Sync the DOMImplementation interface with the spec

Sync the DOMImplementation interface with the spec

https://dom.spec.whatwg.org/#interface-domimplementation

BUG=460722, 238368, 335871, 429536
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=190897

[philipj_slow, 2015-02-25 17:20:46]

philipj@opera.com changed reviewers:
+ haraken@chromium.org

[philipj_slow, 2015-02-25 17:20:47]

PTAL

[haraken, 2015-02-25 17:25:40]

haraken@chromium.org changed reviewers:
+ jl@opera.com

[haraken, 2015-02-25 17:25:41]

LGTM

https://codereview.chromium.org/960533003/diff/1/Source/core/dom/DOMImplement...
File Source/core/dom/DOMImplementation.idl (right):

https://codereview.chromium.org/960533003/diff/1/Source/core/dom/DOMImplement...
Source/core/dom/DOMImplementation.idl:29: [RaisesException] XMLDocument
createDocument(DOMString? namespaceURI, [TreatNullAs=EmptyString] DOMString
qualifiedName, optional DocumentType? doctype = null);

I'm wondering if we need '= null'. I guess it's just redundant. This is a
question not to your CL but to the spec though.

Jens?

[Jens Widell, 2015-02-25 17:32:56]

https://codereview.chromium.org/960533003/diff/1/Source/core/dom/DOMImplement...
File Source/core/dom/DOMImplementation.idl (right):

https://codereview.chromium.org/960533003/diff/1/Source/core/dom/DOMImplement...
Source/core/dom/DOMImplementation.idl:29: [RaisesException] XMLDocument
createDocument(DOMString? namespaceURI, [TreatNullAs=EmptyString] DOMString
qualifiedName, optional DocumentType? doctype = null);
On 2015/02/25 17:25:41, haraken wrote:
> 
> I'm wondering if we need '= null'. I guess it's just redundant. This is a
> question not to your CL but to the spec though.
> 
> Jens?

"= null" says "treat missing the same as if null was passed". The spec prose
then only needs to explain two cases: "null" and "a DocumentType object".

Without "= null", the spec prose would also need to explain the "missing" case.
(Probably by saying "if missing or null, do X".)

In Blink, we'd need to explicitly implement the "missing" case to behave the
same as null as well, without "= null". Now the implementation just sees a
pointer, always with a valid value, possibly null.

Everybody wins. :-)

[haraken, 2015-02-25 17:38:56]

On 2015/02/25 17:32:56, Jens Widell wrote:
>
https://codereview.chromium.org/960533003/diff/1/Source/core/dom/DOMImplement...
> File Source/core/dom/DOMImplementation.idl (right):
> 
>
https://codereview.chromium.org/960533003/diff/1/Source/core/dom/DOMImplement...
> Source/core/dom/DOMImplementation.idl:29: [RaisesException] XMLDocument
> createDocument(DOMString? namespaceURI, [TreatNullAs=EmptyString] DOMString
> qualifiedName, optional DocumentType? doctype = null);
> On 2015/02/25 17:25:41, haraken wrote:
> > 
> > I'm wondering if we need '= null'. I guess it's just redundant. This is a
> > question not to your CL but to the spec though.
> > 
> > Jens?
> 
> "= null" says "treat missing the same as if null was passed". The spec prose
> then only needs to explain two cases: "null" and "a DocumentType object".
> 
> Without "= null", the spec prose would also need to explain the "missing"
case.
> (Probably by saying "if missing or null, do X".)
> 
> In Blink, we'd need to explicitly implement the "missing" case to behave the
> same as null as well, without "= null". Now the implementation just sees a
> pointer, always with a valid value, possibly null.
> 
> Everybody wins. :-)

Thanks for the clarification!

[philipj_slow, 2015-02-26 01:07:46]

keep createHTMLDocument argument default value

[philipj_slow, 2015-02-26 01:09:40]

Oops, I shouldn't have removed the createHTMLDocument title argument default
value, I forgot that crbug.com/335871 has only been fixed when there's a default
value, so my change caused a test to fail. Reverting back and linking to that
bug.

[philipj_slow, 2015-02-26 03:27:28]

The CQ bit was checked by philipj@opera.com

[philipj_slow, 2015-02-26 03:27:28]

The patchset sent to the CQ was uploaded after l-g-t-m from haraken@chromium.org
Link to the patchset: https://codereview.chromium.org/960533003/#ps20001 (title:
"keep createHTMLDocument argument default value")

[commit-bot: I haz the power, 2015-02-26 03:28:25]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/960533003/20001

[commit-bot: I haz the power, 2015-02-26 09:47:10]

Committed patchset #2 (id:20001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=190897