Iterating over code embeded in regex

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

I'd like to run a few lines of code on each instance of a string of a particular form. To accomplish this I've been using a regex of the form:

$s =~ /(string)(?{
  print $^N, $/;
})(?!)/;
[download]

While this works to print all occurances of the string in question (see Global Regex sans //g), I cannot seem to get this to work in list (or perhaps array) context:

my @letters = ('a'..'z');
for my $letter (0..$#letters) {
  $letters[$letter] =~ /(?{ print $letter })(?!)/;
}
[download]

The code seems to run each iteration; however instead of printing the current value of $letter, it returns the following error:

Use of uninitialized value in print at (re_eval 1) line 1.

As I type this I wonder if the regex and the embeded code are being compiled prior to the instantion of the variable? I tried running it through Deparse, but it appeared to be more or less unaltered. Should this be the case, I'd surmise it happens afterwards -- perhaps at runtime?

How can I print the current value of the iteration from within the regex -- preferrably sans eval?

Comment on Iterating over code embeded in regex Select or Download Code

Replies are listed 'Best First'.
Re: Iterating over code embeded in regex by ikegami (Patriarch) on Jun 13, 2006 at 03:47 UTC
Don't use lexicals in code embedded in regexp. Use package variables. `my @letters = ('a'..'z'); for our $letter (0..$#letters) { $letters[$letter] =~ /(?{ print $letter })(?!)/; }` [download]	[reply] [d/l]
Re: Iterating over code embeded in regex by Samy_rio (Vicar) on Jun 13, 2006 at 04:01 UTC
Hi eibwen, Try like this, `use strict; use warnings; my @letters = ('a'..'z'); for (0..$#letters) { $letters[$_] =~ m/(?{ print $_ })(?!)/; }` [download] Regards, Velusamy R. eval"print uc\"\\c$_\""for split'','j)@,/6%@0%2,`e@3!-9v2)/@\|6%,53!-9@2~j';	[reply] [d/l] [select]
Re: Iterating over code embeded in regex by ioannis (Abbot) on Jun 13, 2006 at 05:34 UTC
The snippet bellow shows that the $l inside the embedded code of the regex is a closure. `use Scalar::Util 'refaddr'; for my $l ('a'..'b') { print 'loop variable ', refaddr \$l; $l =~ /(?{ print refaddr(\$l) }) /x; }` [download] Here is the output: `loop variable 135603848 135606104 loop variable 135606092 135606104` [download]	[reply] [d/l] [select]
Re: Iterating over code embeded in regex by ambrus (Abbot) on Jun 13, 2006 at 12:31 UTC
I think you shouldn't use the `(?{...})` construct without a very good reason. To run some code on all matches of a regexp in a string, you can do this for example (untested): `while ($string =~ /(pattern)/g) { do_something_with($1); }` [download] Or, if you want overlapping matches too, then `while ($string =~ /(?=(patter))/g) { do_something_with($1); }` [download]	[reply] [d/l] [select]