I have some text files (several hundred megabytes each) that I am processing. To simplify, I am going through the sections (think of them as paragraphs) and removing lines that are "ref n" (where n is an integer). There will be just a few hundred of these per file.

So I am just reading the whole file into memory, and substituting out the offending lines (I am actually removing the "ref n" lines, and not the lines that begins with n, which I am matching in the first line of code below).

while ($allfile =~ /^(\d+) /mg)
    {
    my $objectref = 'ref' . $1;
    push (@objects, $objectref);
    }

for (@objects) { $allfile =~ s/^ +$_\n//m; }

print $allfile
[download]

This worked fine (although likely far from the best way to do it), taking about ten minutes or so on average. Then I found out that (rarely - maybe once every dozen files or so) some lines that I need to remove will actually be "foo ref n". No problem, I thought. I just changed to code to:

for (@objects) { $allfile =~ s/^ *(foo)? +$_\n//mn; }
[download]

Something is not working as I expected. :) I am ninety minutes into processing the first file after the code change, and there is no sign of any progress. Why is it taking so long, and how can I improve my algorithm / code? Thank you in advance.