RegEx Matching, Loop Breaking...

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

So I have the below code. I know there are some extra bits, but it is not behaving how i would like. basically I have a file i am reading in, then a second file read in, and looping through and saying that if a line in XML contains anything from a line in DOMAIN, set a variable to nothing and exit the innermost loop. But the problem is that nothing ever matches. If I replace the $cleanedDomain line with a literal, it matches. The XML file is xml and the domains file is just a list of domain names... It looks right to me but i think i am missing something.


use strict; use warnings; 
my $linexml;
my $cleanedXML;
my $cleanedDomain;

 open (XML, 'domainList.xml');
 while (my $xmlline = <XML>) {
    
    $linexml = $xmlline;
    $cleanedXML = quotemeta $xmlline;
    open (DOMAINS, 'dead_domains.txt');
    while (my $domainline = <DOMAINS>) {
        
        $cleanedDomain = quotemeta $domainline;
                
        if ($cleanedXML =~ m/$cleanedDomain/) { $linexml = ''; last;}
        
         
    }
    
    
    close (DOMAINS);
    open (MYFILE, '>>new_domainList.xml');
     print MYFILE "$linexml";
    close (MYFILE); 
    
    
    
 }
 close (XML);
[download]

Comment on RegEx Matching, Loop Breaking... Download Code

Replies are listed 'Best First'.
Re: RegEx Matching, Loop Breaking... by toolic (Bishop) on May 28, 2013 at 17:30 UTC
Tip #3 from the Basic debugging checklist: chomp? `while (my $domainline = <DOMAINS>) { chomp $domainline; $cleanedDomain = quotemeta $domainline;` [download]	[reply] [d/l]
Re^2: RegEx Matching, Loop Breaking... by rardoe (Initiate) on May 28, 2013 at 17:49 UTC
chomp doesn't change anything, i've tried it with and without.	[reply]
Re: RegEx Matching, Loop Breaking... by hdb (Monsignor) on May 28, 2013 at 18:56 UTC
Why do you need to "quotemeta" $xmlline? On the lefthand-side of the match this should not be needed. Without the files it is difficult to develop more ideas, the logic seems ok. I would add a print statement just before the match to see the contents of $cleanedDomain and $cleanedXML. I agree with comments above that you should revisit your open/close policy.	[reply]
Re^2: RegEx Matching, Loop Breaking... by rardoe (Initiate) on May 28, 2013 at 19:45 UTC
Taking the quotemeta out fixed my issue. Thanks!	[reply]
Re^3: RegEx Matching, Loop Breaking... by AnomalousMonk (Archbishop) on May 28, 2013 at 20:05 UTC
The reason taking the quotemeta out fixed the problem is that a `quotemeta`-ed string that contained `\W` characters will, in general (there may be some oddball corner cases), not regex-match with itself after the string has been double-quote-ishly interpolated into a regex. The reason is that after a string is `quotemeta`-ed, something like `'\-'` in the `quotemeta`-ed string is literally a backslash-hyphen sequence of characters, but this sequence in a regex only matches a hyphen leaving the backslash in the `quotemeta`-ed string unmatched. `>perl -wMstrict -le "my $s = 'a-b cd'; my $qm_s = quotemeta $s; print qq{raw: '$s' quotemeta-ed: '$qm_s'}; ;; printf qq{%s equal \n}, $qm_s eq $qm_s ? '' : 'NOT '; printf qq{%s match \n}, $qm_s =~ /$qm_s/ ? '' : 'NO '; " raw: 'a-b cd' quotemeta-ed: 'a\-b\ c\*d' equal NO match` [download]	[reply] [d/l] [select]
Re: RegEx Matching, Loop Breaking... by Laurent_R (Canon) on May 28, 2013 at 17:50 UTC
Asides from the fact that you need to chomp the data from your file, I think is is probably not a good idea to open, read and close a file for every single line of input of anoher file. You probably want to read your domain file only once at the beginning, clean and store its content in memory (in an array or a hash) for when you process your other file.	[reply]