Hi Michael, Thanks very much for the assistance, I'll have to investigate further! Kind Regards, Aidan Lister "Michael Sims" <michaels@xxxxxxxxxxxxxx> wrote in message news:EOEIIEMPJGBOGHFJBKAFEEDICBAA.michaels@xxxxxxxxxxxxxxxxx > Aidan Lister wrote: >> Hello list, >> >> I'm pretty terrible with regular expressions, I was wondering if >> someone would be able to help me with this >> http://paste.phpfi.com/31964 >> >> The problem is detailed in the above link. Basically I need to match >> the contents of any HTML tag, except a link. I'm pretty sure a >> lookbehind set is needed in the center (%s) bit. >> >> Any suggestions would be appreciated, but it's not quite as simple as >> it sounds - if possible please make sure you run the above script and >> see if it "PASSED". > > So basically, you want to put a link around "foo", only if it doesn't > already have one, right? > > The problem with look-behind assertions is that they have to be > fixed-width. > If you're certain of what kind of data you're going to be dealing with > then > this may be sufficient. For example, I came up with a regex that will > PASS > your script but I doubt seriously that it'll be very useful to you as it > would be easy to break it by coming up with various test cases. For your > single test case, however, this works: > > /(?<!<a href="foo">)(?<!<a href=")(foo)/ > > The problem is that HTML tags can be split across lines...they have have > any > variable amount of whitespace within the tag...they can have other > attributes (class, id, onClick), etc. Since look behind assertions have > to > be fixed width it'd be impossible (IMHO) to come up with a single regex > that > would match all cases, unless the input data was uniform. For example, > stuff like > > <a href = "foo" ID="id1" class="redlink" > onClick="javascript:someFunction();">foo</a> > > and its infinite variants could not be trapped for with a single regex > since > you cannot have an infinite number of fixed width look-behind assertions. > If quantifying modifiers such as '*', '+', and '?' were allowed in > look-behind assertions it would be possible, but they aren't (see "man > perlre"). > > If your data is coming from unknown sources you'll probably have to use a > full fledged HTML parser to pull out text that isn't already part of an > <a> > tag. I know there are several of these available for perl and I'm sure > there are for PHP too but I'm unaware of them. > > Sorry if this isn't terribly helpful. Maybe I'm overlooking something and > someone else will point out a simple way to accomplish what you're trying > to > do... -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
