remove non alpha and numeric chars+more?

Anon

hi!

i would like to remove all non alpha and numeric chars and some more like:

$string = ereg_replace('[^{0-9a-zA-Z]','',$string);}

works ok for 0-9 a-z A-Z but how do i include !"#$%&'()*+,-./<=>?[]^_' in the string too?

so i want to allow all those mentioned above

cant quite figure it out how to write the "code", since some of the chars are ereg "expressions".

code example appreciated.

best regards: (newbie)
/BN

van__

[^{0-9a-zA-Z!"#\$%&\'()*+,-.\/<=>\?[\]^_\s]}

Anon

thanks for the fast answer, but i quite can't get it to work.

this is how i put it:
$string = ereg_replace('[^{0-9a-zA-Z!"#\$%&\'()*+,-.\/<=>\?[\]^_\s]','',$string);}

and then i just "echo $string;" on the next row. I get $string trough a html-POST variable..

when i enter for example xy~abc123 (tilde) in the input box, the ~ (tilde) will echo with the example above.. but if i run $string = ereg_replace('[^{0-9a-z]','',$string);} it will not. apprently ~ (tilde) is not included in the (longer)string far up, but why does it echo.
am i missing a point here?

i tried with the /:s before i posted the question, but got the same result then too

Best Regards & Thanks!:
/BN

van__

borje wrote:

am i missing a point here?

No, actually you fond a bug with POSIX extensions in PHP posix-extended regex.

Appearantly, <> matches &tilde;
That is not good.

But, fine for me, because I always though ereg sucked ass anyway.

I always use preg.
Before it was just because it was substantially faster executing.

But now, it will be because it's got less bugs too.

<?
$string = 'xy~abc123';
$string = preg_replace('/[^{0-9a-zA-Z!#\$%&.\?*,[]\'\"+-<=^>()]+/','',$string);}
echo $string;
?>

Anon

thanks Van!
i think i am getting this sorted out very soon🙂
propably will stick to preg myself too..

thanks again!
/BN