Walks like a regex, quacks like a regex...

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

Monks ~

I wanted to write a test to see if a given scalar contains a quoted regular expression. I don't care if the regex is valid or what it matches as long as perl thinks it's a real quoted regex. I discovered that it's either much easier than I thought, or I'm missing something. Or both :)

#!/usr/bin/perl
use strict;
use warnings;

my %qrs = (
    string1  => 'foo',
    string2  => '(xism)',
    regex    => qr/foo/,
    'regex?' => '(?-xism:foo)',
);

while( my( $key, $regex ) = each %qrs ) {
    print "'$key' ";
    print $regex =~ /\(\?\-xism:/
        ? 'YES: '
        : 'no: ';
    print $regex;

    print ' (and matches)' if 'foo' =~/$regex/;
    print "\n";
}
[download]

And here's the output:

'regex?' YES: (?-xism:foo) (and matches)
'regex' YES: (?-xism:foo) (and matches)
'string1' no: foo (and matches)
'string2' no: (xism)

It looks like creating a scalar of the pattern "(?-xism:$STUFF)" is enough to make perl believe it's a quoted regex. I've tried to find a definitive in perldoc, man pages, The Camel, etc. answer and haven't.

Is it really this cool and simple?

--
man with no legs, inc.

Comment on Walks like a regex, quacks like a regex... Download Code

Replies are listed 'Best First'.
Re: Walks like a regex, quacks like a regex... by kvale (Monsignor) on Sep 23, 2003 at 03:24 UTC
Well, both `foo` and `(?-xism:$STUFF)` are regular expressions and valid strings. A better way to distinguish the two is to look at the reference: `$reg = qr'foo'; $str = 'foo'; $str2 = '(?-xism:foo)'; $reg_ref = \$reg; $str_ref = \$str; $str2_ref = \$str2; print "$reg -> $reg_ref\n"; print "$str -> $str_ref\n"; print "$str2 -> $str2_ref\n";` [download] yields `(?-xism:foo) -> REF(0x80635ac) foo -> SCALAR(0x80635dc) (?-xism:foo) -> SCALAR(0x8063600)` [download] -Mark	[reply] [d/l] [select]
Re: Re: Walks like a regex, quacks like a regex... by legLess (Hermit) on Sep 23, 2003 at 04:17 UTC
Nice test -- thanks. -- man with no legs, inc.	[reply]
Re: Walks like a regex, quacks like a regex... by BrowserUk (Patriarch) on Sep 23, 2003 at 03:28 UTC
Basically, any string can be used as a regex. Vis. `print 'Matched!' if 'fred' =~ 'fred' Matched!` [download] Or perhaps more bizarrly, `print 'Matched!' if ' ' =~ pack 'V', 538976288 Matched!` [download] Whether an arbitrary string will survive compilation and do anything useful, or what it will do if it does is a harder question to answer. Question: Would a legless duck be a free floating spirit? :) Examine what is said, not who speaks. "Efficiency is intelligent laziness." -David Dunham "When I'm working on a problem, I never think about beauty. I think only how to solve the problem. But when I have finished, if the solution is not beautiful, I know it is wrong." -Richard Buckminster Fuller If I understand your problem, I can solve it! Of course, the same can be said for you.	[reply] [d/l] [select]
Re: Walks like a regex, quacks like a regex... by Abigail-II (Bishop) on Sep 23, 2003 at 08:04 UTC
Perl doesn't do "hmmm, it walks like a regex, it quacks like a regex, so this thing must be a regex". No, Perl is more like "hey, you use this thing as if it's a regex, so it'll be a regex". Strings, numbers, arrays, hashes, anything can be a regex: $ perl -wle 'print "Yes" if 8 =~ @INC' Yes $ [download] It's all in the same line as strings becoming numbers and vica versa, arrays becoming numbers in scalar context, etc. This could lead to bizarre looking code like: $ perl -le 'print "Yes" if %INC =~ %INC' Yes $ [download] Abigail	[reply] [d/l] [select]
Re: Walks like a regex, quacks like a regex... by belg4mit (Prior) on Sep 23, 2003 at 03:25 UTC
Umm technically all of your values are "regexen". `/foo/;` [download] is perfectly legal. So is `qr/foo/;` [download] or for that matter `/qr\/foo\//;` [download] What you seem to be after is regex meta-characters... see perlre `-- I'm not belgian but I play one on TV.`	[reply] [d/l] [select]
Re: Walks like a regex, quacks like a regex... by Roger (Parson) on Sep 23, 2003 at 03:37 UTC
You should also checkout the CPAN module YAPE::Regex, which does regular expression parsing in perl.	[reply]
Re: Walks like a regex, quacks like a regex... by shenme (Priest) on Sep 23, 2003 at 04:58 UTC
Wait, are you asking how to tell if a value is the result of `qr{^[Zz]owie!?$}`. Please don't hit me - aren't you looking for this? `if( ref $regex eq 'Regexp' ) { print "That's an affirmative, base.\n"; }` [download]	[reply] [d/l]
Re: Re: Walks like a regex, quacks like a regex... by Roger (Parson) on Sep 23, 2003 at 05:33 UTC
Hang on, but that only works in the case: `regex1 => qr/foo/` [download] But not in the case: `regex2 => '(?-xism:foo)'` [download] Where regex1 is compiled as a Regexp, and regex2 is compiled as a string. The reference to regex1 yields a Regexp, the reference to regex2 yeilds nothing, so it will not match regex2 which is a valid regular expression.	[reply] [d/l] [select]
Re: Re: Re: Walks like a regex, quacks like a regex... by shenme (Priest) on Sep 23, 2003 at 05:48 UTC
Well that was (is) my confusion at his mentioning the phrase "a real quoted regex". _If_ he means a compiled `qr()` regex then while he's still throwing the result reference value around the `ref()` is the right way to check. But once he forces the qr() reference value to be *stringified*, it'll end up in the "(?-xism:^[Zz]owie!?$)" form and he'll have to recognize that 'unique' format. And then his checking will fail because the leading part shows which options are and are not in effect. Using `qr{^Zzowie!?$}sm` will, when stringified, result in "(?ms-xi:^[Zz]owie!?$)". Zowie!	[reply]
Son of the Bride of Walks like a regex, quacks like a regex... by legLess (Hermit) on Sep 24, 2003 at 03:45 UTC
Re: Son of the Bride of Walks like a regex, quacks like a regex... by diotalevi (Canon) on Sep 24, 2003 at 04:26 UTC
Re: Walks like a regex, quacks like a regex... by legLess (Hermit) on Sep 23, 2003 at 04:17 UTC
So, replying to my own question, I just figured out my disconnect here: cloistering. I'd not used it before, so the "?-xism" didn't look familiar to me. It makes even more sense than I thought. Thanks for the replies. -- man with no legs, inc.	[reply]