Aha -- that does it. I like the approach much better (matching non-backslash and colon or 2-character escape sequences).

Running some quick test code:

    my $re = shift or die "re expected\n";
    while (<>) {
        chomp;
        my ($field, $value) = ($_ =~ $re);
        print "'$_' -> ('$field', '$value')\n";
    }
[download]

We see that it handles all these cases correctly:

 'a:b:c' -> ('a', 'b:c')
 'a\:b:c' -> ('a\:b', 'c')
 'a\\:b:c' -> ('a\\', 'b:c')
 'a\\\:b:c' -> ('a\\\:b', 'c')
 'a:' -> ('a', '')
 ':b' -> ('', 'b')
[download]

The only thing it doesn't handle perfectly is the invalid input case with an escaped colon but no delimiter:

 
 'a\:b' -> ('', 'b')
[download]

Notice that this is indistinguishable from the ':b' input. The fix is simple though, just anchor the start of the match:

 ^((?:[^:\\]+|\\.)*):(.*)
[download]

For my use this works better, as the match fails, telling me there was no field delimiter present.

Thanks
--mike