Re^2: Regex's, parentheses, and the mysterious ( ??{ } ) operator

[Clovis_Sangrail]

Laurent R: "the $stuff variable is not the same in the two programs:"

I must've done that after I copied one program to the other, but it does not matter, in both programs that area is inside the first pair of parenthesis, and it gets included in the match.

Anonymous Monk: "I think you have one too many $parens..."

I do not know how to say what the problem is any differently. I *want* to use $paren a second time and have it find the *second* separate parenthesized text string within the $stuff variable, and set the $2 regex memory variable equal to it. The regex defined in the variable $paren is set works, it successfully finds the first open '(' and balancing close ')' pair of parenthesis. Because I use it as:

/($paren)/

It sets the matching text to the $1 variable. How come I cannot use it again, via something like:

/($paren)[^()]+($paren)/

The sample $stuff variable does have a second, separate parenthesized text string in it ( "(for a while)" ) how come the same $paren regex does not find that second parenthsized string and set $2 to it?

[Anonymous Monk]

Please try reading that again, because our is our

[Clovis_Sangrail]

Ahh! I have deciphered the meaning of your koan. I get a compilation error when I take the "our" out of the middle of the recursive regex:

$ vi p6.pl
"p6.pl" 36 lines, 925 characters 
#!/opt/perl5.16/bin/perl

use strict;
use warnings;

our $paren = qr/                # Need declared variable with use stri
+ct.
      \(
        (
           [^()]+               # Not parens
         |
           (??{ $paren })       # Another balanced group (not interpol
+ated yet)
        )*
      \)
    /x;                         # 'x' means ignore whitespace, comment
+s.


my $stuff = "On the outside now then (we go( in( and in (&stop)(awhile
+) ( furthe
r ))) but still (here) ) and now (for a while) we are out again.";
$stuff =~ /($paren)[^()]*($paren)/;

print "----------\n";
print "$stuff\n";
print "----------\n";
"p6.pl" 27 lines, 650 characters 
$ ./p6.pl 
Variable "$paren" is not imported at (re_eval 1) line 2.
Global symbol "$paren" requires explicit package name at (re_eval 1) l
+ine 2.
Compilation failed in regexp at ./p6.pl line 14.
$

[Clovis_Sangrail]

If I delete the "use strict;" the program compiles/runs, but still does not match the 2nd regex. I also tried making a duplicate of $paren called $par2, and used it instead of $paren to set $2, and that doesn't work either, $2 is still blank.

[Clovis_Sangrail]

"... our = our"

Yes, I am familiar with the reflexive property of equality. (For all X, X == X.) Can you be a little less enigmatic? Are you saying I cannot use the "our" scope? When I used "my" I created an entirely separate $paren and broke the recursion, and because I have "use strict" I cannot leave the variable undeclared. Perhaps I just cannot use "use strict"?

[AnomalousMonk]

... our is our

I think Anonymonk is concerned about the second appearance of  our $re within the  (??{ our $re }) regex subexpression. Although somewhat unfamiliar, I think this is kosher because our acts as a declaration of a package variable and  $re needs to be declared or pre-declared somehow to be referenced within the regex under full strictures. The following variations work identically with strictures (note no capturing groups within $re):

>perl -wMstrict -le
"my $s = 'x(y) (a(b)) ()() q (a(b)c()(d(e(f)g))h) q';
 ;;
 our $r3;
 $r3 = qr{
   \(
     (?:
       [^()]+
       |
       (??{ $r3 })
     )*
   \)
   }xms;
 ;;
 my @p = $s =~ m{ $r3 }xmsg;
 print qq{'$_'} for @p;
 print '--------';
 ;;
 $s =~ m{ ($r3) [^()]* ($r3) }xms;
 print qq{1 '$1'   2 '$2'};
 print '--------';
"
'(y)'
'(a(b))'
'()'
'()'
'(a(b)c()(d(e(f)g))h)'
--------
1 '(y)'   2 '(a(b))'
--------

And the OPed:

our $r3 = qr{
  ...       
    (??{ our $r3 })
  ...
  }xms;