That did the trcik! Nice and clean, thanks Rob > -----Original Message----- > From: Hanson, Rob [mailto:[EMAIL PROTECTED] > Sent: Wednesday, June 04, 2003 2:55 PM > To: Dan Muey; [EMAIL PROTECTED] > Subject: RE: Modify links > > > Try using URI to figure out the absolute URL. > > use URI; > > # the base is the *current absolute page* > my $base_url = 'http://foo.com/documents/help.html'; > > print URI->new_abs('doc1.html', $base_url), "\n"; > print URI->new_abs('./doc2.html', $base_url), "\n"; > print URI->new_abs('../documents/doc3.html', $base_url), > "\n"; print URI->new_abs('http://somewhere.com/', $base_url), "\n"; > > <<< SCRIPT OUTPUT >>> > http://foo.com/documents/doc1.html > http://foo.com/documents/doc2.html > http://foo.com/documents/doc3.html > http://somewhere.com/ > > So your regex *might* look like this (untested)... > > my $base = 'http://foo.com/documents/help.html'; > $html_code =~ s/href="(.*?)"/'href="' . URI->new_abs($1, > $base) . '"'/seg; > > Rob > > > -----Original Message----- > From: Dan Muey [mailto:[EMAIL PROTECTED] > Sent: Wednesday, June 04, 2003 3:26 PM > To: [EMAIL PROTECTED] > Subject: Modify links > > > I'm trying to work out a regex that will do this: > > Take an entire page's html: my $html_code; # all lines in > thes one variable > > And make any href's that are relative absolute by prepending > $url into them: > $url = "http://myclonesite.com";; > make <a href="./documents/help.hml"> into <a > href="http://myclonesite.com/documents/help.html";> > $html_code =~ s/href\=\"\.?\/?(.*)\"/href\=\"$url\/"/ig; > the rpobolem with this is it prepends $url to absolute > url's also > > I need to say : > Put $url in front of relative urls (make ./foo /foo or foo > $url/foo and ../foo would have to be treated differently, > ignored for now) in href if href does not start with https?:// > > Any ideas? > > TIA > Dan > > > -- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] >
-- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
