[^jpg] is a character class - meaning you match either j, p, org.

Update: and the ^ negates the character class, so [^jpg] matches a single character that is not j, p, or g.

Other than that, I feel your list is incomplete. Why won't you allow .svg images?

And are you aware that it's completely possible to write a script in, say some CGI capable language such as Perl indeed, that would be called my-image-script.(txt|html|bananajuice) that would still generate a valid image in whatever format? But your regexp would disallow it. Is that intentional?

Anyway, this should help you along...

m(
    \[img\]    # Match opening tag
    \s*        # Maybe there are spaces inside
    .*?        # Non-greedy match
    \.         # A dot
    (?:        # Non-capturing group
        jpe?g
      | png
      | svg
      | gif
      | bmp
    )          # End of group
    \s*        # Maybe there are more spaces still
    \[/img\]   # Match closing tag
)x
[download]

Thanks, not quite sure how to use this in the context of replacing occurrences in a block of text.

Basically this doesn't work:

$text =~ s~(\[img\]\s*.*?\.(?:jpe?g|png|svg|gif|bmp)\s*\[/img\])~Inval
+id~isg;
[download]

In these cases it is generally easier to write a regexp which matches only image URLs, and then use the negative match operator !~ instead of the match operator =~.

if ($url !~ m{^.*(jpe?g|gif|xbm|png|bmp)$}) {
  say "$url is not an image";
}
[download]

Thanks but I'm trying to get the instances of this out of a chunk of text though. Like this:

s~m(\[img\]\s*.*?\.(?:jpe?g|png|svg|gif|bmp)\s*\[/img\])x~Invalid~isg;
[download]

but it doesn't work, so I'm clearly not using it correctly.

Ah, OK. You never said it was part of a s/// operator.

Maybe something along these lines?

use 5.010;
use strict;
use warnings;
my $text = do { local $/ = <DATA> };

my ($start, $end) = map quotemeta, qw([img] [/img]);
$text =~ s{($start(.+?)$end)}{
    my $link = $1;
    my $uri  = $2;
    $uri =~ /\.(?:jpe?g|png|svg|gif|bmp)$/ ? $link : 'Invalid'
}eg;

print $text;

__DATA__
Foo
[img]http://example.com/[/img]
Bar
[img]http://example.com/logo.jpeg[/img]
[download]

That said, what you are doing is conceptually broken. It is perfectly valid for an image on the web to have a URL ending with ".cgi", ".php", or even ".html". Whether something is an image or not is decided by its HTTP Content-Type header, not by the last few characters of the URL.

perl -E'sub Monkey::do{say$_,for@_,do{($monkey=[caller(0)]->[3])=~s{::}{ }and$monkey}}"Monkey say"->Monkey::do'
/p

Which matches anything except string ending in jpg within the tags, but when I try to add the "or" bits it goes horribly wrong. Please help, thanks

Sorry, no it doesn't , it forbids a single character, one of g or j or p

Read perlintro, perlrequick, perlfaq6, Parse::BBCode