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

Comparing JSON-XS/Changes (file contents):
Revision 1.17 by root, Sat Mar 24 22:55:16 2007 UTC vs.
Revision 1.144 by root, Fri Aug 13 21:00:02 2010 UTC

1Revision history for Perl extension JSON::XS 1Revision history for Perl extension JSON::XS
2 2
3TODO: allow a configurable recursion limit and/or or other forms of resource limiting (string size, memory usage...) 3 - applied patch by DaTa for Data::Dumper support in json_xs.
4TODO: maybe simply use \0 and \1 as true/false when encoding (and overloaded/blessed versions thereof on 4 - added dump output method to json_xs.
5 decoding?)
6 5
62.29 Wed Mar 17 02:39:12 CET 2010
7 - fix a memory leak when callbacks set using filter_json_object
8 or filter_json_single_key_object were called (great testcase
9 by Eric Wilhelm).
10
112.28 Thu Mar 11 20:30:46 CET 2010
12 - implement our own atof function - perl's can be orders of
13 magnitudes slower than even the system one. on the positive
14 side, ours seems to be more exact in general than perl's.
15 (testcase provided by Tim Meadowcroft).
16 - clarify floating point conversion issues a bit.
17 - update jpsykes csrf article url.
18 - updated benchmark section - JSON::PP became much faster!
19
202.27 Thu Jan 7 07:35:08 CET 2010
21 - support relaxed option inside the incremental parser
22 (testcase provided by IKEGAMI via Makamaka).
23
242.26 Sat Oct 10 03:26:19 CEST 2009
25 - big integers could become truncated (based on patch
26 by Strobl Anton).
27 - output format change: indent now adds a final newline, which is
28 more expected and more true to the documentation.
29
302.25 Sat Aug 8 12:04:41 CEST 2009
31 - the perl debugger completely breaks lvalue subs - try to work
32 around the issue.
33 - ignore RMAGICAL hashes w.r.t. CANONICAL.
34 - try to work around a possible char signedness issue on aix.
35 - require common sense.
36
372.24 Sat May 30 08:25:45 CEST 2009
38 - the incremental parser did not update its parse offset
39 pointer correctly when parsing utf8-strings (nicely
40 debugged by Martin Evans).
41 - appending a non-utf8-string to the incremental parser
42 in utf8 mode failed to upgrade the string.
43 - wording of parse error messages has been improved.
44
452.232 Sun Feb 22 11:12:25 CET 2009
46 - use an exponential algorithm to extend strings, to
47 help platforms with bad or abysmal==windows memory
48 allocater performance, at the expense of some memory
49 wastage (use shrink to recover this extra memory).
50 (nicely analysed by Dmitry Karasik).
51
522.2311 Thu Feb 19 02:12:54 CET 2009
53 - add a section "JSON and ECMAscript" to explain some
54 incompatibilities between the two (problem was noted by
55 various people).
56 - add t/20_faihu.t.
57
582.231 Thu Nov 20 04:59:08 CET 2008
59 - work around 5.10.0 magic bugs where manipulating magic values
60 (such as $1) would permanently damage them as perl would
61 ignore the magicalness, by making a full copy of the string,
62 reported by Dmitry Karasik.
63 - work around spurious warnings under older perl 5.8's.
64
652.23 Mon Sep 29 05:08:29 CEST 2008
66 - fix a compilation problem when perl is not using char * as, well,
67 char *.
68 - use PL_hexdigit in favour of rolling our own.
69
702.2222 Sun Jul 20 18:49:00 CEST 2008
71 - same game again, broken 5.10 finds yet another assertion
72 failure, and the workaround causes additional runtime warnings.
73 Work around the next assertion AND the warning. 5.10 seriously
74 needs to adjust it's attitude against working code.
75
762.222 Sat Jul 19 06:15:34 CEST 2008
77 - you work around one -DDEBUGGING assertion bug in perl 5.10
78 just to hit the next one. work around this one, too.
79
802.22 Tue Jul 15 13:26:51 CEST 2008
81 - allow higher nesting levels in incremental parser.
82 - error out earlier in some cases in the incremental parser
83 (as suggested by Yuval Kogman).
84 - improve incr-parser test (Yuval Kogman).
85
862.21 Tue Jun 3 08:43:23 CEST 2008
87 - (hopefully) work around a perl 5.10 bug with -DDEBUGGING.
88 - remove the experimental status of the incremental parser interface.
89 - move =encoding around again, to avoid bugs with search.cpan.org.
90 when can we finally have utf-8 in pod???
91 - add ->incr_reset method.
92
932.2 Wed Apr 16 20:37:25 CEST 2008
94 - lifted the log2 rounding restriction of max_depth and max_size.
95 - make booleans mutable by creating a copy instead of handing out
96 the same scalar (reported by pasha sadri).
97 - added support for incremental json parsing (still EXPERIMENTAL).
98 - implemented and added a json_xs command line utility that can convert
99 from/to a number of serialisation formats - tell me if you need more.
100 - implement allow_unknown/get_allow_unknown methods.
101 - fixed documentation of max_depth w.r.t. higher and equal.
102 - moved down =encoding directive a bit, too much breaks if it's the first
103 pod directive :/.
104 - removed documentation section on other modules, it became somewhat
105 outdated and is nowadays mostly of historical interest.
106
1072.1 Wed Mar 19 23:23:18 CET 2008
108 - update documentation here and there: add a large section
109 about utf8/latin1/ascii flags, add a security consideration
110 and extend and clarify the JSON and YAML section.
111 - medium speed enhancements when encoding/decoding non-ascii chars.
112 - minor speedup in number encoding case.
113 - extend and clarify the section on incompatibilities
114 between YAML and JSON.
115 - switch to static inline from just inline when using gcc.
116 - add =encoding utf-8 to the manpage, now that perl 5.10 supports it.
117 - fix some issues with UV to JSON conversion of unknown impact.
118 - published the yahoo locals search result used in benchmarks as the
119 original url changes so comparison is impossible.
120
1212.01 Wed Dec 5 11:40:28 CET 2007
122 - INCOMPATIBLE API CHANGE: to_json and from_json have been
123 renamed to encode_json/decode_json for JSON.pm compatibility.
124 The old functions croak and might be replaced by JSON.pm
125 comaptible versions in some later release.
126
1272.0 Tue Dec 4 11:30:46 CET 2007
128 - this is supposed to be the first version of JSON::XS
129 compatible with version 2.0+ of the JSON module.
130 Using the JSON module as frontend to JSON::XS should be
131 as fast as using JSON::XS directly, so consider using it
132 instead.
133 - added get_* methods for all "simple" options.
134 - make JSON::XS subclassable.
135
1361.53 Tue Nov 13 23:58:33 CET 2007
137 - minor doc clarifications.
138 - fixed many doc typos (patch by Thomas L. Shinnick).
139
1401.52 Mon Oct 15 03:22:06 CEST 2007
141 - remove =encoding pod directive again, it confuses too many pod
142 parsers :/.
143
1441.51 Sat Oct 13 03:55:56 CEST 2007
145 - encode empty arrays/hashes in a compact way when pretty is enabled.
146 - apparently JSON::XS was used to find some bugs in the
147 JSON_checker testsuite, so add (the corrected) JSON_checker tests to
148 the testsuite.
149 - quite a bit of doc updates/extension.
150 - require 5.8.2, as this seems to be the first unicode-stable version.
151
1521.5 Tue Aug 28 04:05:38 CEST 2007
153 - add support for tied hashes, based on ideas and testcase by
154 Marcus Holland-Moritz.
155 - implemented relaxed parsing mode where some extensions are being
156 accepted. generation is still JSON-only.
157
1581.44 Wed Aug 22 01:02:44 CEST 2007
159 - very experimental process-emulation support, slowing everything down.
160 the horribly broken perl threads are still not supported - YMMV.
161
1621.43 Thu Jul 26 13:26:37 CEST 2007
163 - convert big json numbers exclusively consisting of digits to NV
164 only when there is no loss of precision, otherwise to string.
165
1661.42 Tue Jul 24 00:51:18 CEST 2007
167 - fix a crash caused by not handling missing array elements
168 (report and testcase by Jay Kuri).
169
1701.41 Tue Jul 10 18:21:44 CEST 2007
171 - fix compilation with NDEBUG (assert side-effect),
172 affects convert_blessed only.
173 - fix a bug in decode filters calling ENTER; SAVETMPS;
174 one time too often.
175 - catch a typical error in TO_JSON methods.
176 - antique-ised XS.xs again to work with outdated
177 C compilers (windows...).
178
1791.4 Mon Jul 2 10:06:30 CEST 2007
180 - add convert_blessed setting.
181 - encode did not catch all blessed objects, encoding their
182 contents in most cases. This has been fixed by introducing
183 the allow_blessed setting.
184 - added filter_json_object and filter_json_single_key_object
185 settings that specify a callback to be called when
186 all/specific json objects are encountered.
187 - assume that most object keys are simple ascii words and
188 optimise this case, penalising the general case. This can
189 speed up decoding by 30% in typical cases and gives
190 a smaller and faster perl hash.
191 - implemented simpleminded, optional resource size checking
192 in decode_json.
193 - remove objToJson/jsonToObj aliases, as the next version
194 of JSON will not have them either.
195 - bit the bullet and converted the very simple json object
196 into a more complex one.
197 - work around a bug where perl wrongly claims an integer
198 is not an integer.
199 - unbundle JSON::XS::Boolean into own pm file so Storable
200 and similar modules can resolve the overloading when thawing.
201
2021.3 Sun Jun 24 01:55:02 CEST 2007
203 - make JSON::XS::true and false special overloaded objects
204 and return those instead of 1 and 0 for those json atoms
205 (JSON::PP compatibility is NOT achieved yet).
206 - add JSON::XS::is_bool predicate to test for those special
207 values.
208 - add a reference to
209 http://jpsykes.com/47/practical-csrf-and-json-security.
210 - removed require 5.8.8 again, it is just not very expert-friendly.
211 Also try to be more compatible with slightly older versions,
212 which are not recommended (because they are buggy).
213
2141.24 Mon Jun 11 05:40:49 CEST 2007
215 - added informative section on JSON-as-YAML.
216 - get rid of some c99-isms again.
217 - localise dec->cur in decode_str, speeding up
218 string decoding considerably (>15% on my amd64 + gcc).
219 - increased SHORT_STRING_LEN to 16kb: stack space is
220 usually plenty, and this actually saves memory
221 when !shrinking as short strings will fit perfectly.
222
2231.23 Wed Jun 6 20:13:06 CEST 2007
224 - greatly improved small integer encoding and decoding speed.
225 - implement a number of ยต-optimisations.
226 - updated benchmarks.
227
2281.22 Thu May 24 00:07:25 CEST 2007
229 - require 5.8.8 explicitly as older perls do not seem to offer
230 the required macros.
231 - possibly made it compile on so-called C compilers by microsoft.
232
2331.21 Wed May 9 18:40:32 CEST 2007
234 - character offset reported for trailing garbage was random.
235
2361.2 Wed May 9 18:35:01 CEST 2007
237 - decode did not work with magical scalars (doh!).
238 - added latin1 flag to produce JSON texts in the latin1 subset
239 of unicode.
240 - flag trailing garbage as error.
241 - new decode_prefix method that returns the number
242 of characters consumed by a decode.
243 - max octets/char in perls UTF-X is actually 13, not 11,
244 as pointed out by Glenn Linderman.
245 - fixed typoe reported by YAMASHINA Hio.
246
2471.11 Mon Apr 9 07:05:49 CEST 2007
248 - properly 0-terminate sv's returned by encode to help
249 C libraries that expect that 0 to be there.
250 - partially "port" JSON from C to microsofts fucking broken
251 pseudo-C. They should be burned to the ground for pissing
252 on standards. And I should be stoned for even trying to
253 support this filthy excuse for a c compiler.
254
2551.1 Wed Apr 4 01:45:00 CEST 2007
256 - clarify documentation (pointed out by Quinn Weaver).
257 - decode_utf8 sometimes did not correctly flag errors,
258 leading to segfaults.
259 - further reduced default nesting depth to 512 due to the test
260 failure by that anonymous "chris" whose e-mail address seems
261 to be impossible to get. Tests on other freebsd systems indicate
262 that this is likely a problem in his/her configuration and not this
263 module.
264 - renamed json => JSON in error messages.
265 - corrected the character offset in some error messages.
266
2671.01 Sat Mar 31 16:15:40 CEST 2007
268 - do not segfault when from_json/decode gets passed
269 a non-string object (reported by Florian Ragwitz).
270 This has no effect on normal operation.
271
2721.0 Thu Mar 29 04:43:34 CEST 2007
273 - the long awaited (by me) 1.0 version.
274 - add \0 (JSON::XS::false) and \1 (JSON::XS::true) mappings to JSON
275 true and false.
276 - add some more notes to shrink, as suggested by Alex Efros.
277 - improve testsuite.
278 - halve the default nesting depth limit, to hopefully make it
279 work on Freebsd (unfortunately, the cpan tester did not
280 send me his report, so I cannot ask about the stack limit on fbsd).
281
2820.8 Mon Mar 26 00:10:48 CEST 2007
283 - fix a memleak when decoding hashes.
284 - export jsonToBj and objToJson as aliases
285 to to_json and from_json, to reduce incompatibilities
286 between JSON/JSON::PC and JSON::XS. (experimental).
287 - implement a maximum nesting depth for both en- and de-coding.
288 - added a security considerations sections.
289
2900.7 Sun Mar 25 01:46:30 CET 2007
7 - code cleanup. 291 - code cleanup.
8 - fix a memory overflow bug when indenting. 292 - fix a memory overflow bug when indenting.
9 - pretty-printing now up to 15% faster. 293 - pretty-printing now up to 15% faster.
10 - improve decoding speed of strings by 294 - improve decoding speed of strings by
11 up to 50% by specialcasing short strings. 295 up to 50% by specialcasing short strings.
… …
48 case the called functions do... stuff. 332 case the called functions do... stuff.
49 - croak when encoding to ascii and an out-of-range 333 - croak when encoding to ascii and an out-of-range
50 (non-unicode) codepoint is encountered. 334 (non-unicode) codepoint is encountered.
51 335
520.2 Fri Mar 23 00:23:34 CET 2007 3360.2 Fri Mar 23 00:23:34 CET 2007
53 - the "could not sleep without debuggign release". 337 - the "could not sleep without debugging release".
54 it should basically work now, with many bugs as 338 it should basically work now, with many bugs as
55 no production tests have been run yet. 339 no production tests have been run yet.
56 - added more testcases. 340 - added more testcases.
57 - the expected shitload of bugfixes. 341 - the expected shitload of bugfixes.
58 - handle utf8 flag correctly in decode. 342 - handle utf8 flag correctly in decode.

Diff Legend

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