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

Hi,

The demo: 
#!perl -l
use strict;
use warnings;

my $orig =        'a.dll';
my $replacement = 'b.dll';

my $str1 = "foo $orig bar";
my $str2 = $str1;

$str1 =~ s/\Q$orig\E/\Q$replacement\E/g;
$str2 =~ s/\Q$orig/$replacement/g;

print $str1;
print $str2;

[download]
I expected $str1 and $str2 to be identical after the substitution regexes, but they're not (on both perl-5.16.0 and 5.20.0 at least). For me, the demo outputs: 
foo b\.dll bar
foo b.dll bar

[download]
Why does the first regex put that backslash in there ?
Are we allowed a maximium of *one* \Q per regex ?
Is there documentation that explains this ?

Supposing that instead of the first regex, I had wanted to do: 
$str1 =~ s/\Q$orig\E.+/\Q$replacement\E/g;

[download]
$str1 would then have become foo b\.dll, whereas I expected foo b.dll.

Cheers,
Rob

Replies are listed 'Best First'.
Re: A misunderstanding of quotemeta ?
by LanX (Saint) on Jun 04, 2014 at 02:44 UTC
    quotemeta 
                   Returns the value of EXPR with all non-"word" character
+s
               backslashed.  (That is, all characters not matching
               "/[A-Za-z_0-9]/" will be preceded by a backslash in the
               returned string, regardless of any locale settings.)  T
+his is
               the internal function implementing the "\Q" escape in d
+ouble-
               quoted strings.

    [download]

    So the dot is backslashed.

    Other question, why do you quotemeta the RHS of a s/// ?

    Which problem are you trying to solve?

    Cheers Rolf

    (addicted to the Perl Programming Language)

[reply]
[d/l]
[select]
      Other question, why do you quotemeta the RHS of a s/// ?

      Duh - there's no need to quotemeta the RHS.

      Which problem are you trying to solve?

      The problem of my own idiocy. (It's an ongoing exercise.)
      Thanks, guys.

      Cheers,
      Rob
[reply]
Re: A misunderstanding of quotemeta ?
by Athanasius (Archbishop) on Jun 04, 2014 at 02:52 UTC

    Hello syphilis,

    From quotemeta:

    quotemeta (and \Q ... \E) are useful when interpolating strings into regular expressions, because by default an interpolated variable will be considered a mini-regular expression.

    So, the purpose of using \Q is to prevent “special” characters from having their special (overloaded) meanings within regular expressions. And this is done by backslashing the characters:

    ... all ASCII characters not matching /[A-Za-z_0-9]/ will be preceded by a backslash in the returned string, ...

    Now, in a substitution, only the left-hand side is a regex; the right-hand side is simply the replacement text. So it makes no sense to “de-regexify” the special characters on the right-hand side; they have no special regex meaning in this position anyway. But if you do apply \Q, of course the non-word characters are backslashed, as you are seeing.

    That’s my understanding, anyway. Hope it helps,

    Athanasius <°(((>< contra mundum Iustus alius egestas vitae, eros Piratica,

[reply]
[d/l]
[select]
Re: A misunderstanding of quotemeta ?
by AnomalousMonk (Archbishop) on Jun 04, 2014 at 04:13 UTC

    It may be helpful to consider  \Q...\E quotemeta interpolation in a string-only context, entirely divorced from regexification. These escapes control meta-quoting in any double-quote evaluation, of which a regex is only one instance. 

    c:\@Work\Perl\monks>perl -wMstrict -le
"my $interp = 'a.b';
 my $string = qq{--\Q$interp\E--};
 print qq{'$string'};
 ;;
 my $strung = qq{--$interp--};
 print qq{'$strung'};
"
'--a\.b--'
'--a.b--'

    [download]
[reply]
[d/l]
[select]

Back to Seekers of Perl Wisdom