regexperl

What does the caret ^ in (?^:…) mean in the string form of a Perl qr// Regex?

Consider this script, which is based on an answer to SO 267399 about parsing Roman numbers, though the parsing of Roman numbers is incidental to this question.

#!/usr/bin/env perl
#
# Based on answer to SO 0026-7399

use warnings;
use strict;

my $qr1 = qr/(?i:M{1,3})/;
my $qr2 = qr/(?i:C[MD]|D?C{1,3})/;
my $qr3 = qr/(?i:X[CL]|L?X{1,3})/;
my $qr4 = qr/(?i:I[XV]|V?I{1,3})/;

print "1000s: $qr1\n";
print " 100s: $qr2\n";
print "  10s: $qr3\n";
print "   1s: $qr4\n";

# This $qr is too simple — it matches the empty string
#my $qr = qr/($qr1?$qr2?$qr3?$qr4?)/;

my $qr = qr/\b((?:$qr1$qr2?$qr3?$qr4?)|(?:$qr2$qr3?$qr4?)|(?:$qr3$qr4?)|(?:$qr4))\b/;

print " Full: $qr\n";

while (<>)
{
    chomp;
    print " Line: [$_]\n";
    while ($_ =~ m/$qr/g)
    {
        print "Match: [$1] found in [$_] using qr//\n";
    }
}

Given the data file below, the first three lines each contain a Roman number.

mix in here
no mix in here
mmmcmlxxxix
minimum

When run with (home-built) Perl 5.22.0 on a Mac now running macOS Sierra 10.12.4, I get output like this (but the version of Perl is not critical):

1000s: (?^:(?i:M{1,3}))
 100s: (?^:(?i:C[MD]|D?C{1,3}))
  10s: (?^:(?i:X[CL]|L?X{1,3}))
   1s: (?^:(?i:I[XV]|V?I{1,3}))
 Full: (?^:\b((?:(?^:(?i:M{1,3}))(?^:(?i:C[MD]|D?C{1,3}))?(?^:(?i:X[CL]|L?X{1,3}))?(?^:(?i:I[XV]|V?I{1,3}))?)|(?:(?^:(?i:C[MD]|D?C{1,3}))(?^:(?i:X[CL]|L?X{1,3}))?(?^:(?i:I[XV]|V?I{1,3}))?)|(?:(?^:(?i:X[CL]|L?X{1,3}))(?^:(?i:I[XV]|V?I{1,3}))?)|(?:(?^:(?i:I[XV]|V?I{1,3}))))\b)
 Line: [mix in here]
Match: [mix] found in [mix in here] using qr//
 Line: [no mix in here]
Match: [mix] found in [no mix in here] using qr//
 Line: [mmmcmlxxxix]
Match: [mmmcmlxxxix] found in [mmmcmlxxxix] using qr//
 Line: [minimum]

The only part of the output that I don't understand is the caret ^ in the (?^:…) notation.

I've looked at Perl documentation for perlre and perlref and even the section of perlop on 'Regex quote-like operators' without seeing this exemplified or explained. (I also checked the resources suggested by SO when you ask a question about regexes. The (?^: string is carefully designed to give search engines conniptions.)

There are two parts to my question:

  1. What is the significance of the caret in (?^:…) and what caused it to be added to the qr// regexes?
  2. If it matters, how do you stop it from being added to the qr// regexes?
Solution

  • Basically it means the default flags apply (even if it gets interpolated into a regex that specifies differently). Before it was introduced, qr would produce something like (?-ismx: and a new flag being added to Perl would make that change, which made keeping tests up to date a pain.

    http://perldoc.perl.org/perlre.html#Extended-Patterns:

    Starting in Perl 5.14, a "^" (caret or circumflex accent) immediately after the "?" is a shorthand equivalent to d-imnsx . Flags (except "d" ) may follow the caret to override it. But a minus sign is not legal with it.