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

I need to check the name to conform to the following conventions: valid names are comprised of only alphanumeric characters(a-z A-Z and 0-9) plus the characters '-', '_', and '.'. Names must begin with a letter.

someone told me to use this

$name =~ /\A[a-zA-Z][\w.-]*\z/;

[download]

but it treat "s@--__" and "s!!!@@@___" to valid which should be invalide name, casue \w in perl include @ and ! and punction sign, Any hints on this?

Thanks,
--Whatluo

Replies are listed 'Best First'.
Re: how to state this by REL
by pg (Canon) on Nov 09, 2005 at 06:57 UTC

    The regexp is correct. The problem is your string. Use single quote instead of double quote, to avoid interpolation. Try this code:

    {
    my $name = "s@--__";
    print $name, "\n";
    if ($name =~ /\A[a-zA-Z][\w.-]*\z/) {
        print "match\n";
    } else {
        print "not match\n";
    }
}
{
    my $name = "s!!!@@@___";
    print $name, "\n";
    if ($name =~ /\A[a-zA-Z][\w.-]*\z/) {
        print "match\n";
    } else {
        print "not match\n";
    }
}
{
    my $name = 's@--__';
    print $name, "\n";
    if ($name =~ /\A[a-zA-Z][\w.-]*\z/) {
        print "match\n";
    } else {
        print "not match\n";
    }
}
{
    my $name = 's!!!@@@___';
    print $name, "\n";
    if ($name =~ /\A[a-zA-Z][\w.-]*\z/) {
        print "match\n";
    } else {
        print "not match\n";
    }
}

    [download]

    Which prints:

    Possible unintended interpolation of @___ in string at math1.pl line 1
+1.
s-__
match
s!!!@@
not match
s@--__
not match
s!!!@@@___
not match

    [download]
[reply]
[d/l]
[select]
      >The regexp is correct. The problem is your string. Use single 
>quote instead of double quote, to avoid interpolation.

Yes, I got it, Now I use single quote to make it happy, and I 
want to know what's really means as to \w in perl, do it 
include @ ! or not?

Thanks for the info.

      [download]
[reply]
[d/l]
        According to the perlre documentation, it doesn't:
        A "\w" matches a single alphanumeric character (an alphabetic character, or a decimal digit) or "_", not a whole word.

        Arjen

        All that is gold does not glitter...
[reply]
Re: how to state this by REL
by Errto (Vicar) on Nov 09, 2005 at 06:00 UTC
    casue \w in perl include @ and ! and punction sign,

    This isn't true. See the following example, 

    print "no match" unless 's!!!@@@___' =~ /\A[a-zA-Z][\w.-]*\z/;

    [download]
    which prints "no match." \w only matches letters, digits, and underscores. Maybe you could show us an example program where this is happening so we can provide some better advice.
[reply]
[d/l]
Re: how to state this by REL
by reasonablekeith (Deacon) on Nov 09, 2005 at 11:07 UTC
    While the guy who gave you this regex obviously knows what he's doing, and admittedly, it does function according to spec, I'd still change it.

    The simple reason being I had to lookup what \A and \z do. Far more common anchors for matching the beginning and end of strings are ^ and $. Even the documentation describes them as "The two most common anchors".

    so I'd change it to 

     $name =~ /^[a-zA-Z][\w.-]*$/

    [download]
    I might even be tempted to explicitly escape that '.', even though it's not strictly necessary. 
     $name =~ /^[a-zA-Z][\w\.-]*$/

    [download]
    To me, that makes it more obvious that we're matching a dot.

    I wonder what percentage of 'average ability' Perl programmers could tell you the subtle difference between \z and \Z without reference to a manual?

    ---
    my name's not Keith, and I'm not reasonable.
[reply]
[d/l]
[select]
      Howdy!

      I wouldn't change that RE. Since you had to look up \A and \z, did you also look up ^ and $? They don't do what you say they do. \A and \z do exactly what you attribute to ^ and ?.

      Further, escaping . in a character class is gratuitous.

      To elaborate: ^ matches the beginning of the line, while $ matches the end of the line (unless the line ends with a newline, in which case the anchor matches before the newline. \A is always the start of the string. \z is always the end of the string. ^ and $ depend on whether you are using the /m modifier or not.

      TheDamian makes fair arguments for using \A and \z (along with /x, /m, and /s routinely). The core argument is that doing so makes your regexen behave with less surprising behavior all the time. Consistency, and all that.

      yours,
      Michael
[reply]

Back to Seekers of Perl Wisdom