39642 - Changing HTML attributes does not cause style update

Status: VERIFIED FIXED

(Core :: CSS Parsing and Computation, defect, P2)

Description

[Marc Attinasi]

Changing some HTML attributes (like bgcolor) via the DOM is not causing the 
style to change. This is due to shaver's recent changes that look for attrivutes 
that are relevant to rules - well, many HTML attributes are implicitly mapped to 
CSS and are thus relevant, even though they may not appear in selectors.

I have a patch that fixes this by restricting shaver's changes to XML content 
and not HTML content.

This is critical because many basic HTML attribute changes are not doing 
anything.

Comment 1

[Mike Shaver (:shaver -- probably not reading bugmail closely)]

Yeah, this is sucky.  Let's fix it. =)

Comment 2

[Marc Attinasi]

Attachment: Patch of change to restrict AttributeAffectsStyle to XML elements

Comment 3

[Marc Attinasi]

Marking nsbeta2: this is going to affect a lot of HTML pages. The fix is also 
pretty simple and low-risk. I strongly urge fixing this for beta2.

Comment 4

[Mike Shaver (:shaver -- probably not reading bugmail closely)]

Attachment: Patch to ensure that id and class selectors are properly handled.

Comment 5

[Mike Shaver (:shaver -- probably not reading bugmail closely)]

I think we need my patch, too.  Marc, do you agree?

Comment 6

[Marc Attinasi]

We need shaver's patch too because class and ID selectors will apply to XUL 
elements (since they implement nsIStyledContent), so we need to consider those 
attributes as relevant.

Shaver, for XML elements the ID attribute does not have to be called 'ID' - the 
XML DTD defines what the identifier attribute is, so the patch is not as general 
as it could be. I don't know how to get to the attribute name for the XML 
element's identifier, though :(  Also, according to the CSS spec (5.8.3) class 
selectors are only for HTML documents. We apply them to XML docs too though.

Comment 7

[Marc Attinasi]

Just a nsbeta2 reminder: this is a big problem as it is - changing HTML 
attributes via script will not cause the element's style to be updated.

Comment 8

[Marc Attinasi]

Adding dogfood kwd. 

This is a pretty smelly problem that prevents many DHTML pages from being 
rendered properly (any that change HTML presentational attributes, like bgcolor, 
background, align to name a few).

Comment 9

[Marc Attinasi]

This bug needs some attention! 

It is a recent regression, we have a fix, and this affects all kinds of 
regression tests (ones that modify HTML attributes), and it will break 99% of 
DHTML pages on the web.

What does a poor engineer have to do to get an important fix approved? I'm 
cc'ing everybody I can think of to get some visibility...

Comment 10

[ekrock's old account (dead)]

PDT team -- I recommend [nsbeta2][6/15]. The attributes are HTML 3.2, and they 
should be settable from JavaScript, with effect, through DOM1. 

Marc -- I think the PDT team has probably just fallen behind on the nominations. 
Thanks for pushing to get this in! 

Comment 11

[leger]

Putting on [nsbeta2+][6/15] radar.

Comment 12

[Marc Attinasi]

I checked in the AttributeAffectsStyle change. 

Shaver, can you check in your patch now so we can close this puppy?

Comment 13

[Mike Shaver (:shaver -- probably not reading bugmail closely)]

I'm on it, brendan/waterson willing.

Comment 14

[Mike Shaver (:shaver -- probably not reading bugmail closely)]

In the tree.

Comment 15

[ckritzer (gone)]

Related/dup of bug 40178?

Marc, Mike: setting the bgcolor via JavaScript doesn't work, as of the 2000-06-
01-20/21-M16 builds.  Setting the background via similar methods does appear to
work.

Check out:
http://voodoolady/bugzilla/nobgcolor.html for bgcolor testcase
http://voodoolady/bugzilla/nobackground.html for background testcase

Re-opening.

Comment 16

[ckritzer (gone)]

Argh.
Bad HTML on my part should *not* constitute an emergency on your part.

The bgcolor file I pointed you guys to had a typo.  Pardon my zeal in re-opening.

Marking VERIFIED FIXED on:
- MacOS9 2000-06-01-20-M16 Commercial Build
- Linux6 2000-06-01-20-M16 Commercial Build
- Win98  2000-06-01-20-M16 Commercial Build

Comment 17

[ckritzer (gone)]

Marking VERIFIED (see previous comments)

Comment 18

[ckritzer (gone)]

*** Bug 40178 has been marked as a duplicate of this bug. ***