A little Tie::File fun:

perl -MTie::File -e 'tie @a, q{Tie::File}, shift; chomp $a[-1]' file.t
+xt
[download]

Note that you can get the size of a file using the -s operator.

Re truncate (at Re: How to chomp the last new line character from the file & Re: How to chomp the last new line character from the file): note that the relevant file MUST be writable.

If you take the truncate doc (which doesn't mention my boldfaced qualifier above) as a complete and accurate statement, you'll miss something: the file in FILEHANDLE or EXPR must be writable. truncate will fail or appear to fail without explanation if the file is, for example, flocked by another process, such as an editor -- or if FILEHANDLE or EXPR is opened for reading only (eg, open ($fh, '<', "EXPR") ....

BrowserUk deserves credit here for observing that Perl's truncate may well be merely "a flimsy wrapper around the POSIX call of the same name" and for referring me to http://linux.die.net/man/2/truncate, q.v., for

Here is one way:

    perl -i -pe 'chomp if eof' file1 file2 ...
[download]

--
John.

In the spirit of TIMTOWTDI ...

$ perl -pe'eof&&chomp' file | nextprocess
$ perl -0pe's/\n\Z//'  file | nextprocess
[download]

$count = `wc -l < $file`;
if($x==$count);
{
 chomp($line);
}
[download]

You don't need ($x) when you have ($.). perlvar.

But all that's too much fuss when you have eof. There's no need to keep track of what constitutes the last line number.

You could do it as a one-liner like this:

perl -pi.bak -e 'chomp if eof' file.txt



Dave

1. Verify that the file exists and that it is writable by you.
2. Seek to the end of the file, obtaining its length in bytes.   Exit if zero.
3. Seek one byte back from where you are.   Read that byte.   Exit if not a newline.   (Assuming no double-bye character-set nonsense; otherwise, you must back up one character.)
4. Truncate the file at length - 1.