Is the variable 'tagname'

kazimmerman — Thu, 09 Jul 2009 00:57:17 +0000

Is the variable 'tagname' defined elsewhere in the code? I'm guessing that's the "certain tag" that will be sought to which you were referring?

Anyway, your expression is confusing me a bit. I'm certainly no expert (and I know you didn't write the expression, but why at one point are they searching for a ">" or anything but a ">" (optional) and then a ">". Wouldn't the second search produce the same result?

...(>|[^>]*>)...

Also, maybe I'm not seeing exactly what you want, but I don't see the reasoning behind all the line feeds, tabs, etc...

Here's what I would use, though I haven't tested it completely:

"/^(.*)(<$tagname([^>]*>))(.*)(<\/$tagname>))(.*)$/"

Edit:
Thinking over it again, if you're just looking for the data between the tags, this would probably be a better expression:

"/^.*<$tagname[^>]*>(.*)<\/$tagname>.*$/"

As far as I understand, anything in parenthesis is captured when the expression is matched...

Is the markup likely to be

Abhishek Reddy — Wed, 08 Jul 2009 22:33:49 +0000

Is the markup likely to be well-formed? If so, you can run it through a suitable HTML or XML parser. Traverse the resulting DOM object to access the nodes you're interested in. Look at the SimpleXML extension for a start.

For malformed markup, a regular expression may be best. But if it's feasible, a simpler one which is more specific to the document will be easier to maintain and should work well enough. Very general regexps like you've posted will be hard to debug and adapt to changing inputs.