Anyone a regex guru?

[danielk]

I think this would do the trick:

<img\b[^>]+\bsrc=(("[^"]+")|('[^']+'))

Indeed it would. However, I think using a back reference would be slightly more elegant (tested now):

<img\b[^>]+\bsrc=(["'])(.*?)\1

The trick to get this to work properly was to use a non-greedy quantifier, so something like <img src="foo" alt="bar" /> won't return foo" alt="bar in the captured substring. Note that this still won't handle <img src=foo> as allowed by HTML without the X, but frankly, I wouldn't give a damn if this was my own code.

 

[Paul C. Anagnostopoulos]

I did not know about Perl's nongreedy quantifiers. Cool.

~~ Paul

 

[Floyt]

http://txt2regex.sourceforge.net/ <--- "English" to regex compiler.

Ooh, thankee very much. That might take out some of my more aggravating OpenOffice interludes.