[whatwg] On tag inference

Henri Sivonen wrote:
> On Sep 9, 2005, at 22:22, Blake Kaplan wrote:
>> For compatibility,
> 
> With IE?

With both IE and previous Geckos.

> How can using them with XBL be necessary for compatibility with IE 
> considering that IE does not have XBL? Do those who use XBL with custom 
> elements use something else in IE (HTC perhaps)?

I don't have any numbers or anything. See
https://bugzilla.mozilla.org/show_bug.cgi?id=167493 for the bug when 
this was introduced (note that we've had this behavior since mid 2003).

> 
>> Why should this change,
> 
> Actually, allowing custom stuff in head is a change from Firefox 1.0.x.

No, it isn't. If you omit the <head> tag, then we'll default to putting 
userdefined tags in the body (this is what the fix that I made for 
https://bugzilla.mozilla.org/show_bug.cgi?id=307122), but the testcase: 
data:text/html,<head><foobar><body><baz> gives me the content model:
HEAD
- foobar
BODY
- baz

In both Firefox 1.0.x and current trunk builds (modulo recently 
introduced regressions!).

>> especially considering that IE allows this?
> 
> Opera and Safari (like Firefox 1.0.x) put custom elements in body.

Following Opera and Safari (while notable for interoperability, and for 
new behavior, I'm open to it) is no reason to break compatibility over 
this sort of thing.
-- 
Blake Kaplan

Received on Friday, 9 September 2005 14:19:22 UTC