JRuby unable to create HTML doc if URL arg provided

[nirvdrum]

If a URL is provided when creating a document, Nokogiri errors out on JRuby:

doc = Nokogiri::HTML(open('http://google.com/').read, 'http://google.com/', nil)
pp doc.errors

=> [#<Nokogiri::XML::SyntaxError: /home/nirvdrum/dev/workspaces/mogotest/http:/google.com (No such file or directory)>]

[nirvdrum]

I'm confirming that this is still an issue in 1.5.3.

[flavorjones]

Reproduced using this script:

#! /usr/bin/env ruby

require 'nokogiri'
require 'open-uri'
require 'yaml'

puts Nokogiri::VERSION_INFO.to_yaml
html = open('http://google.com/').read
doc = Nokogiri::HTML html, 'http://google.com/', nil
puts "Doc: #{doc.to_html.inspect}"
puts "Errors: #{doc.errors}"

output for me:

--- 
warnings: []

nokogiri: 1.5.3
ruby: 
  version: 1.8.7
  platform: java
  description: jruby 1.6.5.1 (ruby-1.8.7-p330) (2011-12-27 1bf37c2) (Java HotSpot(TM) 64-Bit Server VM 1.7.0) [linux-amd64-java]
  engine: jruby
  jruby: 1.6.5.1
Doc: ""
Errors: /home/mike/code/oss/nokogiri/http:/google.com (No such file or directory)

@yokolet - any ideas?

[jvshahid]

The problem is that the url is assumed to be a local file path. I modified the code to use the [URI.create](http://docs.oracle.com/javase/6/docs/api/java/net/URI.html#create(java.lang.String\)) method to parse the url and fall back to the old behavior in case of failure.

[yokolet]

@jvshahid's patch looks ok. But, before merging the change, let me clarify this.

@flavorjones
Now, nonet is a default setting. Should this prevent internet connection by "nonet" setting? Only when donet is specified, this example should work?

I still don't understans "nonet" option should prevent what.

[yokolet]

OK. I figured out Nokogiri is allowed to connect the Internet in this case with nonet option.

@jvshahid would you give us pull request for this fix as well?

The test case needs a bit tweak when there's not internet connection available. I'll do that after merging pull request.

[jvshahid]

I created a pull request here

[yokolet]

I merged the pull request. Thanks!

[flavorjones]

in 1.5.5.rc1