The easiest way to cope with unpredictable line-termination patterns in data files is to avoid using "chomp", and instead do this:

my $infile = "some_indeterminate_kindof.txt";
open( IN, "<", $infile ) or die "$infile: $!\n";

my @raw_lines = <IN>;
close IN;

s/[\r\n]+$// for ( @raw_lines );  # "chomp" for both dos and unix

my @text_lines = map { split /\r+/ } @raw_lines;  # for (old or oddbal
+l) mac files
[download]

Regarding that comment about "old or oddball" mac files: since macosx is built atop BSD UNIX, it's entirely likely (and normal) for text files on current macs to use LF rather than CR line termination (i.e. when they've been created by perl or unix shell operations). But there are some "legacy style" operations and tools still operating in macosx that may create CR-terminated text files as well.

That snippet should also work as intended in case you happen to get "hybrid" input (e.g. in cases where a single text stream is really a "raw" concatenation of files with different line-termination patterns).