107904 - Problems with XHTML as text/html in HTML tokenizer

Status: RESOLVED WONTFIX

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

Description

[Christopher Hoess (gone)]

From conversation with harishd on IRC tonight, CParserContext::SetMimeType()
isn't setting the XML parsing flag for XHTML served as text/html, which it will
eventually need to parse XML-style empty tags properly.

Solution is probably to modify doctype-sniffing code upstream of tokenizer so
that XHTML DTDs set both the strict and XML flags. (Actually, layout is handling
text/html XHTML as quirks right now, but that's another bug).

Comment 1

[Christopher Hoess (gone)]

Looks like the most likely solution would be to hack nsParser.cpp around line
1105 so documents with an XHTML doctype are set to, say, aDocType = eXHTML. In
"if (aDocType == eXML)", s/XML/XHTML/, and the tokenizer should work correctly.
Parse Mode will still be set to strict; however, I'm not sure exactly how the
parser decides to use HTML or XML tokenizer, so I don't know how this would
affect that mechanism.

Comment 2

[harishd]

mass move to 1.1

Comment 3

[Jonas Jørgensen]

Pardon my ignorance, but is this bug requesting that XHTML documents served as
text/html be parsed as XML (as they would if they were served with an XML mime
type)?

Comment 4

[Christopher Hoess (gone)]

I should hope not! This is in reference to XML-style empty elements being
incompatible with strict SGML parsing: see
<URL:http://www.cs.tut.fi/~jkorpela/html/empty.html>.

Comment 5

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

[was 1.1alpha]

Comment 6

[Hixie (not reading bugmail)]

According to the HTMLWG, XHTML sent as text/html must not be parsed any
differently than HTML sent as text/html. So this would be WONTFIX, no?

I wish we implemented the NET feature...

Comment 7

[Boris Zbarsky [:bzbarsky]]

I'm going to go out on a limb here and wontfix this.