Implement URLUtils and URL interface

Implement URLUtils and URL interface

This does not implement the URLQuery part of the URL spec yet.

This moves the URLUtils WebIDL attributes from HTMLAnchorElement
to URLUtils and let both URL and HTMLAnchorElement implement
URLUtils.

This allows creating new URL objects using `new URL`, which can
later be mutated using the same property names as provided by
HTMLAnchorElement and Location.

BUG=303152
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=159337

[arv (Not doing code reviews), 2013-10-08 22:31:31]

PTAL

Eric: You mentioned doing this behind a runtime flag but as far as I know we
cannot yet apply runtime flags to "X implements Y" directives. I could of course
duplicate the IDL attribute definitions.

[eseidel, 2013-10-08 22:33:23]

I'm not sure what X implements Y translates to.  Does it just expose the
properties? or does it chagne the prototype chain?

Is it observable from JS?

[arv (Not doing code reviews), 2013-10-08 22:36:40]

On 2013/10/08 22:33:23, eseidel wrote:
> I'm not sure what X implements Y translates to.  Does it just expose the
> properties? or does it chagne the prototype chain?

It generates code that does, in this case:

  DOMURLUtils::setPort(impl, value)

Instead of

  impl->setPort(value)

> Is it observable from JS?

It is not observable from JS.

[eseidel, 2013-10-08 22:49:05]

If it's not observable from JS than I don't care at all about the runtime flag.
:)  My assertion in the thread about using a runtime flag was about guarding
from unintentionally exposed JS changes during feature development, so that we
can intentionally change our JS api all at once.

[arv (Not doing code reviews), 2013-10-09 00:58:22]

On 2013/10/08 22:49:05, eseidel wrote:
> If it's not observable from JS than I don't care at all about the runtime
flag.
> :)  My assertion in the thread about using a runtime flag was about guarding
> from unintentionally exposed JS changes during feature development, so that we
> can intentionally change our JS api all at once.

Just to make it explicit. There is no change to any existing API. (except for
URL.length changing to 1 because it now takes an argument.) The changes in the
existing tests are actually due to https://codereview.chromium.org/26506003/

[arv (Not doing code reviews), 2013-10-09 21:29:49]

This is now as polished as I can get it.

[abarth-chromium, 2013-10-10 04:11:55]

Can you put these tests in fast/domurl ?  There's no need to put them in the DOM
directory or to use ALLCAPSYELLING

[abarth-chromium, 2013-10-10 04:17:34]

LGTM

We can't share this code with Location too?

https://codereview.chromium.org/24066010/diff/18039/Source/core/html/DOMURL.h
File Source/core/html/DOMURL.h (right):

https://codereview.chromium.org/24066010/diff/18039/Source/core/html/DOMURL.h...
Source/core/html/DOMURL.h:71: virtual void setInput(const String&);
Can you mark these four functions as OVERRIDE?

https://codereview.chromium.org/24066010/diff/18039/Source/core/html/HTMLAnch...
File Source/core/html/HTMLAnchorElement.h (right):

https://codereview.chromium.org/24066010/diff/18039/Source/core/html/HTMLAnch...
Source/core/html/HTMLAnchorElement.h:73: virtual void setInput(const String&);
OVERRIDE

https://codereview.chromium.org/24066010/diff/18039/Source/core/html/HTMLAnch...
File Source/core/html/HTMLAnchorElement.idl (left):

https://codereview.chromium.org/24066010/diff/18039/Source/core/html/HTMLAnch...
Source/core/html/HTMLAnchorElement.idl:25: [Reflect, TreatNullAs=NullString,
URL] attribute DOMString href;
We're losing Reflect here, but I guess we're just implementing that manually
now.

[commit-bot: I haz the power, 2013-10-10 04:17:54]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/arv@chromium.org/24066010/18039

[arv (Not doing code reviews), 2013-10-10 12:57:27]

On 2013/10/10 04:11:55, abarth wrote:
> Can you put these tests in fast/domurl ?  There's no need to put them in the
DOM
> directory or to use ALLCAPSYELLING

Done.

[arv (Not doing code reviews), 2013-10-10 13:03:12]

I was planning on handling Location and WorkerLocation next. 

Location uses some extended attributes that do not work in a worker so I'll have
to look into finding some other way to get the activeWindow and firstWindow
through.

https://codereview.chromium.org/24066010/diff/18039/Source/core/html/DOMURL.h
File Source/core/html/DOMURL.h (right):

https://codereview.chromium.org/24066010/diff/18039/Source/core/html/DOMURL.h...
Source/core/html/DOMURL.h:71: virtual void setInput(const String&);
On 2013/10/10 04:17:34, abarth wrote:
> Can you mark these four functions as OVERRIDE?

Done.

https://codereview.chromium.org/24066010/diff/18039/Source/core/html/HTMLAnch...
File Source/core/html/HTMLAnchorElement.h (right):

https://codereview.chromium.org/24066010/diff/18039/Source/core/html/HTMLAnch...
Source/core/html/HTMLAnchorElement.h:73: virtual void setInput(const String&);
On 2013/10/10 04:17:34, abarth wrote:
> OVERRIDE

Done.

https://codereview.chromium.org/24066010/diff/18039/Source/core/html/HTMLAnch...
File Source/core/html/HTMLAnchorElement.idl (left):

https://codereview.chromium.org/24066010/diff/18039/Source/core/html/HTMLAnch...
Source/core/html/HTMLAnchorElement.idl:25: [Reflect, TreatNullAs=NullString,
URL] attribute DOMString href;
On 2013/10/10 04:17:34, abarth wrote:
> We're losing Reflect here, but I guess we're just implementing that manually
> now.

Yup

[commit-bot: I haz the power, 2013-10-10 13:16:55]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/arv@chromium.org/24066010/37001

[commit-bot: I haz the power, 2013-10-10 14:46:13]

Change committed as 159337