Remove QualifiedName argument from most HTMLElement::create functions

Remove QualifiedName argument from most HTMLElement::create functions

This patch modifies HTMLElementFactory (and the few other callers of HTML element
creation functions) to pass a tagName only for those elements, like HTMLHeadingElement,
where HTML specifies a single DOM interface for more than one tagName.

R=abarth@chromium.org, eseidel@chromium.org
BUG=316964
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=161742

[esprehn, 2013-11-09 01:43:48]

lgtm

[abarth-chromium, 2013-11-09 01:54:49]

lgtm

https://codereview.chromium.org/66643004/diff/1/Source/build/scripts/make_ele...
File Source/build/scripts/make_element_factory.py (right):

https://codereview.chromium.org/66643004/diff/1/Source/build/scripts/make_ele...
Source/build/scripts/make_element_factory.py:81:
interface_counts[tag['interface']] += 1
There's some fancier Pythony way of doing this.

[adamk, 2013-11-11 20:33:27]

https://codereview.chromium.org/66643004/diff/1/Source/build/scripts/make_ele...
File Source/build/scripts/make_element_factory.py (right):

https://codereview.chromium.org/66643004/diff/1/Source/build/scripts/make_ele...
Source/build/scripts/make_element_factory.py:81:
interface_counts[tag['interface']] += 1
On 2013/11/09 01:54:50, abarth wrote:
> There's some fancier Pythony way of doing this.

Found two ways to do this. defaultdict seems to be the older one, so I used that
(Counter is the new stuff, but I don't need any of its fancy features).

[commit-bot: I haz the power, 2013-11-11 20:39:23]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/adamk@chromium.org/66643004/180001

[adamk, 2013-11-11 22:35:34]

Sadly I think XML is going to kill this patch. Turns out we need to pass the
QualifiedName down so any prefix passed in by the caller is maintained.

I tried to patch over this problem with a hack in createElementNS, but it's
wider-reaching than that, so I'm inclined to just drop this patch for now.

[adamk, 2013-11-11 22:49:17]

Perhaps I spoke too soon: I think I just got the wrong Document::createElement
method. Latest patch passes the tests that failed for me locally. It simply
checks to see if the prefixes match after creating the element, and resets the
QualifiedName if they don't. PTAL.

[abarth-chromium, 2013-11-11 22:56:35]

LGTM

I take it the XMLDocumentParser goes through this codepath?

[adamk, 2013-11-11 22:58:55]

On 2013/11/11 22:56:35, abarth wrote:
> LGTM
> 
> I take it the XMLDocumentParser goes through this codepath?

Yep. I agree with the FIXME above Document::createElement(const QualifiedName&)
that says it ought to be elsewhere.

One other possible fix here is just to call Element::setPrefix. The only
downside there is that it does some redundant checking on the prefix string. But
it'd avoid polluting the Element API. Let me know if you'd prefer that approach.

[commit-bot: I haz the power, 2013-11-11 23:02:09]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/adamk@chromium.org/66643004/240002

[abarth-chromium, 2013-11-11 23:04:05]

This seems fine.

[eseidel, 2013-11-11 23:31:03]

lgtm

This is such space-aged technology compared to where this code came from. 
Amazing.

https://codereview.chromium.org/66643004/diff/240002/Source/core/dom/Element.h
File Source/core/dom/Element.h (right):

https://codereview.chromium.org/66643004/diff/240002/Source/core/dom/Element....
Source/core/dom/Element.h:867: inline void
Element::setTagNameForCreateElementNS(const QualifiedName& tagName)
We could also just call this setPrefix... :)

[adamk, 2013-11-11 23:33:03]

https://codereview.chromium.org/66643004/diff/240002/Source/core/dom/Element.h
File Source/core/dom/Element.h (right):

https://codereview.chromium.org/66643004/diff/240002/Source/core/dom/Element....
Source/core/dom/Element.h:867: inline void
Element::setTagNameForCreateElementNS(const QualifiedName& tagName)
On 2013/11/11 23:31:04, eseidel wrote:
> We could also just call this setPrefix... :)

I agree that the name could use improvement...but can we wait till this lands
via the CQ and fix up naming then?

[adamk, 2013-11-12 00:03:03]

Committed patchset #5 manually as r161742 (presubmit successful).