7656 - {css1} Canvas should have root element background

Status: VERIFIED FIXED

(Core :: Layout, defect, P3)

Description

[Christopher W. Curtis]

This was interesting.  My window was too small to properly display the test
and the first rendering was not corrent.  The bottom text appeared at the top
and was garbled with other elements.  Maximizing the window forced it to
render properly, even after making the browser smaller and reloading the test.
However, in all cases of the window being too small, the blue background does
not extend beyond the width of the screen, so that scrolling to the right shows
a grey background instead of blue on the outside.  Mozilla M6.

Comment 1

[Peter Linss]

The background color is a dup of another bug. There are some issues with
overflow here.

Comment 2

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

See also bug 6153 and bug 7400

Comment 3

[troy]

The initial reflow problem is a DUP of bug 7400 so that's how I'm marking this.

The blue background not painting all the way isn't a bug it's intentional and
Kipp and I think that this is correct with the way CSS specifies (although
different from Navigator and IE behavior).

I don't see the overflow issues Peter is referring to

*** This bug has been marked as a duplicate of 7400 ***

Comment 4

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

I'm reopening this, since I'm seeing the overflow issue described at the end of
the original report.  I think this may be something you want to fix, especially
if it applies in more general cases (i.e., no width on BODY or HTML).

What happens is, when the window is small, the HTML element is *smaller* than
the specified width, and scrolling to the side, you get the gray background
instead of the background of the HTML element, which should extend out to the
side.

If this only applies to cases where width is specified, I think it's still a
bug, but then not very serious.

Comment 5

[troy]

Why do you think that the background should extend out to the side? I agree it
would be nice, but what we're doing now seems to be consistent with the way CSS2
has defined things.

The particular issue in question is where should the background of the HTML
element be displayed (that's where the blue comes from).

The CSS2 spec (10.3.3) defines the computed width to be based on the containing
block width. That means that the width of the HTML element is the width of the
window only. The fact that the BODY element overflows doesn't change the width
of the HTML element.

Because the 'overflow' property is 'visible', the overflow is displayed and
rendered outside of the HTML box.

14.2 says 'background' refers to the content and padding areas. That means we
don't render the background in the overflow area

Kipp and I were unhappy about this, because it isn't backwards compatible, but
that's how we interpreted the spec

I would like to see an 'overflow' value that specifies that the box should
"expand" (or grow) to encompass its overflow

Comment 6

[leger]

Clearing Duplicate resolution since the bug is now Reopened.

Comment 7

[troy]

Per my explanation in the description log, the blue background shouldn't display
all the way to the right when the window is resized very small and the BODY
content overflows the HTML element's box

That's my interpretation of the spec, and maybe David has a different
interpretation

Comment 8

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

You're right.  For some reason, I thought the boxacidtest had an explicit width
on the HTML element.

Marking verified.

Comment 9

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

I'm reopening this (again) based on the following statement in CSS2, section
12.2 (http://www.w3.org/TR/REC-CSS2/colors.html#q2):

The background of the box generated by the root element covers the entire
canvas.

I think the entire canvas includes horizontal scrolling.  This makes a lot of
sense and is a good thing, since it can cause illegible text otherwise.  (Note
that this will apply to the BODY background when HTML has transparent
background, because of the HTML/BODY background rules...)

Comment 10

[troy]

*** Bug 2977 has been marked as a duplicate of this bug. ***

Comment 11

[troy]

I think you're right that the canvas should include the entire scrolled area.
From 2.3.1:

For all media, the term canvas describes "the space where the formatting
structure is rendered." The canvas is infinite for each dimension of the
space...

Comment 12

[Hixie (not reading bugmail)]

Here is another test case:

   http://www.bath.ac.uk/%7Epy8ieh/internet/eviltests/htmlbodyrendering1.html

I've changed the summary to more accurately match the remaining bug.

Comment 13

[leger]

Clearing Won't Fix.

Comment 14

[Christine Hoffman]

Using 8/5 Apprunner, verified bug fixed.

Comment 15

[Hixie (not reading bugmail)]

Migrating from {css1} to css1 keyword. The {css1}, {css2}, {css3} and {css-moz}
radars should now be considered deprecated in favour of keywords.
I am *really* sorry about the spam...