Bug 159490 – Align Document.body setter with the HTML specification

• Component: DOM
• Status: RESOLVED FIXED

[Chris Dumez]

Document.body setter should be accept a frameset element as input:
- https://html.spec.whatwg.org/multipage/dom.html#dom-document-body

We currently throw an exception when given a frameset element.

Both Firefox and Chrome match the HTML specification.

[Chris Dumez]

Created attachment 282954 [details]
Patch

[Chris Dumez]

Created attachment 282955 [details]
Patch

[Alex Christensen]

Comment on attachment 282955 [details]
Patch

It seems like you should make newBody a Ref or keep the null check.

[Chris Dumez]

(In reply to comment #3)
> Comment on attachment 282955 [details]
> Patch
> 
> It seems like you should make newBody a Ref or keep the null check.

Why?

if (!is<HTMLBodyElement>(newBody.get()) && !is<HTMLFrameSetElement>(newBody.get()))
Will be true if newBody is null and therefore we will throw a HIERARCHY_REQUEST_ERR as expected.

We cannot easily take a Ref<> because this function is used by the bindings and the type is nullable in the IDL. If we updated the type in the bindings to be non-nullable then we would no longer match the spec and we would throw a TypeError instead of a HIERARCHY_REQUEST_ERR.

[Chris Dumez]

Created attachment 282981 [details]
Patch

[Chris Dumez]

(In reply to comment #4)
> (In reply to comment #3)
> > Comment on attachment 282955 [details]
> > Patch
> > 
> > It seems like you should make newBody a Ref or keep the null check.
> 
> Why?
> 
> if (!is<HTMLBodyElement>(newBody.get()) &&
> !is<HTMLFrameSetElement>(newBody.get()))
> Will be true if newBody is null and therefore we will throw a
> HIERARCHY_REQUEST_ERR as expected.
> 
> We cannot easily take a Ref<> because this function is used by the bindings
> and the type is nullable in the IDL. If we updated the type in the bindings
> to be non-nullable then we would no longer match the spec and we would throw
> a TypeError instead of a HIERARCHY_REQUEST_ERR.

I added a test case to cover "document.body = null" to confirm that it still throws a HIERARCHY_REQUEST_ERR.

[Alex Christensen]

Comment on attachment 282981 [details]
Patch

Sure enough.  I should be more familiar with is<>.  r=me
That spec is very specific, and now our code reads like the spec.

[WebKit Commit Bot]

Comment on attachment 282981 [details]
Patch

Clearing flags on attachment: 282981

Committed r202893: <http://trac.webkit.org/changeset/202893>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.