235768 - mozilla 1.6 and up don't always display fixed background correctly. moz 1.4 was fine

Status: RESOLVED FIXED

(Core :: Web Painting, defect, P1)

Description

[Jonathan Brown]

User-Agent:       
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.5) Gecko/20040114 Firebird/0.7

On the page http://emergence.uk.net/about/ - you cannot see the smoke and
railtrails through the main content section with moz 1.6. moz 1.4 is fine.

Reproducible: Always
Steps to Reproduce:
1.go to http://emergence.uk.net/about/
2.look at it
3.feel sad

Actual Results:  
i felt more and more sad

Expected Results:  
displayed the background image under the text

Comment 1

[Boris Zbarsky [:bzbarsky]]

The problem here is that fixed-attachment backgrounds should just be placed
relative to the viewport, per CSS2.1, without the "nearest scrolling ancestor"
stuff we do.

Comment 2

[Boris Zbarsky [:bzbarsky]]

Attachment: Maybe like this?

Comment 3

[Boris Zbarsky [:bzbarsky]]

Attachment: Same as diff -w

Comment 4

[Boris Zbarsky [:bzbarsky]]

Comment on attachment 142393 [details] [diff] [review]
Same as diff -w

roc, what do you think?

Comment 5

[Robert O'Callahan (:roc) (email my personal email if necessary)]

I think Hixie was planning to change CSS, or something. At least I blame him for
advocating the "nearest scrollframe" behaviour.

Comment 6

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Comment on attachment 142393 [details] [diff] [review]
Same as diff -w

This looks like a good patch to make us CSS2.1 compliant.

Comment 7

[Boris Zbarsky [:bzbarsky]]

Heh.  Well, the nearest scrollframe changed when we changed the impl of
overflow:hidden....

Taking bug.

Comment 8

[Boris Zbarsky [:bzbarsky]]

Checked in.  I also filed bug 235788 on some perf issues I was seeing on that page.

Comment 9

[Jonathan Brown]

Thanks guys!

Comment 10

[Hixie (not reading bugmail)]

(In reply to comment #5)
> I think Hixie was planning to change CSS, or something. At least I blame him 
> for advocating the "nearest scrollframe" behaviour.

That rings a bell. Seems I lost the argument in the WG though. :-)