From: TAKAHASHI Masayoshi <maki@...>
Date: 2001-11-21T20:14:21+09:00
Subject: [ruby-talk:26002] Re: REXML parser bug(?)

Hi,

Tobias Reif <tobiasreif@pinkjuice.com> wrote:
> > The script to parse string with REXML::Document,
> > 
> >    require 'rexml/document'
> >    doc = REXML::Document.new("<foo bar=\"foo'bar\"></foo>")
> >    doc.write(STDOUT)
> > 
> > returns:
> > 
> >   <foo bar2='foo' bar='foo'/>
> 
> 
> really? where does the '2' come from?

Oops, sorry. I've copy&pasted old script.

Correct sample is:

  require 'rexml/document'
  doc = REXML::Document.new("<foo bar=\"foo'bar\" bar2='foo\"bar'></foo>")
  doc.write(STDOUT)


I think regex pattern in rexml/attribute.rb (line 15)

>		PATTERN = /\s*([\w-]*:?[\w-]+)\s*=\s*["\'](.*?)["\']/m

causes this behavior.


Regards,

TAKAHASHI Masayoshi (maki@inac.co.jp)