Is there any reason you happened to anchor your regex with an "end-of-line" $ code? Because that will depend on the definition of a "line" too.

If I'm stripping a single line thing, I'd use \z to anchor to the end of the string.

Similarly, the \n and \r are semantic, referring to this platform's idea of newlines and returns. Specify exact encodings that you're encountering. For sources using ASCII or UTF-8, that's probably \x0A and \x0D.

$text =~ s/[\x0D\x0A]+\z//;
[download]