Path: | README.rdoc |
Last Update: | Sun Dec 01 07:46:27 -0800 2013 |
This gem provides a (hopefully) high quality http parser library that can build request information iteratively as data comes over the line without requiring the caller to maintain the entire body of the request as a single string in memory.
It will also have a full set of specs and a Ruby-native reference library so that it can be used in implementations or environments that do not support C extensions.
Simple usage example:
p = Http::Parser.new p.parse("GET / HTTP/1.1\r\n") p.parse("Host: blah.com\r\n") p.parse("Cookie: blorp=blah\r\n") p.parse("\r\n") p.method => "GET" p.version => [1,1] p.path => "/" p.headers["HOST"] => "blah.com" p.headers["COOKIE"] => "blorp=blah"
If the request is a type that has a body, the body will be available as a stream object via p.body:
p = Http::Parser.new p.parse("PUT / HTTP/1.1\r\n") p.parse("Host: blah.com\r\n") p.parse("Content-Type: text/text\r\n") p.parse("Content-Length: 5\r\n") p.parse("\r\n") p.parse("stuff") p.body.read => "stuff"
If you use p.parse!, any trailing text that isn‘t immediately parseable will be left in the string object you pass in while what was parsed will be removed. This allows for you to hand the parser a large glob of data and allow it to figure out what it needs and what it doesn‘t. When you get more data, you can append it to your existing string and pass that in again until the request is done. You can test if the request is done by using p.done?
p = Http::Parser.new s = "GET / HTTP/1.1\r\nHost:" p.parse!(s) s => "Host:" p.done? => false s << " blah.com\r\n" p.parse!(s) s => "" p.done? => false s << "\r\n" p.parse!(s) s => "" p.done? => true
Copyright (c) 2010 Graham Batty. See LICENSE for details.