If you have ever tried to get a job as a Perl programmer you have probably been asked to create a tail replacement. Well I have. I was bored last night so I made one. Its pretty straight forward, the one issue I can see is that of detecting where the 10th line from the end is, for large files my method may be slow. A character by character search from the tail of the file looking for newline characters could be a faster solution.

#!/usr/bin/perl
use strict;

#Get filename and optional lines to tail from
my $filepath = shift @ARGV;
my $lines = shift @ARGV || 10;

#Get the inital inode and size
our ($inode,$size) = filestats($filepath);

open (FILE,$filepath) or die "Could not open $filepath for reading\n";

my @lineEnds = ();

#Get a list of all the line ending locations
while() {
   push @lineEnds,tell();
}

my $startloc = $lineEnds[($lines*-1)-1];
@lineEnds = (); #No more need;

#Seek to the line requested
seek FILE,$startloc,0;

#Loop Forever, like tail -f
for(;;) {

#Clear Out any EOF flags
seek FILE,0,1; #Seek 0 bytes from current pos

 #Print out any new file contents, or inital requested lines
 while() {
   print;
 }
 sleep(1); #Sleep to allow other procs to run.
 my ($ninode, $nsize) = filestats($filepath); #Get the current inode and file size;

 if ($ninode != $inode) {
    print STDERR "File Rotated Re-opening file @ $filepath\n";
    close(FILE);
    open(FILE,$filepath) or die "File Could Not be Re-Opened!\n";
    $inode = $ninode;
    $nsize = $nsize;
 }

 if($nsize < $size) {
    print STDERR "File Cleared Out Seeking to Top\n";
    seek FILE,0,0; #Seek to Top
    $size = $nsize;
 }

 #Update File size
 $size = $nsize;

}

#Small untility sub for grabing inode and file size;
sub filestats {
    my $file = shift;
    my @stats = stat $file;
    #Inode and file size
    return ($stats[1],$stats[7]);
}