I recently made a little script to come up with an index for the increasingly unwieldy Name Space thread. This led me into an area I've not explored before (one of many, I hasten to add), that of extracting info from HTML. I wanted something that wd go through the page and pull out a node number and name for the first post by each monk who had contributed to the thread. What I came up with was this:

#!/usr/bin/perl -w

use strict;
use CGI qw(:standard :cgi-lib);
use LWP::Simple;

my $url ="http://perlmonks.org/index.pl?node_id=110166";
my $html = get $url or die "can't get url $!";
my %names;

#find the names and node ids:
while ($html =~ s/=(\d*?)&lastnode_id=110166">[^<]*<\/A><BR> by <A HRE
+F="\/index\.pl\?node_id=\d*&lastnode_id=110166">(.*?)<\/A> on \w{3} \
+d{2}, \d{4} at \d{2}:\d{2}//s) {
    $names{$2} = $1 unless $names{$2};
}

# print out a page of links to nodes:
for (sort { lc($a) cmp lc($b) } keys %names) {
    print "<A HREF=\"/index.pl?node_id=$names{$_}&lastnode_id=110166\"
+>$_</A> | ";
}
[download]

... which does the job, BUT the regex is big and fat and ugly, and I just wondered whether there was a more elegant, less impenetrable way to do it (i.e. I wondered how *many* such ways there were). I looked at HTML::Parser, but (and perhaps my inspection was too cursory) it didn't seem as though it wd help me much in pulling out bits of tags, as I need to here. Also, I felt that the while loop was a bit clumsy... but couldn't see a quicker way to capture two matches into a hash. I'd be very interested in any suggestions how to do this better.

§ George Sherston

Edit: chipmunk 2001-11-11