Pattern Matching II

[Anon]

This statement is always returning true, even I have nothing in $username or just basic characters. Can anyone help me here?

if (ereg("\~|`|!|@|#|\$|\%|^|\&|*|(|)|+|\=|{|}|[|]|\;|:|\"|\'|\/|\?|>|.|\,|<", $username)){
$uerror = 1;
}

[Anon]

First, an editorial comment : Ay yi yi! (I'm not getting on your case, it's an easy enough sort of trap to fall into, which I can say having done it myself!) But oy veh, is that hard to read!

Instead of using that mess of alternations, use a character class.

Better yet, since it seems your aim is to exclude everything that's neither an underscore, space, or alphanumeric, why not say what you want and say if it finds anything that's not in that set, reject it.

so

if (ereg("[^{[:alnum:]{:space:]_]",} $username)
$uerror=1;

The caret ^ at the beginning of the character class says "not", roughly, so the condition reads

if $username contains anything that's not a space, alphanumeric, or underscore character ...

(you don't need the braces around "$uerror=1;" because it's only one line.)

I hope that helps.

AC

[Anon]

I don't know why, but PHP seems to get in trouble parsing "\$" (this might be because PHP is able to parse variables within strings if enclosed in "s). To avoid this, just write 's instead of the "s like this:
if(ereg('\~|`|!|@|#|\$|\%|^|\&|*|(|)|+|\=|{|}|[|]|\;|:|\"|\'|\/|\?|>|.|\,|<', $username)){......

Michael aka Unique
PS: does anybody know why the following doesn't work:
if(ereg("^{([a-zA-Z0-9-]+)",} $username)){ dosomething(); }
the if-statement should go into the block when $username contains other characters than a-z, A-Z, 0-9, the underscore and the hyphen -

[Anon]

Hi Adam,
as you can see I tried something similar to your suggestion. I found out that
if(ereg("[^{[a-zA-Z0-9-]+|]",} $username))
seems to work, but
if(ereg("^(a-zA-Z0-9-]+)", $username))
don't. Seems to me as the caret has to be enclosed by brackets [ ] but without the pipe, it doesn't work either. As I currently cannot figure out why, can you tell me why?
Thanks in advance,
Michael aka Unique

[Anon]

The caret has two meanings -- outside of a character class, it anchors the pattern to the start of the string, so "^[a-zA-Z]" means "match any alphabetic character at the start of the string. INSIDE a character class, it takes on the "negation" meaning, i.e.
"[^a-zA-Z]" matches any character that is not alphabetic.

HTH,

AC

[Anon]

Yea, this helps. It's a good explanation of why I had to use the pipe and so on... thanks a lot!
Michael aka Unique