Not matching REGEX

cutter has asked for the wisdom of the Perl Monks concerning the following question:

This is probably something blindingly simple but I have been knocking myself out for about a day now. Adding some functioning to old code and now a subroutine that validates a user id will not work correctly.

This is the existing code, works fine.

sub validate_owner{

    my ($str) = @_;
    print "Owner is $str in sub_validate_owner\n" if ($debug);
    # Remove leading and trailing whitespace.
    $str =~ s{\A \* | \s* \z}{}gxm;

     if ($str =~ /nbk[A-z0-9]{4}/ or /nbd[A-z0-9]{4}/ or /root/){
        print "Error $str is not a valid job owner\n";
        $error = 1;
        return 1;
        }
        else {     #No match for string.
        return;
        }

}

    my $owner = get_jobname $_;
            print "Owner line is $owner\n" if ($debug);

            # Split id@hostname and get the id value.
            $owner = (split /@/,$owner)[0];
            print "Owner ID is $owner\n" if ($debug);

            validate_owner($owner);
[download]

New code that doesn't work. No change to the sub, just that the value is coming from a hash. The sub will find the first match, but not the nbs or root match in the regex.

sub validate_owner{

    my ($str) = @_;
    print "Owner is $str in sub_validate_owner\n" if ($debug);
    # Remove leading and trailing whitespace.
    $str =~ s{\A \* | \s* \z}{}gxm;

     if ($str =~ /nbk[A-z0-9]{4}/ or /nbd[A-z0-9]{4}/ or /root/){
        print "Error $str is not a valid job owner\n";
        $error = 1;
        return 1;
        }
        else {     #No match for string.
        return;
        }

}
if (defined $jobs{'owner'}){

        my $owner = $jobs{'owner'};

        # Split id@hostname and get the id value.
        $owner = (split /@/,$owner)[0];
        print "Owner ID is $owner\n" if ($debug);

        validate_owner($owner);

    }
[download]

Now for the fun part. If you change the subroutine to this, it works

sub validate_owner{

    my ($str) = @_;
    print "Owner is $str in sub_validate_owner\n" if ($debug);
    # Remove leading and trailing whitespace.
    $str =~ s{\A \* | \s* \z}{}gxm;


    print "Owner is $str in sub_validate_owner\n" if ($debug);

    #
        if ($str =~ /nbk[A-z0-9]{4}/){
        print "Error $str is not a valid job owner\n";
        $error = 1;
        return 1;
        }
if ($str =~ /nbd[A-z0-9]{4}/){
        print "Error $str is not a valid job owner\n";
        $error = 1;
        return 1;
}        
if ($str =~ /root/){
        print "Error $str is not a valid job owner\n";
        $error = 1;
        return 1;
        }
[download]

Comment on Not matching REGEX Select or Download Code

Replies are listed 'Best First'.
Re: Not matching REGEX by FunkyMonk (Bishop) on Aug 30, 2007 at 18:04 UTC
This `if ($str =~ /nbk[A-z0-9]{4}/ or /nbd[A-z0-9]{4}/ or /root/){` [download] is the same as `if ($str =~ /nbk[A-z0-9]{4}/ or $_ =~ /nbd[A-z0-9]{4}/ or $_ =~ / +root/){` [download] which not what I think you're after. Perhaps you meant `if ($str =~ /nbk[A-z0-9]{4}/ or $str =~ /nbd[A-z0-9]{4}/ or $str +=~ /root/){` [download]	[reply] [d/l] [select]
Re^2: Not matching REGEX by cutter (Acolyte) on Aug 30, 2007 at 18:21 UTC
That was it. In the legacy code the $_ held the owner value, so it worked there. Thanks a bunch.	[reply]
Re: Not matching REGEX by ikegami (Patriarch) on Aug 30, 2007 at 18:48 UTC
By the way, `/root/` means "contains 'root'", not "is 'root'". You'll get an undesired error if the user is `astrootter`. `/root/` should be `/^root\z/`. The same problem exists in the other regexps too.	[reply] [d/l] [select]
Re: Not matching REGEX by suaveant (Parson) on Aug 30, 2007 at 18:17 UTC
FunkyMonk points out the problem, I'll point out a minor improvement... You can actually put those all together in one RegEx like so `$str =~ /nb[dk][A-z0-9]{4}\|root/;` [download] Which should be quite a bit more efficient than 3 regexes. - Ant - Some of my best work - (1 2 3)	[reply] [d/l]
Re^2: Not matching REGEX by cutter (Acolyte) on Aug 30, 2007 at 18:40 UTC
Thanks. I'm just happy I got the hash of hashs to work that triggered this problem. Now a improved REGEX, whee.	[reply]
Re: Not matching REGEX by jwkrahn (Abbot) on Aug 30, 2007 at 20:35 UTC
`# Remove leading and trailing whitespace. $str =~ s{\A \* \| \s* \z}{}gxm;` [download] That doesn't actually remove leading whitespace, it removes a leading *`` character. You probably want `s{\A \s+ \| \s+ \z}{}gx` instead. `if ($str =~ /nbk[A-z0-9]{4}/ or /nbd[A-z0-9]{4}/ or /root/){` [download] The character class `[A-z0-9]` also matches the characters [ \ ] ^ _ `** Did you intend to include those characters or did you really mean to use `[A-Za-z0-9]`?	[reply] [d/l] [select]