1161752 - Overconstrained values in 'direction: rtl' must ignore and solve for 'left'

Status: RESOLVED DUPLICATE of bug 817406

(Core :: Layout: Positioned, defect)

Description

[Gérard Talbot]

Test
----

http://www.gtalbot.org/BrowserBugsSection/css21testsuite/abs-pos-non-replaced-icb-overconstrain-001.xht

with the embedded file

http://www.gtalbot.org/BrowserBugsSection/css21testsuite/embedded-doc-abs-pos-non-replaced-icb-overconstrain-001.html


Expected result
---------------

http://www.gtalbot.org/BrowserBugsSection/css21testsuite/abs-pos-non-replaced-icb-overconstrain-001-ref.xht

Explanation
-----------

"
The 'direction' property of the initial containing block is the same as for the root element. 
"
10.1 Definition of "containing block"
http://www.w3.org/TR/CSS21/visudet.html#containing-block-details

"
If the values are over-constrained, ignore the value for 'left' (in case the 'direction' property of the containing block is 'rtl') or 'right' (in case 'direction' is 'ltr') and solve for that value.
"
10.3.7 Absolutely positioned, non-replaced elements
http://www.w3.org/TR/CSS21/visudet.html#abs-non-replaced-width


            20% : left
        +
            0px : margin-left
        +
            0px : border-left-width
        +
            0px : padding-left
        +
            20% : width
        +
            0px : padding-right
        +
            0px : border-right-width
        +
            0px : margin-right
        +
            20% : right
        ====================
           100% : width of containing block (width of ICB)


Notes
-----

- There are at least a few (many?) other overconstrained situations not covered by CSS 2.1 test suite which could be found to be incorrectly implemented. The CSS 2.1 test suite does not systematically and methodically test every possible overconstrained situations.
- Chrome 42.0.2311.135, Chrome 44.0.2383.0 and IE11 pass this test.
- Firefox 37.0.2 and Firefox 40.0a1 buildID=20150430100104 fail this test.
- I use Linux 3.13.0-51-generic x86_64, Qt: 4.8.6, KDE 4.13.3; Kubuntu (trusty) 14.04.02
- I've searched for duplicates and did not find any.

Comment 1

[Gérard Talbot]

I intend to submit this test to the CSS2.1 test suite.

Comment 2

[Mats Palmgren (inactive)]

The layout code that handles this overconstrained situation is correct.
It's just that the style data is wrong.  It says the ICB 'direction'
is LTR so we solve for 'right'.  That's bug 817406, IIUC.

Comment 3

[Mats Palmgren (inactive)]

Gérard, do you have a similar test that uses position:fixed?
I'm asking because those take a different code path in Print Preview,
so it would be good to test that too before landing bug 817406.

Comment 4

[Gérard Talbot]

Mats: no, I do not have a similar test that uses position: fixed. 
I am adding this task into my to-do-list; I'll get back to you when such tests are created.

Comment 5

[Gérard Talbot]

> Gérard, do you have a similar test that uses position:fixed?

Is that test okay, suitable for you?

Test
----

http://www.gtalbot.org/BrowserBugsSection/css21testsuite/fixed-pos-non-replaced-icb-overconstrain-001.xht

with the embedded file

http://www.gtalbot.org/BrowserBugsSection/css21testsuite/embedded-doc-fixed-pos-non-replaced-icb-overconstrain-001.html


Expected result
---------------

http://www.gtalbot.org/BrowserBugsSection/css21testsuite/abs-pos-non-replaced-icb-overconstrain-001-ref.xht

Comment 6

[Gérard Talbot]

Here a single-file test (with 'position: fixed'), with proportional values only so that it will work regardless of width of viewport:

http://www.gtalbot.org/BrowserBugsSection/css21testsuite/fixed-pos-non-replaced-icb-overconstrain-009.xht

although this test is CSS3 compliant but not CSS2.1 compliant.

I have printed it with Firefox 37.0.2 and the test failed when printed.

IE11 and Chrome 42.0.2311.152 pass the test when viewing on screen. I think the test failed when trying to print the test in Chrome 42.0.2311.152. I have not tried to print the test with IE11.

Hope this test helps you.

Gérard