[Millennium]

I'm having some trouble with some of my regular expressions. Most of the ones I work with, I can get to do what I want, but I'm having problems with three of them that don't even seem to come close...

Take a chunk of text, likely to contain multiple paragraphs (like a messageboard entry). I want to do the following with it:

1) Get every chunk of text which is both preceded and followed by at least two newlines.
2) Get every chunk of text which is preceded by an arbitrary number of newlines, but followed by only one.
3) Get every URL in a document.

This would, I assume, take three expressions. The only problem is, I can't get anything I can think of to work. Anyone here have any ideas?

[sadie]

Most of my regex experience was in Perl, which i haven't used in a year or two, but my guesses would be something like:

<BLOCKQUOTE><font size="1"face="Geneva, Verdana, Arial">code:</font><HR><pre><font size=1 face=courier>
$var ~ m/\n\n([^\n]*)\n\n/;
$result = $<font color = blue>0</font>; <font color = brown>// or possibly $<font color = blue>1</font>, can't remember</font>
</font>[/code]

That will match any string that does not contain a newline, but which is preceded and followed by two newlines. I can't remember how you match all of them, rather than just one - figure that bit out yourself.

<BLOCKQUOTE><font size="1"face="Geneva, Verdana, Arial">code:</font><HR><pre><font size=1 face=courier>
$var ~ m/\n*([^\n]*)\n[^\n]/;
$result = $<font color = blue>0</font>; <font color = brown>// again, may be $<font color = blue>1</font></font>
</font>[/code]

This matches anything preceded by any number of newlines, and followed by a newline and a non-newline character. This fails to match when the final newline is the last character in the string, but it's early morning and i can't think straight. None of this code has been tested, so it's probably all a big pile of schite.

The last one is more complicated - you have to match several known strings, including http:, mailto:, ftp:, www., *@*.*, etc.

Regular expressions can do an incredible amount of stuff, including everything you asked for, if you only know how to use them. What language are you using, anyway?

[seb2]

to add a little to the code above, "\n{2}" looks a bit nicer than "\n\n", but that's only cosmetic.
the result will be in $1, $0 is the name of your perl program.

might be an approach to grabbing urls would be something like ([\w]?/\/\S?), this would take everything that has "://" in it as a url. might be something like (((https?|ftp)/\/|mailto\S?) would be a better approach, but that's just a guess...

[edit: disabled smilies, didn't turn out the way i expected it to... ;-)]

[ 01-09-2002: Message edited by: seb2 ]