improve testsuite.
374 - halve the default nesting depth limit, to hopefully make it
375 work on Freebsd (unfortunately, the cpan tester did not
376 send me his report, so I cannot ask about the stack limit on fbsd).
377
378 0.8 Mon Mar 26 00:10:48 CEST 2007
379 - fix a memleak when decoding hashes.
380 - export jsonToBj and objToJson as aliases
381 to to_json and from_json, to reduce incompatibilities
382 between JSON/JSON::PC and JSON::XS. (experimental).
383 - implement a maximum nesting depth for both en- and de-coding.
384 - added a security considerations sections.
385
386 0.7 Sun Mar 25 01:46:30 CET 2007
387 - code cleanup.
388 - fix a memory overflow bug when indenting.
389 - pretty-printing now up to 15% faster.
390 - improve decoding speed of strings by
391 up to 50% by specialcasing short strings.
392 - further decoding speedups for strings using
393 lots of \u escapes.
394 - improve utf8 decoding speed for U+80 .. U+7FF.
395
396 0.5 Sat Mar 24 20:41:51 CET 2007
397 - added the UTF-16 encoding example hinted at in previous
398 versions.
399 - minor documentation fixes.
400 - fix a bug in and optimise canonicalising fastpath
401 (reported by Craig Manley).
402 - remove a subtest that breaks with bleadperl (reported
403 by Andreas König).
404
405 0.31 Sat Mar 24 02:14:34 CET 2007
406 - documentation updates.
407 - do some casting to hopefully fix Andreas' problem.
408 - nuke bogus json rpc stuff.
409
410 0.3 Fri Mar 23 19:33:21 CET 2007
411 - remove spurious PApp::Util reference (John McNamara).
412 - adapted lots of tests from other json modules
413 (idea by Chris Carline).
414 - documented mapping from json to perl and vice versa.
415 - improved the documentation by adding more examples.
416 - added short escaping forms, reducing the created
417 json texts a bit.
418 - added shrink flag.
419 - when flag methods are called without enable argument
420 they will by default enable their flag.
421 - considerably improved string encoding speed (at least
422 with gcc 4).
423 - added a test that covers lots of different characters.
424 - clarified some error messages.
425 - error messages now use correct character offset
426 with F_UTF8.
427 - improve the "no bytes" and "no warnings" hacks in
428 case the called functions do... stuff.
429 - croak when encoding to ascii and an out-of-range
430 (non-unicode) codepoint is encountered.
431
432 0.2 Fri Mar 23 00:23:34 CET 2007
433 - the "could not sleep without debugging release".
434 it should basically work now, with many bugs as
435 no production tests have been run yet.
436 - added more testcases.
437 - the expected shitload of bugfixes.
438 - handle utf8 flag correctly in decode.
439 - fix segfault in decoder.
440 - utf8n_to_uvuni sets retlen to -1, but retlen is an
441 unsigned types (argh).
442 - fix decoding of utf-8 strings.
443 - improved error diagnostics.
444 - fix decoding of 'null'.
445 - fix parsing of empty array/hashes
446 - silence warnings when we prepare the croak message.
447
448 0.1 Thu Mar 22 22:13:43 CET 2007
449 - first release, very untested, basically just to claim
450 the namespace.
451
452 0.01 Thu Mar 22 06:08:12 CET 2007
453 - original version; cloned from Convert-Scalar
454