Take a look at File::ReadBackwards. It's far easier and quicker than reversing the file in memory.

# ls -lh /var/lib/mysql/ibdata1
-rw-rw----  1 mysql mysql 66M Mar 19 11:05 /var/lib/mysql/ibdata1

# wc -l /var/lib/mysql/ibdata1
323833 /var/lib/mysql/ibdata1

# time tac /var/lib/mysql/ibdata1 > /tmp/bar
real    0m0.643s
user    0m0.173s
sys     0m0.322s

# time /usr/bin/perl > /tmp/foo
use warnings;
use strict;
use File::ReadBackwards;
my $bw = File::ReadBackwards->new( '/var/lib/mysql/ibdata1' ) or
  die "$!" ;
while( defined( my $log_line = $bw->readline ) ) {
  print $log_line ;
}
__END__
real    0m29.076s
user    0m11.910s
sys     0m10.819s

# md5sum /tmp/foo
2b31d9f47525853842d5dbce584bd95c  /tmp/foo
# md5sum /tmp/bar
2b31d9f47525853842d5dbce584bd95c  /tmp/bar
[download]

Update Wed Mar 26 10:11:31 CET 2008: It seems that the tac + process file forwards using Perl approach combined is still much faster than process file using Perl and File::ReadBackwards:

# time /usr/bin/perl > /tmp/foo
use warnings;
use strict;
use File::ReadBackwards;
my $bw = File::ReadBackwards->new( '/var/lib/mysql/ibdata1' ) or
  die "$!" ;
while( defined( my $log_line = $bw->readline ) ) {
  $log_line =~ s/1/2/g;
  print $log_line;
}
__END__

real    0m27.431s
user    0m12.906s
sys     0m10.701s
[root@afflinux aff]# time /usr/bin/perl > /tmp/bar
use warnings;
use strict;
my $FH = undef;
open($FH,  '/var/lib/mysql/ibdata1' ) or
  die "$!" ;
while( my $log_line = <$FH> ) {
  $log_line =~ s/1/2/g;
  print $log_line;
}
__END__

real    0m5.249s
user    0m1.535s
sys     0m0.293s
[download]

It's much faster than using Perl and File::ReadBackwards:

True if the only aim is to reverse the file.

If however, the aim is to process the data in the file in the reverse order using perl, you'd have to compare F::RB with using tac and then reading every record in the file using Perl.

Also, how many lines are there in your test file?

---

Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.

"Science is about questioning the status quo. Questioning authority".

In the absence of evidence, opinion is indistinguishable from prejudice.

"Too many [] have been sedated by an oppressive environment of political correctness and risk aversion."

Wow. Might be interesting to take a peak at the tac source code to try to figure out what it's doing that's so much better than File::ReadBackwards. There must be a trick here, that's way more than the standard C-versus-Perl difference.

-sam

If the file he used doesn't contain any newlines, all it does is a straight copy.

---

Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.

"Science is about questioning the status quo. Questioning authority".

In the absence of evidence, opinion is indistinguishable from prejudice.

"Too many [] have been sedated by an oppressive environment of political correctness and risk aversion."

tac is also available for win32 e.g.

C:\mycode>perl -le "print $_ for qw[spam bacon eggs]" > things.txt

C:\mycode>cat things.txt
spam
bacon
eggs

C:\mycode>tac things.txt
eggs
bacon
spam
[download]

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

my $file = shift || $0;

my $pid = open(FH,"tac $file |") or die "$!\n";

my $count = 0;
while (<FH>){
 #do your line processing here
 print $count++ .' '. $_; 
}
[download]

That's completely untested, right? Secondly, even if you did fix the most obvious mistakes, it does not do what you want it to (reverse each line vs. print file reversed).

Thirdly, if one was to use a load-the-whole-file-into-an-array approach it most likely would be very slow for large files.

--
Andreas