External links are broken with several multilines about wikicloth

[avaranovich]

The issue appears once you have \n\n before http:// ...
Text fragment example: namespace:\nhttp://mediawiki.org\nAll

For now this regex works for me :

 url_check = Regexp.new( '(^|[\n]*)([\w]+?://[\w]+[^ \"\r\n\t<]*)', Regexp::MULTILINE | Regexp::IGNORECASE )

but it breaks other external link format witth "[...]"

Here is the original regex:

url_check = Regexp.new( '(^|[\n ])([\w]+?://[\w]+[^ \"\r\n\t<]*)', Regexp::MULTILINE | Regexp::IGNORECASE )

and it seems explicitly requires a whitespace before http://

[avaranovich]

@nricciar Seems to be a different issue,

 See a designated [https://github.com/repo/README.md README].

rendered as

<p>See a designated <a href="<a href="https://github.com/repo/README.md">https://github.com/repo/README.md</a>" target="_blank"><a href="https://github.com/repo/README.md">https://github.com/repo/README.md</a></a>.</p></body></html>

I think the previous fix breaks this use case.

I also asked the question here: http://stackoverflow.com/questions/16011555/ruby-regex-doesnt-recognize-beginning-of-string

so regex by itself seems to be OK, but it is not working in the parsing context. Perhaps I am missing something important.

here is my test wiki markup:

 == Usage ==

See a designated [https://github.com/repo/README.md README].

[https://github.com/repo/README.md]

This is entrance page for the service (using HTML):

http://foobar.org/resources?format=html

Let us pick the "contributions" namespace: http://foobar.org/

[nricciar]

The previous update should fix that regression. Should also work a lot better than a regular expression.