On my `JSON.dump` benchmark it shows up as 6% of runtime, compared
to 40% for `convert_UTF8_to_JSON`.
Since the vast majority of the time this function is called we
still have some buffer capacity, we might as well check that
first and skip the expensive loop etc.
With this change my profiler now report this function as 0.7%,
so almost 10x better.
a7206bf2db
002ac2771c...e09ffc0d7d
* Remove Rubinius exception since transcoding should be working now.
* Fix https://github.com/flori/json/issues/162 reported by Marc-Andre
Lafortune <github_rocks@marc-andre.ca>. Thanks!
* Applied patches by Yui NARUSE <naruse@airemix.jp> to suppress
warning with -Wchar-subscripts and better validate UTF-8 strings.
* Applied patch by ginriki@github to remove unnecessary if.
* Add load/dump interface to JSON::GenericObject to make
serialize :some_attribute, JSON::GenericObject
work in Rails active models for convenient
SomeModel#some_attribute.foo.bar access to serialised JSON data.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@43731 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* ext/bigdecimal/bigdecimal.c, ext/digest/md5/md5.c,
ext/json/fbuffer/fbuffer.h, ext/json/generator/generator.c:
Eliminate less-than-zero checks for unsigned variables.
According to section 4.1.5 of C89 standard, size_t is an unsigned
type. These checks were found with 'cppcheck' static analysis tool.
[ruby-core:57117] [Feature #8890]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@42920 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This issue is fixed in upper stream as issue #142. (https://github.com/flori/json/issues/142)
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@36541 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This breaks tests:
20120507T070102Z.log.html.gz
And don't change local codes, contribute it to upstream:
https://github.com/flori/json
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@35562 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
