40944 - doctype.systemId always empty in HTML

Status: VERIFIED FIXED

(Core :: DOM: Core & HTML, defect, P3)

Description

[Heikki Toivonen (remove -bugzilla when emailing directly)]

This is not implemented. Even HTML documents can have system id, see for example 
the W3C doc in the URL.

(I am working on the implementation of doctype for XML.)

Comment 1

[Johnny Stenback (:jst)]

This is needed for the editor people to be able to output the doctype when a
document is saved. This should be really easy to fix and the fix should be very
low risk, nominating for beta2.

Comment 2

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

What about that http://www.w3.org/TR/DOM-Level-2/core.html#ID-B63ED1A31 says
that doctype is always null in HTML?

Comment 3

[Johnny Stenback (:jst)]

David, we decided to expose the doctype even for HTML document since the editor
needs to access the doctype info, and since the DOM defines the doctype property
we decided to use that existing interface, even if the DOM spec says it's always
null  for HTML documents (which doesn't make sense to me anyway, it made more
sense in DOM1 since the only thing you could access was the doctype name which
is always "HTML", and entities and notations which are not supported in HTML).

Comment 4

[Hixie (not reading bugmail)]

David: What would we break if we implemented this? I can't think of any code
that would _rely_ on expecting NULL... You can't even use a NULL check to make
sure that it's an HTML document, since some XML documents also return NULL.

(This should not be taken to mean that I'm in favour of us breaking the spec
here, it's just that I'm wondering if it would break anything.)

Comment 5

[Heikki Toivonen (remove -bugzilla when emailing directly)]

I don't understand either why DOM 2 wants HTML always return NULL for doctype. 
doctype is a DOM core property, why would we want to return NULL when we can use 
it for something meaningful? The public id is the only "reliable" method of 
detecting the HTML version in the document, right? In Mozilla's case, it can 
also be checked by scripts to guess if Mozilla is using strict or quirks mode.

Comment 6

[rickg]

The DOM spec doesn't say the doctype is *always* null, it says it will report 
NULL if the doctype was absent in the source document. 

Mozilla will *always* insert a doctype to support the editor (and script 
writers) who care to know how we behave by default. 

approved for nsbeta2.

Comment 7

[Johnny Stenback (:jst)]

I have a fix for this sitting in my tree, I'll need to do some more testing
before checking in.

Comment 8

[Johnny Stenback (:jst)]

The fix for this is now checked in, marking FIXED.

Comment 9

[leger]

Adding verifyme keyword.

Comment 10

[Heikki Toivonen (remove -bugzilla when emailing directly)]

Marking verified.