Simple regex gone wierd

digger has asked for the wisdom of the Perl Monks concerning the following question:

Esteemed Monks,

I am working on another script that makes changes to postscript files. This time I am trying to delete lines that ask for paper types that aren't going to be used in the document. My sub takes the current line and a list of media types to remove.

Here's my sub and the sample data I copied out of a larger file for testing.

sub kill_media {
    my $line = shift;
    my $media;
    if ($line =~ m!^%%DocumentMedia:!){
        $media_killed=1;
        print OUT $line;
        while ($line = <IN>){
            foreach $media (@_) {
                if ($line =~ m!$media!){
                    $line = "";
                    print "Killed media type: $media\n";
                }
                elsif ($line !~ m!^%%+!){
                    return $line;
                }
            }
            print OUT $line;
        }
    }
----SAMPLE DATA-----
%XRXbegin: 001.014
%XRXtitle: JOB P04063001.X0026 - Wps
%XRXsenderName: CompuSet
%XRXdisposition: PRINT
%XRXcopyCount: 1
%XRXrequirements: duplex
%XRXsignature: False
%XRXorientation: Portrait
%XRXxImageShift: 0
%XRXyImageShift: 0
%XRXend
%!PS-Adobe-3.0
%%DocumentMedia: anything 612 792  0 white  ()
%%+ anything 612 792  0 blue  ()
%%+ anything 612 792  0 green  ()
%%+ anything 612 792  0 pinky  () 
%%+ anything 612 792  0 UNIVERSALID  ()
%%Creator: CompuSet Version  8.0.0 
%%CreationDate:  3/03/2004   16:04:28
%KDKHost: Binder
%%BoundingBox: 0 0 612 792
%%Pages: 182
[download]

I want to kill the pinky and UNIVERSALID media calls. The strange thing is the line for UNIVERSALID doesn't get deleted when it is the last one. If I swap the pinky and UNIVERSALID lines, both get deleted correctly. This isn't making any sense to me, so I am seeking the wisdom of the monks.

digger

Comment on Simple regex gone wierd Download Code

Replies are listed 'Best First'.
Re: Simple regex gone wierd by Roy Johnson (Monsignor) on Apr 19, 2004 at 15:08 UTC
In your foreach loop, on a successful match, you empty line. The next pass through the foreach loop, you're going to fail to match `^%%+` and thus return. The fix is to put a `last` after the `print "Killed media type: $media\n"`. Once you've killed, after all, you needn't bother with more checking. Alternatively, and probably better, is to put the check for `^%%+` before the foreach loop. The PerlMonk `tr///` Advocate	[reply] [d/l] [select]

Replies are listed 'Best First'.

Re: Simple regex gone wierd
by Roy Johnson (Monsignor) on Apr 19, 2004 at 15:08 UTC

^%%+

The fix is to put a last after the print "Killed media type: $media\n". Once you've killed, after all, you needn't bother with more checking.

Alternatively, and probably better, is to put the check for ^%%+ before the foreach loop.

The PerlMonk tr/// Advocate

[reply]
[d/l]
[select]