Need help with a regex

[Sandy]

UPDATE!!

It isn't a regex problem at all!. It was the way I was reading in the file.

I erroneously had $/='';, but then it only read my data up to, but not including the blank line.

When I changed the above code to undef $/, everything works fine.

Oops!

END_OF_UPDATE

Hello all,

I have a file that I need to parse. Each page has a header and a page number.

At this point, I am having difficulties just grabbing the header for each page (part of a more complex regex).

Problem:

If there is an extra line in front of page 54's header, the regex does not find this page. If the blank line is replaced with an 'x', the page header will be found.

Code & output when it doesn't work:

#!/usr/bin/perl
use warnings;
use strict;
my $header =
    join("\\s*\\n",
        'My\s+Header',
        'Page\s+\d+',
        );
$/ = '';
my $file = <DATA>;
while ($file =~/($header\n+)/g)
{
    print "pos = ", pos($file),"\n";
    print $1,"\n";
}

__DATA__
My Header
Page 53
Some Text
Some More Text
Some More Text
Some More Text

My Header
Page 54
3 Chapter Title
My Header
Page 55
Some Text
Some More Text
Some More Text
Some More Text

Result:

pos = 18
My Header
Page 53

Code & output when it does work:

same code, slightly different data

#!/usr/bin/perl
use warnings;
use strict;
my $header =
    join("\\s*\\n",
        'My\s+Header',
        'Page\s+\d+',
        );
$/ = '';
my $file = <DATA>;
while ($file =~/($header\n+)/g)
{
    print "pos = ", pos($file),"\n";
    print $1,"\n";
}

__DATA__
My Header
Page 53
Some Text
Some More Text
Some More Text
Some More Text
x
My Header
Page 54
3 Chapter Title
My Header
Page 55
Some Text
Some More Text
Some More Text
Some More Text

Result:

pos = 18
My Header
Page 53

pos = 93
My Header
Page 54

pos = 127
My Header
Page 55

I have checked the data with a binary editor, and nothing weird is on the blank line.

I'm using active state perl V5.6.1 on a windows 2000 professional machine.

Any insight would be appreciated.

Sandy

UPDATE:

Just a note: the code above was part of a series of test. The same result is obtained even if I use the 'sgmx' modifiers for the regex.

[GrandFather]

One of the ways PerlMonks works best is when you distill your problem down to a small test case. If you don't find the answer during distilation, then there's a 90% chance you will see it 10 seconds after clicking "create".

Well posted node!