ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/JSON-XS/Changes
(Generate patch)

Comparing JSON-XS/Changes (file contents):
Revision 1.40 by root, Wed May 9 16:35:21 2007 UTC vs.
Revision 1.165 by root, Sun Mar 2 22:09:38 2014 UTC

1Revision history for Perl extension JSON::XS 1Revision history for Perl extension JSON::XS
2 2
3TODO: maybe use some overload magic to decode true and false into something 3TODO: maybe detetc and croak on more invalid inputs (e.g. +-inf/nan)
4 not 0 or 1 in perl. or use attached magic? 4TODO: maybe avoid the reblessing and better support readonly objects.
5TODO: get rid of dynamic arrays to support so-called C compilers 5TODO: http://stevehanov.ca/blog/index.php?id=104 compression
6 by fucking microsoft that do not do C. argl, snprintf is also 6
7 missing. 7TODO: how to cope with tagged values and standard json decoders
8TODO: work around <=5.16 bugs (Reini Urban)
9TODO: decide what to do with RFC7158 incompatibilities.
103.01 Tue Oct 29 16:55:15 CET 2013
11 - backport to perls < 5.18 (reported by Paul Howarth).
12
133.0 Tue Oct 29 01:35:37 CET 2013
14 - implemented an object tagging extension (using the
15 Types::Serialiser serialisation protocol).
16 - reworked the documentation regarding object serialisation,
17 add a new OBJECT SERIALISATION section that explains th
18 whole process.
19 - new setting: allow_tags.
20 - switch to Types::Serialiser booleans.
21 - remove to_json/from_json.
22 - other minor improvements to the documentation.
23
242.34 Thu May 23 11:30:34 CEST 2013
25 - work around bugs in perl 5.18 breaking more than 100
26 widely used modules, without a fix in sight because
27 p5pers don't care about CPAN anymore.
28 - when canonicalising, only allocate up to 64 hash key
29 pointers on the stack. for larger hashes, use the heap,
30 to avoid using too much stackspace.
31 - discuss the problem with setlocale (reported by a few victims).
32
332.33 Wed Aug 1 21:03:52 CEST 2012
34 - internal encode/decode XS wrappers did not expect stack
35 moves caused by callbacks (analyzed and testcase by Jesse Luehrs).
36 - add bencode as to/from option in bin/json_xs.
37 - add -e option to json_xs, and none and string in/out formats.
38
392.32 Thu Aug 11 19:06:38 CEST 2011
40 - fix a bug in the initial whitespace accumulation.
41
422.31 Wed Jul 27 17:53:05 CEST 2011
43 - don't accumulate initial whitespace in the incremental buffer
44 (this can be useful to allow whitespace-keepalive on a tcp
45 connection without triggering the max_size limit).
46 - properly croak on some invalid inputs that are not strings
47 (e.g. undef) when trying to decode a json text (reported
48 and analyzed by Goro Fuji).
49
502.3 Wed Aug 18 01:26:47 CEST 2010
51 - make sure decoder doesn't change the decoding in the incremental
52 parser (testcase provided by Hendrik Schumacher).
53 - applied patch by DaTa for Data::Dumper support in json_xs.
54 - added -t dump support to json_xs, using Data::Dump.
55 - added -f eval support to json_xs.
56
572.29 Wed Mar 17 02:39:12 CET 2010
58 - fix a memory leak when callbacks set using filter_json_object
59 or filter_json_single_key_object were called (great testcase
60 by Eric Wilhelm).
61
622.28 Thu Mar 11 20:30:46 CET 2010
63 - implement our own atof function - perl's can be orders of
64 magnitudes slower than even the system one. on the positive
65 side, ours seems to be more exact in general than perl's.
66 (testcase provided by Tim Meadowcroft).
67 - clarify floating point conversion issues a bit.
68 - update jpsykes csrf article url.
69 - updated benchmark section - JSON::PP became much faster!
70
712.27 Thu Jan 7 07:35:08 CET 2010
72 - support relaxed option inside the incremental parser
73 (testcase provided by IKEGAMI via Makamaka).
74
752.26 Sat Oct 10 03:26:19 CEST 2009
76 - big integers could become truncated (based on patch
77 by Strobl Anton).
78 - output format change: indent now adds a final newline, which is
79 more expected and more true to the documentation.
80
812.25 Sat Aug 8 12:04:41 CEST 2009
82 - the perl debugger completely breaks lvalue subs - try to work
83 around the issue.
84 - ignore RMAGICAL hashes w.r.t. CANONICAL.
85 - try to work around a possible char signedness issue on aix.
86 - require common sense.
87
882.24 Sat May 30 08:25:45 CEST 2009
89 - the incremental parser did not update its parse offset
90 pointer correctly when parsing utf8-strings (nicely
91 debugged by Martin Evans).
92 - appending a non-utf8-string to the incremental parser
93 in utf8 mode failed to upgrade the string.
94 - wording of parse error messages has been improved.
95
962.232 Sun Feb 22 11:12:25 CET 2009
97 - use an exponential algorithm to extend strings, to
98 help platforms with bad or abysmal==windows memory
99 allocater performance, at the expense of some memory
100 wastage (use shrink to recover this extra memory).
101 (nicely analysed by Dmitry Karasik).
102
1032.2311 Thu Feb 19 02:12:54 CET 2009
104 - add a section "JSON and ECMAscript" to explain some
105 incompatibilities between the two (problem was noted by
106 various people).
107 - add t/20_faihu.t.
108
1092.231 Thu Nov 20 04:59:08 CET 2008
110 - work around 5.10.0 magic bugs where manipulating magic values
111 (such as $1) would permanently damage them as perl would
112 ignore the magicalness, by making a full copy of the string,
113 reported by Dmitry Karasik.
114 - work around spurious warnings under older perl 5.8's.
115
1162.23 Mon Sep 29 05:08:29 CEST 2008
117 - fix a compilation problem when perl is not using char * as, well,
118 char *.
119 - use PL_hexdigit in favour of rolling our own.
120
1212.2222 Sun Jul 20 18:49:00 CEST 2008
122 - same game again, broken 5.10 finds yet another assertion
123 failure, and the workaround causes additional runtime warnings.
124 Work around the next assertion AND the warning. 5.10 seriously
125 needs to adjust it's attitude against working code.
126
1272.222 Sat Jul 19 06:15:34 CEST 2008
128 - you work around one -DDEBUGGING assertion bug in perl 5.10
129 just to hit the next one. work around this one, too.
130
1312.22 Tue Jul 15 13:26:51 CEST 2008
132 - allow higher nesting levels in incremental parser.
133 - error out earlier in some cases in the incremental parser
134 (as suggested by Yuval Kogman).
135 - improve incr-parser test (Yuval Kogman).
136
1372.21 Tue Jun 3 08:43:23 CEST 2008
138 - (hopefully) work around a perl 5.10 bug with -DDEBUGGING.
139 - remove the experimental status of the incremental parser interface.
140 - move =encoding around again, to avoid bugs with search.cpan.org.
141 when can we finally have utf-8 in pod???
142 - add ->incr_reset method.
143
1442.2 Wed Apr 16 20:37:25 CEST 2008
145 - lifted the log2 rounding restriction of max_depth and max_size.
146 - make booleans mutable by creating a copy instead of handing out
147 the same scalar (reported by pasha sadri).
148 - added support for incremental json parsing (still EXPERIMENTAL).
149 - implemented and added a json_xs command line utility that can convert
150 from/to a number of serialisation formats - tell me if you need more.
151 - implement allow_unknown/get_allow_unknown methods.
152 - fixed documentation of max_depth w.r.t. higher and equal.
153 - moved down =encoding directive a bit, too much breaks if it's the first
154 pod directive :/.
155 - removed documentation section on other modules, it became somewhat
156 outdated and is nowadays mostly of historical interest.
157
1582.1 Wed Mar 19 23:23:18 CET 2008
159 - update documentation here and there: add a large section
160 about utf8/latin1/ascii flags, add a security consideration
161 and extend and clarify the JSON and YAML section.
162 - medium speed enhancements when encoding/decoding non-ascii chars.
163 - minor speedup in number encoding case.
164 - extend and clarify the section on incompatibilities
165 between YAML and JSON.
166 - switch to static inline from just inline when using gcc.
167 - add =encoding utf-8 to the manpage, now that perl 5.10 supports it.
168 - fix some issues with UV to JSON conversion of unknown impact.
169 - published the yahoo locals search result used in benchmarks as the
170 original url changes so comparison is impossible.
171
1722.01 Wed Dec 5 11:40:28 CET 2007
173 - INCOMPATIBLE API CHANGE: to_json and from_json have been
174 renamed to encode_json/decode_json for JSON.pm compatibility.
175 The old functions croak and might be replaced by JSON.pm
176 comaptible versions in some later release.
177
1782.0 Tue Dec 4 11:30:46 CET 2007
179 - this is supposed to be the first version of JSON::XS
180 compatible with version 2.0+ of the JSON module.
181 Using the JSON module as frontend to JSON::XS should be
182 as fast as using JSON::XS directly, so consider using it
183 instead.
184 - added get_* methods for all "simple" options.
185 - make JSON::XS subclassable.
186
1871.53 Tue Nov 13 23:58:33 CET 2007
188 - minor doc clarifications.
189 - fixed many doc typos (patch by Thomas L. Shinnick).
190
1911.52 Mon Oct 15 03:22:06 CEST 2007
192 - remove =encoding pod directive again, it confuses too many pod
193 parsers :/.
194
1951.51 Sat Oct 13 03:55:56 CEST 2007
196 - encode empty arrays/hashes in a compact way when pretty is enabled.
197 - apparently JSON::XS was used to find some bugs in the
198 JSON_checker testsuite, so add (the corrected) JSON_checker tests to
199 the testsuite.
200 - quite a bit of doc updates/extension.
201 - require 5.8.2, as this seems to be the first unicode-stable version.
202
2031.5 Tue Aug 28 04:05:38 CEST 2007
204 - add support for tied hashes, based on ideas and testcase by
205 Marcus Holland-Moritz.
206 - implemented relaxed parsing mode where some extensions are being
207 accepted. generation is still JSON-only.
208
2091.44 Wed Aug 22 01:02:44 CEST 2007
210 - very experimental process-emulation support, slowing everything down.
211 the horribly broken perl threads are still not supported - YMMV.
212
2131.43 Thu Jul 26 13:26:37 CEST 2007
214 - convert big json numbers exclusively consisting of digits to NV
215 only when there is no loss of precision, otherwise to string.
216
2171.42 Tue Jul 24 00:51:18 CEST 2007
218 - fix a crash caused by not handling missing array elements
219 (report and testcase by Jay Kuri).
220
2211.41 Tue Jul 10 18:21:44 CEST 2007
222 - fix compilation with NDEBUG (assert side-effect),
223 affects convert_blessed only.
224 - fix a bug in decode filters calling ENTER; SAVETMPS;
225 one time too often.
226 - catch a typical error in TO_JSON methods.
227 - antique-ised XS.xs again to work with outdated
228 C compilers (windows...).
229
2301.4 Mon Jul 2 10:06:30 CEST 2007
231 - add convert_blessed setting.
232 - encode did not catch all blessed objects, encoding their
233 contents in most cases. This has been fixed by introducing
234 the allow_blessed setting.
235 - added filter_json_object and filter_json_single_key_object
236 settings that specify a callback to be called when
237 all/specific json objects are encountered.
238 - assume that most object keys are simple ascii words and
239 optimise this case, penalising the general case. This can
240 speed up decoding by 30% in typical cases and gives
241 a smaller and faster perl hash.
242 - implemented simpleminded, optional resource size checking
243 in decode_json.
244 - remove objToJson/jsonToObj aliases, as the next version
245 of JSON will not have them either.
246 - bit the bullet and converted the very simple json object
247 into a more complex one.
248 - work around a bug where perl wrongly claims an integer
249 is not an integer.
250 - unbundle JSON::XS::Boolean into own pm file so Storable
251 and similar modules can resolve the overloading when thawing.
252
2531.3 Sun Jun 24 01:55:02 CEST 2007
254 - make JSON::XS::true and false special overloaded objects
255 and return those instead of 1 and 0 for those json atoms
256 (JSON::PP compatibility is NOT achieved yet).
257 - add JSON::XS::is_bool predicate to test for those special
258 values.
259 - add a reference to
260 http://jpsykes.com/47/practical-csrf-and-json-security.
261 - removed require 5.8.8 again, it is just not very expert-friendly.
262 Also try to be more compatible with slightly older versions,
263 which are not recommended (because they are buggy).
264
2651.24 Mon Jun 11 05:40:49 CEST 2007
266 - added informative section on JSON-as-YAML.
267 - get rid of some c99-isms again.
268 - localise dec->cur in decode_str, speeding up
269 string decoding considerably (>15% on my amd64 + gcc).
270 - increased SHORT_STRING_LEN to 16kb: stack space is
271 usually plenty, and this actually saves memory
272 when !shrinking as short strings will fit perfectly.
273
2741.23 Wed Jun 6 20:13:06 CEST 2007
275 - greatly improved small integer encoding and decoding speed.
276 - implement a number of ยต-optimisations.
277 - updated benchmarks.
278
2791.22 Thu May 24 00:07:25 CEST 2007
280 - require 5.8.8 explicitly as older perls do not seem to offer
281 the required macros.
282 - possibly made it compile on so-called C compilers by microsoft.
283
2841.21 Wed May 9 18:40:32 CEST 2007
285 - character offset reported for trailing garbage was random.
8 286
91.2 Wed May 9 18:35:01 CEST 2007 2871.2 Wed May 9 18:35:01 CEST 2007
10 - decode did not work with magical scalars (doh!). 288 - decode did not work with magical scalars (doh!).
11 - added latin1 flag to produce JSON texts in the latin1 subset 289 - added latin1 flag to produce JSON texts in the latin1 subset
12 of unicode. 290 of unicode.
… …
105 case the called functions do... stuff. 383 case the called functions do... stuff.
106 - croak when encoding to ascii and an out-of-range 384 - croak when encoding to ascii and an out-of-range
107 (non-unicode) codepoint is encountered. 385 (non-unicode) codepoint is encountered.
108 386
1090.2 Fri Mar 23 00:23:34 CET 2007 3870.2 Fri Mar 23 00:23:34 CET 2007
110 - the "could not sleep without debuggign release". 388 - the "could not sleep without debugging release".
111 it should basically work now, with many bugs as 389 it should basically work now, with many bugs as
112 no production tests have been run yet. 390 no production tests have been run yet.
113 - added more testcases. 391 - added more testcases.
114 - the expected shitload of bugfixes. 392 - the expected shitload of bugfixes.
115 - handle utf8 flag correctly in decode. 393 - handle utf8 flag correctly in decode.

Diff Legend

– Removed lines
+ Added lines
< Changed lines
> Changed lines