okjson,rwz

okjson's Introduction

This is a json parser, written in ruby.
It has no dependencies other than ruby.
See the file tested-on for a complete list
of ruby interpreters used for testing.


To Use:

  OkJson.encode({"abc" => [123, -456.789]})
  => "{\"abc\":[123,-456.789]}"

  OkJson.decode('{"abc":[123,-456.789e0]}')
  => {"abc"=>[123, -456.789]}

That's it.


This library is intended to be "vendored".
It is not a gem; instead, copy okjson.rb
into your project and "require" it directly.
This method helps you avoid an external
dependency. It's only about 575 lines of
source; about half of that is UTF-8 coding.
If you bundle okjson with your library,
please change the module's name to something
like MyLib::OkJson, to avoid conflicts with
other libraries.

OkJson is not particularly fast or flexible,
but it is designed to be simple and portable.
It should suffice for most needs.

It provides two guarantees:
 1. okjson will never generate invalid json
 2. OkJson.decode(OkJson.encode(x)) == x

Note for people who've used other ruby json
libraries: json does not represent symbols,
so okjson does not encode them. It won't
convert them to strings; that would silently
corrupt your data. The same holds for all of
ruby's other data types that don't exist in
json, including Range, Regexp, Proc, Time,
and user-defined classes.

Patches to increase the correctness,
readability, or portability of the
source code will be gratefully accepted.
Other patches (for example, extra features
or optional behavior) likely won't.


To Test:

  $ ./rubies ruby rbx /opt/ruby192/bin/ruby

(Note: these tests originated at
https://github.com/genki/json
but have been changed to follow
RFC 4627 and revised substantially.)

Recommend Projects