URL valid characters. java to validate

Question

a string like: 'www.test.com' is good. a string like: 'www.888.com' is good. a string like: 'stackoverflow.com' is good. a string like: 'GOoGle.Com' is good.

why ? because those are valid urls. it does not necessarely matter if they have been registered or not.

now bad strings are:

'goog*d\x' 'manydots...com'

why because you can't register those urls.

if I have a string in java which is supposed to be a good url what's the best way to validate it ?

thanks a lot

Answer 1

use UrlValidator from the Apache Commons library. Binary package: http://www.mirrorservice.org/sites/ftp.apache.org/commons/validator/binaries/commons-validator-1.3.1.zip (zip contains .jar files)

Example of usage (Construct a UrlValidator with valid schemes of "http", and "https"):

String[] schemes = {"http","https"}.
UrlValidator urlValidator = new UrlValidator(schemes);
if (urlValidator.isValid("ftp://foo.bar.com/")) {
   System.out.println("url is valid");
} else {
   System.out.println("url is invalid");
}

prints "url is invalid"

If instead the default constructor is used.

UrlValidator urlValidator = new UrlValidator();
if (urlValidator.isValid("ftp://foo.bar.com/")) {
   System.out.println("url is valid");
} else {
   System.out.println("url is invalid");
}

prints out "url is valid"

Answer 2

Those examples are hostnames. They're not valid URLs in themselves.

Hostnames are made of .-separated ‘labels’. Each label must be up to 63 characters of letters, digits and hyphens, but a hyphen must not be the first or last character. It is optional to follow the whole hostname with another dot.

You can match this with a pattern like (assuming case-insensitive):

([a-z0-9]|[a-z0-9][a-z0-9\-]{0,61}[a-z0-9])(\.[a-z0-9]|[a-z0-9][a-z0-9\-]{0,61}[a-z0-9])*\.?

However this matches strings like 1.2.3.4 as well, which although they technically could be host/domain names will actually act as direct IP addresses. You may want to allow that. If you do, you may also want to allow IPv6 addresses, which are colon-separated hex; when embedded in a URL, they also have square brackets around them.

And then of course there's IDNA. Nowadays, 例え.テスト is a valid IDNA domain name, corresponding to xn--r8jz45g.xn--zckzah. If you want to allow those you'll need some Unicode support.

Summary: it's quite a bit more difficult than you might think. And that's just hostnames. ‘Validating’ a whole URL is even more work. A simple regex isn't going to hack it. Use a pre-existing library.