When you use constant Foo . . . , 'Foo' becomes the name of a sub. Inside a regex, you need to use a form which makes the sub be called by evaluation. The \Q is not helping, either.

while (<>) {
    s/${\NEWLINE()}$//;     # should work
    print;
}
[download]

In a substitute string, the /e modifier would be needed to call the sub.

Thanks to both of you for your responses!

> It looks like you're dealing with foreign line ends...

I'm dealing with cleaning up the output of 'script' which sometimes has a ^M at EOL & '^M' in other cases. Changing $/ would be great if everything was consistent, but I see examples of both the control code & ASCII representations existing in the output.

FWIW.

I recommend the Readonly module:

  #!/usr/bin/perl -w

  use strict;
  use warnings;

  use Readonly;
  Readonly my $NEWLINE => '^M';

  while (<>) {
      s/\Q$NEWLINE$//;
      print;
  }
[download]

The usual (ugly) trick for interpolating subs inside double quotes is:

"@{ [ NEWLINE ] }"
[download]

That is, dereference an array ref containing your sub call. It works because strings recognize @.

People who see this problem are likely using some form of Unix. You likely have a program called dos2unix on your system. Use that on your files.

Barring that, here's a perl one-liner from the Unix prompt that will do the same thing.

%  perl  -pi~ -e 's/\r//'  *.txt
[download]

If you really need to do this in your own code, the bit between the single quotes is useful to you.

Update: From above replies, it looks like you do know the difference between '^M' and '\r', but I hope the information is still useful for yourself or others.