Attempting to pull just filename from backtick - perl

Question

I have an issue where I am attempting to pull just a filename from the output of running a backtick, my code is as follows:
$var = `munpack -f filename`;
If anyone is familiar with mpack the output will be something like:
tempdesc.txt: File exists
file_20130620.zip (application/octet-stream)
I am trying to just get the filename, however, all my attempted regexes have failed. I have even tried to just remove the linebreaks and then attempt to process the information and I cannot. I thought they could just be whitespace and remove the whitespace but those regexes have failed. I could go through and list every regex I have tried to pull this data and I can provide that if necessary, but maybe someone has something that could work. I can't produce any matches that id like nor alter the output in any way. So just to be clear im looking for something that will output me just the filename ex: file_20130620.zip
Some suggestions given with output:
$var =~ m{^(.+?)\(}m and print "$1\n";
output:
tempdesc.txt: File exists
file_20130620.zip
($filename) = $var =~ /(?s:.*\n)?(.*) \([^)]+\)\n/;
output:
tempdesc.txt: File exists
file_20130620.zip
if($var =~/\S+: [^\n]+\n(\S+) [^\n]+\n/) { printf $1; }
output:
tempdesc.txt: File exists
Fix per ysth:
$var = `munpack -f filename 2>/dev/null`; #will remove 'tempdesc.txt: File exists'

Answer 1

Assuming the filename is before a space before a parenthesized mimetype on the last line of output:
($filename) = $var =~ /(?s:.*\n)?(.*) \([^)]+\)\n/;
Though I'd rather create a temporary directory (use File::Temp) and unpack in it and just look for what file(s) are there than parse the output.
It is possible that the File Exists warning isn't actually in $var, but is appearing in your output because munpack is writing it to stderr (which doesn't get captured by backticks.)
Try doing munpack -q -f ... or munpack -f ... 2>/dev/null.

Answer 2

If you can assume that the filename is followed by a parenthesized description, something like this works:
$var =~ m{^(.+?)\(}m and print "$1\n";
The \m modifier treats a string as one with multiple lines so that you can match ^ and $ on any line. See perlre

Answer 3

I put your sample output to file1.txt as I dont have mpack utility installed.
And this regexp works
#!/usr/bin/perl
my $var = `more file1.txt`;
if($var =~/\S+: [^\n]+\n(\S+) [^\n]+\n/)
{
printf $1;
}