Want a faster replace, Please

[SFLEX]

I have been working on my module to speed it up and have gained a lot of speed, but I see there are still some spots I can gain a lot more.

This is the code I want to speed up, I would like to have it run in a loop but those loops seems to like to eat time.

my @AUBBC_TAGS = ('b','i','sub','sup','pre','ul','ol','small','big','l
+i','p','h1','h2','h3','h4','h5','h6','div','strong');

my $message = '[b] stuff [i] stuff [/i] [/b]';
foreach (@AUBBC_TAGS) { $message =~ s{\[(.?)?$_\]}{<$1$_>}xg; }
print $message;

I know this will leave open tags in HTML and I dont realy care. Speed is what Im after here.

The code that I made that was faster is.

my $message = '[b] stuff [i] stuff [/i] [/b]';
$message =~ s{\[(.?)?([\w]+)\]}{<$1$2>}xgo;
print $message;

But that code doesn't work off a list and I would like if it did.

Update 2: Everyones post helped so ++ to all.
Since the array was so small I ended up just doing this.
$message =~ s/\[(\/?(?:big|h[123456]|[ou]?li?|pre|s(?:mall|trong|u[bp])|[bip]))\]/<$1>/gxo;
I guess I will stick with that for now.

[Kanji]

my $AUBBC_TAGS = join '|', @AUBBC_TAGS;

my $message = '[b] stuff [i] stuff [/i] [/b]';

$message =~ s{\[(.?)?($AUBBC_TAGS)\]}{<$1$2>}xgo;
#        or s{\[(/?(?:$AUBBC_TAGS))\]}{<$1>}xgo;

print $message;

[Khen1950fx]

Is this any better?

#!/usr/bin/perl

use strict;
use warnings;

my @AUBBC_TAGS = (
    'b',  'i',     'sub', 'sup', 'pre', 'ul',
    'ol', 'small', 'big', 'li',  'p',   'h1',
    'h2', 'h3',    'h4',  'h5',  'h6',  'div',
    'strong'
);

my $message = '[b] stuff [i] stuff [/i] [/b]';
foreach (@AUBBC_TAGS) {
    $message =~ s/\[(.?)?$_\]/<$1$_>/gx;
}
print $message, "\n";

[JavaFan]

I've looked really hard, but what it the difference between your suggestion and the code the OP posted?

[jwkrahn]

Change:

foreach (@AUBBC_TAGS) { $message =~ s{\[(.?)?$_\]}{<$1$_>}xg; }

To:

$message =~ s{\[(.?)$_\]}{<$1$_>}g foreach @AUBBC_TAGS;

[JavaFan]

Note that your regexp is wrong, it will replace [li] with <i>, as "li" matches /.?i/.

Of course, if it's speed you're after, and not correctness, you maybe able to get away with:

   $message =~ y/[]/<>/;

[Anonymous Monk]

Regexp::Assemble, Regex::PreSuf
Read more... (828 Bytes)

(?-xism:(?:s(?:trong|u[bp]|mall)|h[123456]|b(?:ig)?|p(?:re)?|[ou]l|div
+|l?i))

(?:big|div|h[123456]|li|ol|pre|s(?:mall|trong|u[bp])|ul|[bip])