regexp::assemble and dispatch-table with unicode

[ajan]

Dear Monks,

Im trying to make some regex mainly changing spaces between special words to be smaller spaces (needed for nice pdf printout).

Because there are many cases to match and I have do do it on large textfiles for each paragraph (depending on the individual style there are different sets of replacement) i'm just playing with regex::assemble.

Sadly, the script crashes when i try to replace space with eq Unicode Thin Space chr (0x2009).

Here is my code so far:

#!c:\strawberry\perl\bin\perl

use strict;
use warnings;
use feature qw/:5.10/;
use Regexp::Assemble;

binmode( STDOUT, ":utf8" );

my $ra;

say "SGB".chr(0x2009)."I";   #gives the expected output

my %dispatch = (
    '(SGB)[\s\xa0](I)'     => sub { $ra->mvar(1) . chr(65) . $ra->mvar
+(2) }, #works
#    '(SGB)[\s\xa0](I)'     => sub { $ra->mvar(1) . chr(0x2009) . $ra-
+>mvar(2) }, #crashes
    '(Art\.)[\s\xa0]?(\d)' => sub { $ra->mvar(1) . 'x' . $ra->mvar(2) 
+},
);


$ra = Regexp::Assemble->new( track => 1 )->add( keys %dispatch );

    while (<DATA>) {
        while ( $ra->match($_) ) {
            my $m = $ra->matched;
            s/$m/&{$dispatch{$m}}/e;
        }

                print;
    }

__DATA__
Dies ist ein Test für SGB I und spacing    
Art. 1 zum Beispiel ist hier
es könnte aber auch Art 3 sein

I would be glad if i can get some enlightment from you

Thanks in advance

ajan

Update: just do say it correct, the script does not crash, it hangs (and i suppose it will hang forever...

[ikegami]

U+2009 matches \s, so you go in an infinite loop.

First pass of while loop: "SGB\x{0020}I" gets replaced with "SGB\x{2009}I"
Second pass of while loop: "SGB\x{2009}I" gets replaced with "SGB\x{2009}I"
Third pass of while loop: "SGB\x{2009}I" gets replaced with "SGB\x{2009}I"
...

Sorry, I can't help find you a solution right now. I gotta run.

[ajan]

Uups

I was not aware of that

Sometimes one can really be blind

Thanks, that helps me further

ajan

[ikegami]

RA has no obvious support for global searching. As a quick fix, you could replace \s with something more specific.