Front page | perl.perl5.porters |
Postings from October 2014
On Tue, Oct 14, 2014 at 02:51:19PM -0400, Marco Moreno wrote:
> Using a \Q inside a bracketed character class, should this return true or
> false?
>
> perl -e 'print "foo"=~ /\Q[a-z]\E/ ? "true" : "false"'
>
> It returns "false" for my 5.18.2 and 5.20.0 installs.
You talk about "\Q inside a bracketed character class" but your example
above is actually a bracketed character class inside a \Q. Which one are
you concerned about?
In any case, I would expect that
/\Q[a-z]\E/
is equivalent to
/\[a\-z\]/
(which it is), while I would expect that
/[\Qa-z\E]/
is equivalent to
/[a\-z]/
(which it is), but which seems to contract the documentation you quote:
> However in perlrebackslash under "All the sequences and escapes" it states:
>
>
> Those not usable within a bracketed character class (like [\da-z] ) are
> marked as Not in [].
> \E Turn off \Q, \L and \U processing. Not in [].
> \Q Quote (disable) pattern metacharacters till \E. Not in
> [].
>
> According to these docs, I would have expected \Q and \E to lose their
> meaning and be ignored, thus returning true instead of false. What is the
> proper behavior?
So I think the docs are wrong. For example, the following:
$r = qr/[a\Q]b\E]/;
print $r,"\n";
outputs
...[a\]b]...
at least as far back as 5.6.1
--
More than any other time in history, mankind faces a crossroads. One path
leads to despair and utter hopelessness. The other, to total extinction.
Let us pray we have the wisdom to choose correctly.
-- Woody Allen
Thread Previous
|
Thread Next