ereg and TLD's

Pig

I'm trying to set up an ereg validation for a URL. Everything is fine, but I cannot seem to get it to accept "anything" for the TLD portion. Whenever I put in .. or .+ or (.*), it accepts anything, but it also stops checking for the . before it. If I have it check a list (us|com|net) it works fine.

Here is what I have:
ereg("^{http://(www.)?.+.(.*)",} $url) //this ought to work, but it doesn't. It accepts both:
http://www.blah.com
http://www.blahcom

ereg("^{http://(www.)?.+.(com|org|net)",} $url) //this works

Thanks much.

ir4z0r

maybe use [^.]+ instead of .*

but aren't urls like "http://ibm.com" or "http://hdd.ibm.com" valid?

Pig

That wouldn't work. the ^ denotes that the characters occurs at the beginning of the string. Also, the format for this ereg takes those kind of URLs. The only issue is with the . between the domain and the tld.

ir4z0r

inside classes [ ], ^ means NOT, if used as first character

so [^c] means: a,b,d,e,f,...

I thought of excluding the . , hence the [^.]+ would contain the domainname

hmm, or do I mix up the Perl/PCRE thingie 😕

Pig

Hmm... I've never heard of that, but I am no expert. Even so, I don't think that would be the trick. I need to make sure there is a ., not prevent one.

Thanks though.

ir4z0r

i tried a little, with preg_*, too but one problem remains:

how do you want to distinguish between

"http://domain.tld" and "http://sub.domaintld" ?

Pig

That's the problem I'm hitting. In theory, you should be able to just add . between the first and second parts. It works fine if you specify that the last part is us, com, net, org, etc.

For some reason as soon as I make the us/com/org part be "anything", it stops checking for that period.

This works that way:
ereg("^{[url]http:///url?.+.(com|org|net)",} $url)

Why the heck wouldn't it work if you just replaced taht other garbage?

ir4z0r

I don't understand the ? completely, but I think it's an option,
which mustn't be contained

this option is left out when matching http://www.blahcom

^[url]http://[/url]     --> [url]http://[/url]
(www\.)?   -->
.+               --> www
\.                --> .
(.*)             --> blahcom

and like I said, you just CAN'T distinguish between the 2 cases, so you have to make compromises
like [url]http://[/url](a group with a . at the end - once or twice)(2-3 chars, the tld)

sorry, I can't test it right now, due to technical probs

😃

Pig

No no. That is what the ? does. ? means that the preceeding group or character can be included once, or not at all.

The expression will test true if there is zero or one www. and will fail if there is more than one.

http://www.blah.com //true
http://blah.com //true
http://www.www.blah.com //false

ir4z0r

okay, but that's basically what I said 😃

notice that my listing above got nothing behind "(www.)?" ?
the regex matches the string (or vice versa)

this only an answer, no solution 🙁