NAME

Regexp::Compare - partial ordering for regular expressions

SYNOPSIS

  use Regexp::Compare qw(is_less_or_equal);

  if (is_less_or_equal($rx[i], $rx[j])) {
      print "duplicate: $rx[i]\n";
  }

DESCRIPTION

This module implements a function comparing regular expressions: it returns true if all strings matched by the first regexp are also matched by the second. It's meant to be used for optimization of blacklists implemented by regular expressions (like, for example, "http://www.communitywiki.org/cw/BannedContent" ).

Both arguments of "is_less_or_equal" are strings - IOW the call

  $rv = is_less_or_equal($rx, /hardcoded/i);

probably won't do what you want - use

  $rv = is_less_or_equal($rx, '(?i:hardcoded)');

instead.

False return value does not imply that there's a string matched by the first regexp which isn't matched by the second - many regular expressions (i.e. those containing Perl code) are impossible to compare, and this module doesn't even implement all possible comparisons.

BUGS

EBCDIC-based platforms not supported
comparison of character classes is simplified and probably has some incorrect corner cases
comparison fails for locale-specific constructs
comparison fails for regexps with backreferences
global variables affecting regexp matching are ignored
function may die for unusual (legal but unexpected) regexp constructs

AUTHOR

Vaclav Barta, <vbarta@mangrove.cz>

COPYRIGHT AND LICENSE