ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/JSON-XS/Changes
Revision: 1.195
Committed: Sun Jan 26 04:52:57 2025 UTC (3 months ago) by root
Branch: MAIN
CVS Tags: HEAD
Changes since 1.194: +12 -0 lines
Log Message:
*** empty log message ***

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 TODO: replace bool_stash by BOOL_STASH seems to work with mod_perl, make a compile time option?
9
10 TODO: validate_utf8, apparewntly some people confuse themselves.
11 TODO: security considerations
12
13 TODO: perl 5.36 has builtin::true/false/is_bool. Not sure how to integrate those.
14 TODO: Peter Juhasz <peter.juhasz@comnica.com>: '11111111111111111111111e1111111111111111111' is parsed as 0
15 # copy freed scalart()??)
16 <config.json runIP-radar eval '
17 use JSON::XS;
18 local $/;
19 my $cfg = decode_json scalar <>;
20 $cfg->{login_background} = Types::Serialiser::as_bool $ARGV[0] eq "";
21 print JSON::XS->new->pretty->utf8->encode ($cfg);
22 ' "$LBUI_ENABLE"
23
24
25 4.03 Tue Oct 27 19:05:01 CET 2020
26 - when parsing comments in relaxed mode, JSON::XS would detect garbage
27 after the JSON text if the comment is after the end and does not end in
28 a newline (reported by Felipe Gasper).
29
30 4.02 Wed Mar 6 08:31:24 CET 2019
31 - undo the fix from 4.01, it breaks more things than it fixes
32 (another testcase by Wesley Schwengle).
33 - try a proper fix this time.
34
35 4.01 Sun Feb 24 05:03:30 CET 2019
36 - fix some stack corruption caused mostly when calling methods
37 in list context (testcase by Wesley Schwengle).
38
39 4.0 Fri Nov 16 00:06:54 CET 2018
40 - SECURITY IMPLICATION: this release enables allow_nonref by default
41 for compatibility with RFC 7159 and newer. See "old" vs. "new"
42 JSON under SECURITY CONSIDERATIONS.
43 - reworked the "old" vs. "new" JSON section.
44 - add ->boolean_values to provide the values to which booleans
45 decode (requested by Aristotle Pagaltzis).
46 - decode would wrongly accept ASCII NUL characters instead of
47 reporting them as trailing garbage.
48 - work around what smells like a perl bug w.r.t. exceptions
49 thrown in callbacks.
50 - incremental parser now more or less respects allow_nonref.
51 - json_xs json-pretty now enables canonical mode.
52 - add documentation section about I-JSON.
53 - minor documentation fixes/updates.
54
55 3.04 Thu Aug 17 04:30:47 CEST 2017
56 - change exponential realloc algorithm on encoding and string decoding to be
57 really exponential (this helps slow/debugging allocators such as libumem)
58 (reported by Matthew Horsfall).
59 - string encoding would needlessly overallocate output space
60 (testcase by Matthew Horsfall).
61 - be very paranoid about extending buffer lengths and croak if buffers get too large,
62 which might (or might not) improve security.
63 - add cbor-packed type to json_xs.
64 - switch from YAML to YAML::XS in json_xs, as YAML is way too buggy and outdated.
65
66 3.03 Wed Nov 16 20:20:59 CET 2016
67 - fix a bug introduced by a perl bug workaround that would cause
68 incremental parsing to fail with a sv_chop panic.
69 - json_xs: toformat failure error message fixed.
70 - json_xs: allow cyclic data structures in CBOR.
71
72 3.02 Fri Feb 26 22:45:20 CET 2016
73 - allow_nonref now affects booleans (\1, $Types::Serialiser::Boolean)
74 as well (reported by Alex Efros).
75 - allow literal tabs in strings in relaxed mode (patch by
76 lubo.rintel@gooddata.com).
77 - support "cbor" format in json_xs tool.
78 - support (and fix) calling encode and decode in list context
79 (reported by Вадим Власов).
80 - work around a bug in older perls crashing when presented
81 with shared hash keys (Reini Urban).
82 - use stability canary.
83
84 3.01 Tue Oct 29 16:55:15 CET 2013
85 - backport to perls < 5.18 (reported by Paul Howarth).
86
87 3.0 Tue Oct 29 01:35:37 CET 2013
88 - implemented an object tagging extension (using the
89 Types::Serialiser serialisation protocol).
90 - reworked the documentation regarding object serialisation,
91 add a new OBJECT SERIALISATION section that explains th
92 whole process.
93 - new setting: allow_tags.
94 - switch to Types::Serialiser booleans.
95 - remove to_json/from_json.
96 - other minor improvements to the documentation.
97
98 2.34 Thu May 23 11:30:34 CEST 2013
99 - work around bugs in perl 5.18 breaking more than 100
100 widely used modules, without a fix in sight because
101 p5pers don't care about CPAN anymore.
102 - when canonicalising, only allocate up to 64 hash key
103 pointers on the stack. for larger hashes, use the heap,
104 to avoid using too much stackspace.
105 - discuss the problem with setlocale (reported by a few victims).
106
107 2.33 Wed Aug 1 21:03:52 CEST 2012
108 - internal encode/decode XS wrappers did not expect stack
109 moves caused by callbacks (analyzed and testcase by Jesse Luehrs).
110 - add bencode as to/from option in bin/json_xs.
111 - add -e option to json_xs, and none and string in/out formats.
112
113 2.32 Thu Aug 11 19:06:38 CEST 2011
114 - fix a bug in the initial whitespace accumulation.
115
116 2.31 Wed Jul 27 17:53:05 CEST 2011
117 - don't accumulate initial whitespace in the incremental buffer
118 (this can be useful to allow whitespace-keepalive on a tcp
119 connection without triggering the max_size limit).
120 - properly croak on some invalid inputs that are not strings
121 (e.g. undef) when trying to decode a json text (reported
122 and analyzed by Goro Fuji).
123
124 2.3 Wed Aug 18 01:26:47 CEST 2010
125 - make sure decoder doesn't change the decoding in the incremental
126 parser (testcase provided by Hendrik Schumacher).
127 - applied patch by DaTa for Data::Dumper support in json_xs.
128 - added -t dump support to json_xs, using Data::Dump.
129 - added -f eval support to json_xs.
130
131 2.29 Wed Mar 17 02:39:12 CET 2010
132 - fix a memory leak when callbacks set using filter_json_object
133 or filter_json_single_key_object were called (great testcase
134 by Eric Wilhelm).
135
136 2.28 Thu Mar 11 20:30:46 CET 2010
137 - implement our own atof function - perl's can be orders of
138 magnitudes slower than even the system one. on the positive
139 side, ours seems to be more exact in general than perl's.
140 (testcase provided by Tim Meadowcroft).
141 - clarify floating point conversion issues a bit.
142 - update jpsykes csrf article url.
143 - updated benchmark section - JSON::PP became much faster!
144
145 2.27 Thu Jan 7 07:35:08 CET 2010
146 - support relaxed option inside the incremental parser
147 (testcase provided by IKEGAMI via Makamaka).
148
149 2.26 Sat Oct 10 03:26:19 CEST 2009
150 - big integers could become truncated (based on patch
151 by Strobl Anton).
152 - output format change: indent now adds a final newline, which is
153 more expected and more true to the documentation.
154
155 2.25 Sat Aug 8 12:04:41 CEST 2009
156 - the perl debugger completely breaks lvalue subs - try to work
157 around the issue.
158 - ignore RMAGICAL hashes w.r.t. CANONICAL.
159 - try to work around a possible char signedness issue on aix.
160 - require common sense.
161
162 2.24 Sat May 30 08:25:45 CEST 2009
163 - the incremental parser did not update its parse offset
164 pointer correctly when parsing utf8-strings (nicely
165 debugged by Martin Evans).
166 - appending a non-utf8-string to the incremental parser
167 in utf8 mode failed to upgrade the string.
168 - wording of parse error messages has been improved.
169
170 2.232 Sun Feb 22 11:12:25 CET 2009
171 - use an exponential algorithm to extend strings, to
172 help platforms with bad or abysmal==windows memory
173 allocater performance, at the expense of some memory
174 wastage (use shrink to recover this extra memory).
175 (nicely analysed by Dmitry Karasik).
176
177 2.2311 Thu Feb 19 02:12:54 CET 2009
178 - add a section "JSON and ECMAscript" to explain some
179 incompatibilities between the two (problem was noted by
180 various people).
181 - add t/20_faihu.t.
182
183 2.231 Thu Nov 20 04:59:08 CET 2008
184 - work around 5.10.0 magic bugs where manipulating magic values
185 (such as $1) would permanently damage them as perl would
186 ignore the magicalness, by making a full copy of the string,
187 reported by Dmitry Karasik.
188 - work around spurious warnings under older perl 5.8's.
189
190 2.23 Mon Sep 29 05:08:29 CEST 2008
191 - fix a compilation problem when perl is not using char * as, well,
192 char *.
193 - use PL_hexdigit in favour of rolling our own.
194
195 2.2222 Sun Jul 20 18:49:00 CEST 2008
196 - same game again, broken 5.10 finds yet another assertion
197 failure, and the workaround causes additional runtime warnings.
198 Work around the next assertion AND the warning. 5.10 seriously
199 needs to adjust it's attitude against working code.
200
201 2.222 Sat Jul 19 06:15:34 CEST 2008
202 - you work around one -DDEBUGGING assertion bug in perl 5.10
203 just to hit the next one. work around this one, too.
204
205 2.22 Tue Jul 15 13:26:51 CEST 2008
206 - allow higher nesting levels in incremental parser.
207 - error out earlier in some cases in the incremental parser
208 (as suggested by Yuval Kogman).
209 - improve incr-parser test (Yuval Kogman).
210
211 2.21 Tue Jun 3 08:43:23 CEST 2008
212 - (hopefully) work around a perl 5.10 bug with -DDEBUGGING.
213 - remove the experimental status of the incremental parser interface.
214 - move =encoding around again, to avoid bugs with search.cpan.org.
215 when can we finally have utf-8 in pod???
216 - add ->incr_reset method.
217
218 2.2 Wed Apr 16 20:37:25 CEST 2008
219 - lifted the log2 rounding restriction of max_depth and max_size.
220 - make booleans mutable by creating a copy instead of handing out
221 the same scalar (reported by pasha sadri).
222 - added support for incremental json parsing (still EXPERIMENTAL).
223 - implemented and added a json_xs command line utility that can convert
224 from/to a number of serialisation formats - tell me if you need more.
225 - implement allow_unknown/get_allow_unknown methods.
226 - fixed documentation of max_depth w.r.t. higher and equal.
227 - moved down =encoding directive a bit, too much breaks if it's the first
228 pod directive :/.
229 - removed documentation section on other modules, it became somewhat
230 outdated and is nowadays mostly of historical interest.
231
232 2.1 Wed Mar 19 23:23:18 CET 2008
233 - update documentation here and there: add a large section
234 about utf8/latin1/ascii flags, add a security consideration
235 and extend and clarify the JSON and YAML section.
236 - medium speed enhancements when encoding/decoding non-ascii chars.
237 - minor speedup in number encoding case.
238 - extend and clarify the section on incompatibilities
239 between YAML and JSON.
240 - switch to static inline from just inline when using gcc.
241 - add =encoding utf-8 to the manpage, now that perl 5.10 supports it.
242 - fix some issues with UV to JSON conversion of unknown impact.
243 - published the yahoo locals search result used in benchmarks as the
244 original url changes so comparison is impossible.
245
246 2.01 Wed Dec 5 11:40:28 CET 2007
247 - INCOMPATIBLE API CHANGE: to_json and from_json have been
248 renamed to encode_json/decode_json for JSON.pm compatibility.
249 The old functions croak and might be replaced by JSON.pm
250 comaptible versions in some later release.
251
252 2.0 Tue Dec 4 11:30:46 CET 2007
253 - this is supposed to be the first version of JSON::XS
254 compatible with version 2.0+ of the JSON module.
255 Using the JSON module as frontend to JSON::XS should be
256 as fast as using JSON::XS directly, so consider using it
257 instead.
258 - added get_* methods for all "simple" options.
259 - make JSON::XS subclassable.
260
261 1.53 Tue Nov 13 23:58:33 CET 2007
262 - minor doc clarifications.
263 - fixed many doc typos (patch by Thomas L. Shinnick).
264
265 1.52 Mon Oct 15 03:22:06 CEST 2007
266 - remove =encoding pod directive again, it confuses too many pod
267 parsers :/.
268
269 1.51 Sat Oct 13 03:55:56 CEST 2007
270 - encode empty arrays/hashes in a compact way when pretty is enabled.
271 - apparently JSON::XS was used to find some bugs in the
272 JSON_checker testsuite, so add (the corrected) JSON_checker tests to
273 the testsuite.
274 - quite a bit of doc updates/extension.
275 - require 5.8.2, as this seems to be the first unicode-stable version.
276
277 1.5 Tue Aug 28 04:05:38 CEST 2007
278 - add support for tied hashes, based on ideas and testcase by
279 Marcus Holland-Moritz.
280 - implemented relaxed parsing mode where some extensions are being
281 accepted. generation is still JSON-only.
282
283 1.44 Wed Aug 22 01:02:44 CEST 2007
284 - very experimental process-emulation support, slowing everything down.
285 the horribly broken perl threads are still not supported - YMMV.
286
287 1.43 Thu Jul 26 13:26:37 CEST 2007
288 - convert big json numbers exclusively consisting of digits to NV
289 only when there is no loss of precision, otherwise to string.
290
291 1.42 Tue Jul 24 00:51:18 CEST 2007
292 - fix a crash caused by not handling missing array elements
293 (report and testcase by Jay Kuri).
294
295 1.41 Tue Jul 10 18:21:44 CEST 2007
296 - fix compilation with NDEBUG (assert side-effect),
297 affects convert_blessed only.
298 - fix a bug in decode filters calling ENTER; SAVETMPS;
299 one time too often.
300 - catch a typical error in TO_JSON methods.
301 - antique-ised XS.xs again to work with outdated
302 C compilers (windows...).
303
304 1.4 Mon Jul 2 10:06:30 CEST 2007
305 - add convert_blessed setting.
306 - encode did not catch all blessed objects, encoding their
307 contents in most cases. This has been fixed by introducing
308 the allow_blessed setting.
309 - added filter_json_object and filter_json_single_key_object
310 settings that specify a callback to be called when
311 all/specific json objects are encountered.
312 - assume that most object keys are simple ascii words and
313 optimise this case, penalising the general case. This can
314 speed up decoding by 30% in typical cases and gives
315 a smaller and faster perl hash.
316 - implemented simpleminded, optional resource size checking
317 in decode_json.
318 - remove objToJson/jsonToObj aliases, as the next version
319 of JSON will not have them either.
320 - bit the bullet and converted the very simple json object
321 into a more complex one.
322 - work around a bug where perl wrongly claims an integer
323 is not an integer.
324 - unbundle JSON::XS::Boolean into own pm file so Storable
325 and similar modules can resolve the overloading when thawing.
326
327 1.3 Sun Jun 24 01:55:02 CEST 2007
328 - make JSON::XS::true and false special overloaded objects
329 and return those instead of 1 and 0 for those json atoms
330 (JSON::PP compatibility is NOT achieved yet).
331 - add JSON::XS::is_bool predicate to test for those special
332 values.
333 - add a reference to
334 http://jpsykes.com/47/practical-csrf-and-json-security.
335 - removed require 5.8.8 again, it is just not very expert-friendly.
336 Also try to be more compatible with slightly older versions,
337 which are not recommended (because they are buggy).
338
339 1.24 Mon Jun 11 05:40:49 CEST 2007
340 - added informative section on JSON-as-YAML.
341 - get rid of some c99-isms again.
342 - localise dec->cur in decode_str, speeding up
343 string decoding considerably (>15% on my amd64 + gcc).
344 - increased SHORT_STRING_LEN to 16kb: stack space is
345 usually plenty, and this actually saves memory
346 when !shrinking as short strings will fit perfectly.
347
348 1.23 Wed Jun 6 20:13:06 CEST 2007
349 - greatly improved small integer encoding and decoding speed.
350 - implement a number of µ-optimisations.
351 - updated benchmarks.
352
353 1.22 Thu May 24 00:07:25 CEST 2007
354 - require 5.8.8 explicitly as older perls do not seem to offer
355 the required macros.
356 - possibly made it compile on so-called C compilers by microsoft.
357
358 1.21 Wed May 9 18:40:32 CEST 2007
359 - character offset reported for trailing garbage was random.
360
361 1.2 Wed May 9 18:35:01 CEST 2007
362 - decode did not work with magical scalars (doh!).
363 - added latin1 flag to produce JSON texts in the latin1 subset
364 of unicode.
365 - flag trailing garbage as error.
366 - new decode_prefix method that returns the number
367 of characters consumed by a decode.
368 - max octets/char in perls UTF-X is actually 13, not 11,
369 as pointed out by Glenn Linderman.
370 - fixed typoe reported by YAMASHINA Hio.
371
372 1.11 Mon Apr 9 07:05:49 CEST 2007
373 - properly 0-terminate sv's returned by encode to help
374 C libraries that expect that 0 to be there.
375 - partially "port" JSON from C to microsofts fucking broken
376 pseudo-C. They should be burned to the ground for pissing
377 on standards. And I should be stoned for even trying to
378 support this filthy excuse for a c compiler.
379
380 1.1 Wed Apr 4 01:45:00 CEST 2007
381 - clarify documentation (pointed out by Quinn Weaver).
382 - decode_utf8 sometimes did not correctly flag errors,
383 leading to segfaults.
384 - further reduced default nesting depth to 512 due to the test
385 failure by that anonymous "chris" whose e-mail address seems
386 to be impossible to get. Tests on other freebsd systems indicate
387 that this is likely a problem in his/her configuration and not this
388 module.
389 - renamed json => JSON in error messages.
390 - corrected the character offset in some error messages.
391
392 1.01 Sat Mar 31 16:15:40 CEST 2007
393 - do not segfault when from_json/decode gets passed
394 a non-string object (reported by Florian Ragwitz).
395 This has no effect on normal operation.
396
397 1.0 Thu Mar 29 04:43:34 CEST 2007
398 - the long awaited (by me) 1.0 version.
399 - add \0 (JSON::XS::false) and \1 (JSON::XS::true) mappings to JSON
400 true and false.
401 - add some more notes to shrink, as suggested by Alex Efros.
402 - improve testsuite.
403 - halve the default nesting depth limit, to hopefully make it
404 work on Freebsd (unfortunately, the cpan tester did not
405 send me his report, so I cannot ask about the stack limit on fbsd).
406
407 0.8 Mon Mar 26 00:10:48 CEST 2007
408 - fix a memleak when decoding hashes.
409 - export jsonToBj and objToJson as aliases
410 to to_json and from_json, to reduce incompatibilities
411 between JSON/JSON::PC and JSON::XS. (experimental).
412 - implement a maximum nesting depth for both en- and de-coding.
413 - added a security considerations sections.
414
415 0.7 Sun Mar 25 01:46:30 CET 2007
416 - code cleanup.
417 - fix a memory overflow bug when indenting.
418 - pretty-printing now up to 15% faster.
419 - improve decoding speed of strings by
420 up to 50% by specialcasing short strings.
421 - further decoding speedups for strings using
422 lots of \u escapes.
423 - improve utf8 decoding speed for U+80 .. U+7FF.
424
425 0.5 Sat Mar 24 20:41:51 CET 2007
426 - added the UTF-16 encoding example hinted at in previous
427 versions.
428 - minor documentation fixes.
429 - fix a bug in and optimise canonicalising fastpath
430 (reported by Craig Manley).
431 - remove a subtest that breaks with bleadperl (reported
432 by Andreas König).
433
434 0.31 Sat Mar 24 02:14:34 CET 2007
435 - documentation updates.
436 - do some casting to hopefully fix Andreas' problem.
437 - nuke bogus json rpc stuff.
438
439 0.3 Fri Mar 23 19:33:21 CET 2007
440 - remove spurious PApp::Util reference (John McNamara).
441 - adapted lots of tests from other json modules
442 (idea by Chris Carline).
443 - documented mapping from json to perl and vice versa.
444 - improved the documentation by adding more examples.
445 - added short escaping forms, reducing the created
446 json texts a bit.
447 - added shrink flag.
448 - when flag methods are called without enable argument
449 they will by default enable their flag.
450 - considerably improved string encoding speed (at least
451 with gcc 4).
452 - added a test that covers lots of different characters.
453 - clarified some error messages.
454 - error messages now use correct character offset
455 with F_UTF8.
456 - improve the "no bytes" and "no warnings" hacks in
457 case the called functions do... stuff.
458 - croak when encoding to ascii and an out-of-range
459 (non-unicode) codepoint is encountered.
460
461 0.2 Fri Mar 23 00:23:34 CET 2007
462 - the "could not sleep without debugging release".
463 it should basically work now, with many bugs as
464 no production tests have been run yet.
465 - added more testcases.
466 - the expected shitload of bugfixes.
467 - handle utf8 flag correctly in decode.
468 - fix segfault in decoder.
469 - utf8n_to_uvuni sets retlen to -1, but retlen is an
470 unsigned types (argh).
471 - fix decoding of utf-8 strings.
472 - improved error diagnostics.
473 - fix decoding of 'null'.
474 - fix parsing of empty array/hashes
475 - silence warnings when we prepare the croak message.
476
477 0.1 Thu Mar 22 22:13:43 CET 2007
478 - first release, very untested, basically just to claim
479 the namespace.
480
481 0.01 Thu Mar 22 06:08:12 CET 2007
482 - original version; cloned from Convert-Scalar
483