Perl Regex multiline match issues

adymcc · 03-28-2008, 08:07 AM

Hi,

I have a file which contains multiple lines which I want to be able to search as one string, but I am having problems getting the regular expression match to continue searching following a newline character.

The contents of the file I'm testing with is:

This is my
multi-line string

For completeness here is the file printing the control characters:

Code:

user@host~]$ cat -vet file
This is my$
multi-line string$

Here's the regex that seems correct:

Code:

#!/usr/bin/perl

open (FILE, "<", file) or die "can't open file";

if (<FILE> =~ /.*(This.*)/s) { print "$1";}
else {print "failed";}

close FILE

Here's the output:

Code:

user@host~]$ ./perl.pl | cat -vet
This is my$

Surely with the /s modifier it should continue past the newline character and display the whole file?

Telemachos · 03-28-2008, 10:08 AM

The problem is not the match or the regex, I think, but how you are opening the file. The context for if is scalar, so it just takes just the first, single line of the file and it tests that against your regex. Try it this way instead:

Code:

perl -e 'print "$1\n" if "This is my\nmulti-line string" =~ /.*(This.*)/s'

The match operator tests against a scalar - ie, it takes one item and tests that against your regex - for example, if you open the file in a while loop, then perl treats each line, one at a time, as $_ and tests $_ against the regex. To test the whole file at once you would have to first slurp the whole file into a scalar.

Code:

#!/usr/bin/perl
use strict;
use warnings;

open(my $file, "<", "file")
    or die "Can't open file: $!";
my $text = do { local $/; <$file> };

if ($text =~ /.*(This.*)/s) {
    print $1;
}

Edit: see this from Perldoc's FAQ http://perldoc.perl.org/5.8.8/perlfa...all-at-once%3f

adymcc · 03-31-2008, 09:45 AM

Thanks for this Telemachos, thats nailed it. And all the time I was looking at the regex match!