117624 - computed value of 'background-position' should be in pixels

Status: RESOLVED INVALID

(Core :: CSS Parsing and Computation, defect, P5)

Description

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

It was decided at the Redwood City F2F of the CSS Working Group (Dec. 2001) that
we would not change the spec to change the computed value of the
'background-position' property, even though we didn't really like the results. 
See http://lists.w3.org/Archives/Member/w3c-css-wg/2001OctDec/0259.html .

This means that our current behavior for inheritance of 'background-position' is
incorrect.  However, the correct behavior is quite difficult to implement since
computing background-position to a pixel value requires knowing both the
computed 'width' of the container and the intrinsic size of the image.

I don't think it's realistic to devote time to fixing it, but I'm filing a bug
on it anyway, mainly so that I can re-resolve bug 91054.

Testcase: attachment 43130 [details].

Comment 1

[fantasai]

How difficult would it be for us to treat 'inherit' as an unrecognized value for 
background-position (i.e. ignore it)?

Comment 2

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

Probably not too hard.

Comment 3

[basic]

status on this? Isn't this a standards compliant issue? And does the summary
need updating?

Comment 4

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

The status on this is that it's an extremely unimportant bug.

Comment 5

[basic]

remove self

Comment 6

[Boris Zbarsky [:bzbarsky]]

caillon, this sounds like something you may be interested in...

Comment 7

[Christopher Aillon (sabbatical, not receiving bugmail)]

Definitely.  I've got an early implementation of this already in my tree, but I
want to finish a few other things (which I probably should file bugs against) first.

Comment 8

[Boris Zbarsky [:bzbarsky]]

This bug is basically in flat contradiction of the current CSS2.1 draft, which
says "Computed value:  	for <length> the absolute value, otherwise a percentage".

Has the CSS WG changed its mind again?

Comment 9

[Hixie (not reading bugmail)]

The computed value of every property got changed when the WG decided to make
computed values not depend on layout, so that UAs could optimise their cascade
and reduce the number of bugs related to "inherit" of length properties.

Comment 10

[Boris Zbarsky [:bzbarsky]]

OK... So this bug is "invalid"?  Or what?

Comment 11

[Hixie (not reading bugmail)]

I guess. The CSS2.1 changes are covered by bug 205790.