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

I'll try to get straight to the problem: 
$attempt1 = "get_stuff()";
$attempt2 = "get_stuff\(\)";
$attempt3 = qq(get_stuff());

$newstr = "apple";

$data = $attempt1;
$data =~ s/$attempt1/$newstr/;
print "\n$data";

$data = $attempt2;
$data =~ s/$attempt2/$newstr/;
print "\n$data";

$data = $attempt3;
$data =~ s/$attempt3/$newstr/;
print "\n$data";

[download]
Output:
apple()
apple()
apple()

 Problem: I'm trying to search and replace the entire contents of the $attempt strings above, but it always stops at the (). Is there any elegant way to do this? I've looked all over my little back book of perl :(


Thanks,
Desert coder

Replies are listed 'Best First'.
Re: Quickie Question: search and replace problem
by indigo (Scribe) on Apr 19, 2001 at 00:19 UTC
    Basically, $data is intepreted as a string, and $attempt1 as a regex, so if you have special regex characters, they will never match up, no matter how you escape them.

    To make a regex treat $attempt1 as a string, use \Q and \E: 
    $data =~ s/\Q$attempt1\E/$newstr/

    [download]
    man perlre for more info.

    Updated: Added rant, as per tye :)
[reply]
[d/l]

      I've created a monster! ;)

      Actually, you can eventually quote them enough to get the same effect as \Q and \E. I personally like the /\Q$string\E/ construct and recommend it over the other methods I'm about to mention.

      But you can also do any of these: 

      $attempt= "get_stuff\\(\\)";
$attempt= "\Qget_stuff()\E";
$attempt= quotemeta( "get_stuff()" );
$attempt= qr/get_stuff\(\)/;
$attempt= 'get_stuff\(\)';

      [download]
      but I still find the original suggestion makes the most sense to me. In a regex (that means only the first half of s///), $x means interpret the contents of $x as a regex while \Q$x\E means interpret the contents of $x as a string.

      Update: Thanks, indigo, I missed that!

              - tye (but my friends call me "Tye")
[reply]
[d/l]
        All these do the escaping properly. But the code went like this: 
        $data = $attempt;
$data =~ s/$attempt/$newstr/;

        [download]
        The problem isn't simply bad escaping. Either $attempt gets munged, and the match fails, or you have to premunge $data to get $attempt right, and the match still fails.

        So /\Q$attempt\E/ really is the best way to go.
[reply]
[d/l]
[select]
      Thanks! Desert coder
[reply]

Back to Seekers of Perl Wisdom