Re:A Liberal, Accurate Regex Pattern for Matching URLs
by wingo on Friday November 27, 2009 @03:58PM
(#964364)
It looks like line noise. But it's funny though, I can read regexps better than I can read formal semantics. Having tried once this evening to read http://matt.might.net/papers/might2007diss.pdf, regexps are refreshing :)
Re:A Liberal, Accurate Regex Pattern for Matching URLs
by mahmud on Friday November 27, 2009 @08:57PM
(#964687)
Please do not link to random papers on semantics, you don't know whose weekend you might ruin.
Horrible nerd-snipping there; I grok half of the paper, and have no choice but to study the formal semantics of the other half :-(
Re:A Liberal, Accurate Regex Pattern for Matching URLs
by PStamatiou on Saturday November 28, 2009 @02:08AM
(#964899)
Matt Might (author of linked PDF) was my Computer Science TA four or five years ago at Georgia Tech... for a java class. small world haha
Re:A Liberal, Accurate Regex Pattern for Matching URLs
by riffraff on Friday November 27, 2009 @05:30PM
(#964457)
what is wrong with URI.parse in the stdlib? the articles' re goes out of url validation to get urlish things in free text, (e.g. " lokk at http://goo.com/bat, lovely" the comma would actually be part of a valid url, but the regex tries to detect this. Yet, for url validation ruby's lib is enough, I believe.
Re:A Liberal, Accurate Regex Pattern for Matching URLs
by ehsanul on Friday November 27, 2009 @09:07PM
(#964698)
Well, with URI.parse, you can send in a ftp uri, or just "something.com" (which is parsed as just a path). So not really what I need from URL validation. If there's something else I'm missing, I'm happy to know.
Edit: And btw, for an industrial-strength Ruby URI library (to replace the standard library), check out Addressable: http://addressable.rubyforge.org/
/^(https?):\/\/((?:[a-z0-9.\-]|%[0-9A-F]{2}){3,})(?::(\d+))?((?:\/(?:[a-z0-9\-._~!$&'()+,;=:@]|%[0-9A-F]{2})))(?:\?((?:[a-z0-9\-._~!$&'()+,;=:\/?@]|%[0-9A-F]{2})))?(?:#((?:[a-z0-9\-._~!$&'()+,;=:\/?@]|%[0-9A-F]{2})*))?$/i
Yeah, more involved. Though it parses it into the url parts, and it does work.