HTMLness of the createDocument() return value should be false regardless of doctype

NEW
Unassigned

Status

()

Core
DOM: Core & HTML
8 years ago
a year ago

People

(Reporter: hsivonen, Unassigned)

Tracking

(Depends on: 1 bug, {html5})

Trunk
html5
Points:
---

Firefox Tracking Flags

(Not tracked)

Details

(Whiteboard: btpp-backlog, URL)

(Reporter)

Description

8 years ago
The code at
http://hg.mozilla.org/mozilla-central/file/4d5760a4f96f/content/xml/document/src/nsXMLDocument.cpp#l123
is wrong per HTML5
http://www.whatwg.org/specs/web-apps/current-work/#html-documents

HTML5 requires documents created with createDocument not to have the HTMLness bit set. Neither WebKit nor Opera sets the HTMLness bit from doctype.

(Also: HTML5 requires all document objects to implement HTMLDocument and SVGDocument.)

Comment 1

8 years ago
Webkit creates XHTMLDocument only if namespace uri of the root element is XHTML.
That is pretty wrong.
Not sure what Opera does.

Comment 2

8 years ago
And Gecko uses doctype for document creation, because that is what should be used
based on DOM 3 Core.

I don't know why HTML5 doesn't allow creating HTML documents, only XHTML documents.
Per actual DOM this method should always create XMLDocument object, without additional treatment for the doctype. Other browsers works like that (I test Chrome, IE11, Opera-Presto), and only Firefox is still inconsistent here (and for example, can output HTMLDocument or SVGDocument).
https://dom.spec.whatwg.org/#dom-domimplementation-createdocument

I catch only one case, when Opera-Presto return HTMLness, but other browser still don't do this:
<script>
	alert(document.implementation.createDocument("http://www.w3.org/1999/xhtml", "html", null));
</script>
Assignee: nobody → ayg
Status: NEW → ASSIGNED
Whiteboard: btpp-active
I'm not sure if we want to do this without fixing bug 897815.  Making the document not HTML doesn't change much in spec-land, but for us it means you lose all the HTMLDocument-specific properties like .body.
Assignee: ayg → nobody
Status: ASSIGNED → NEW
Depends on: 897815
Whiteboard: btpp-active → btpp-backlog
Or, alternatively, if DOM issue 217 <https://github.com/whatwg/dom/issues/217> is resolved and we implement it, it should also be okay to fix this, because passing an HTML DTD but not the HTML namespace is silly and it probably doesn't matter what we do in that case.

An example of a wpt test that newly fails if we implement this as-is is dom/nodes/Node-isEqualNode.html, which tests that .createHTMLDocument() returns the same as a suitably adjusted .createDocument().
(Try run for reference: https://treeherder.mozilla.org/#/jobs?repo=try&revision=f8ba1feb3c25)
You need to log in before you can comment on or make changes to this bug.