Remove url fragments from host name

[jhnns]

When working with tld.js I thought it might be useful if tld.js would also work with host names containing url fragments:

tld.getDomain("http://fr.google.com/some/path")
// should return 'google.com' imho
tld.tldExists('https://user:password@example.co.uk:8080/some/path?and&query#hash');
// should return `true`
tld.getSubdomain('https://moar.foo.google.co.uk');
// should return `moar.foo`

The pull-request contains tests for this feature and updates the README accordingly.

[thom4parisot]

Hey thanks!

It addresses an issue raised by #38. I wonder why you do not use directly the require('url').parse native Node's function?

[jhnns]

Yeah, I could also use require('url').parse. I thought that we don't need all the other features of the url-module and thus only need a regex which captures the relevant parts. I think the regex is safe to use, I used it in other projects too.

But if you feel more comfortable with the url-module, I can change that 😄

[jhnns]

I wonder if that extra logic should be the responsibility of the consumer or the library.

I thought about this question too. Usually I'm a big fan of small libs which do one thing well. But I also like clear and simple APIs and imho it's ok to provide a little bit extra logic if I can save some lines in the application. It's definitely not an uncommon use-case to extract the domain portion of an url.

But I guess this is a matter of taste. 😉

BTW: Travis CI seems to fail because of node 0.8 incompatible dependencies.

[thom4parisot]

@jhnns okay got it :-) I wanted to hear more about that before merging.

Rebase from master, I have removed Node 0.8 compatibility and updated modules anyway.

[jhnns]

Ok, ready to merge

[jhnns]

👍