Validating a Regexp

[PetaMem]

Fellow Monks,

I wonder if there is a way to validate a regexp in Perl. It'd need just a YES/NO functionality. Let's say you have something like a[hjg].* and *a[hj. The routine (my guess is that this could even be another regexp), should identify the first regexp as valid, the second as invalid.

I've been peeking around in the monastery, but only found references to the much bigger (if not impossible) task of validating perl code. But it seems to me, that a plain YES/NO regexp validation should be feasible.

Any clues? TIA

Bye
 PetaMem

[broquaint]

sub is_valid_re {
    eval { qr/$_[0]/ };
    return $@ ? 'NO' : 'YES';
}

my @regexes = qw(a[hjg].* *a[hj);
print is_valid_re($_), $/ for @regexes;

__output__

YES
NO

That seems to do the job.
HTH

_________
broquaint

[Juerd]

sub is_valid_re {
    eval { qr/$_[0]/ };
    return $@ ? 'NO' : 'YES';
}

This eval returns a regex or undef, so you could write as follows:

sub is_valid_re { eval { qr/$_[0] } ? 'YES' : 'NO' }

But I'd _never_ return words when a boolean value is needed. Digits are so much nicer for that. If you want words, you could create yet another sub to turn the truth number into a word.

sub is_valid_re { eval { qr/$_[0] } }
print is_valid_re('abc') ? 'YES' : 'NO';

sub human_bool { shift ? 'YES' : 'NO' }
print human_bool(is_valid_re('abc'));

- Yes, I reinvent wheels.
- Spam: Visit eurotraQ.