Question

Is it possible to write this without the [^...] but with using the \P{...}?

#!/usr/bin/env perl 
use warnings;
use 5.012;
use utf8;

my $string = '_${Hello}?${World}!';

$string =~ s/[^\p{Alphabetic}\p{Mark}\p{Decimal_Number}\p{Connector_Punctuation}]/-/g; 

say "<$string>";

Answer 1

Well, it’s possible but I don’t think I’d call it an improvement:

#!/usr/bin/env perl
use warnings;
use 5.012;
use utf8;

my $string = '_${Hello}?${World}!';

$string =~ s/(?=\P{Alphabetic})
             (?=\P{Mark})
             (?=\P{Decimal_Number})
             (?=\P{Connector_Punctuation}) . /-/xgs;

say "<$string>";

With multiple positive lookaheads, they all have to succeed. So it matches one character (the .) that is not Alphabetic and not Mark and not Decimal_Number and not Connector_Punctuation, just like the negated character class would.

I added the /s modifier because the original regex would match a newline (although your sample string doesn’t have one). I added /x so I could add some whitespace and break it over multiple lines.

What do you have against character classes, anyway?