Bug 163535 – [css-grid] Different width of grid container between initial load and refresh

• Component: Layout and Rendering
• Status: RESOLVED FIXED

[Javier Fernandez]

If you open the following Layout Test [1]:
LayoutTests/fast/css-grid-layout/grid-track-sizing-with-percentages-and-orthogonal-flows.html

The first time the width of the first grid container is 110px, but if you resize the window or reload the page, the width is 120px.

We should investigate why we're getting different results depending on the layout.

I tried to create a reduced test case but it seems that I cannot reproduce it without the JavaScript involved on the test.

I managed to reproduce it inside JS Bin too (resize the output frame to see it):
http://jsbin.com/milawa/edit?html,output

[1] https://trac.webkit.org/browser/trunk/LayoutTests/fast/css-grid-layout/grid-track-sizing-with-percentages-and-orthogonal-flows.html

[Javier Fernandez]

Created attachment 291795 [details]
Patch

[Build Bot]

Comment on attachment 291795 [details]
Patch

Attachment 291795 [details] did not pass mac-ews (mac):
Output: http://webkit-queues.webkit.org/results/2302965

New failing tests:
fast/css-grid-layout/grid-track-sizing-with-orthogonal-flows.html
fast/css-grid-layout/grid-track-sizing-with-percentages-and-orthogonal-flows.html

[Build Bot]

Created attachment 291798 [details]
Archive of layout-test-results from ews103 for mac-yosemite

The attached test failures were seen while running run-webkit-tests on the mac-ews.
Bot: ews103  Port: mac-yosemite  Platform: Mac OS X 10.10.5

[Build Bot]

Comment on attachment 291795 [details]
Patch

Attachment 291795 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.webkit.org/results/2302969

New failing tests:
fast/css-grid-layout/grid-track-sizing-with-orthogonal-flows.html
fast/css-grid-layout/grid-track-sizing-with-percentages-and-orthogonal-flows.html

[Build Bot]

Created attachment 291799 [details]
Archive of layout-test-results from ews105 for mac-yosemite-wk2

The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews.
Bot: ews105  Port: mac-yosemite-wk2  Platform: Mac OS X 10.10.5

[Build Bot]

Comment on attachment 291795 [details]
Patch

Attachment 291795 [details] did not pass mac-debug-ews (mac):
Output: http://webkit-queues.webkit.org/results/2302970

New failing tests:
fast/css-grid-layout/grid-track-sizing-with-orthogonal-flows.html
fast/css-grid-layout/grid-track-sizing-with-percentages-and-orthogonal-flows.html

[Build Bot]

Created attachment 291801 [details]
Archive of layout-test-results from ews117 for mac-yosemite

The attached test failures were seen while running run-webkit-tests on the mac-debug-ews.
Bot: ews117  Port: mac-yosemite  Platform: Mac OS X 10.10.5

[Build Bot]

Comment on attachment 291795 [details]
Patch

Attachment 291795 [details] did not pass ios-sim-ews (ios-simulator-wk2):
Output: http://webkit-queues.webkit.org/results/2302978

New failing tests:
fast/css-grid-layout/grid-track-sizing-with-orthogonal-flows.html
fast/css-grid-layout/grid-track-sizing-with-percentages-and-orthogonal-flows.html

[Build Bot]

Created attachment 291802 [details]
Archive of layout-test-results from ews124 for ios-simulator-wk2

The attached test failures were seen while running run-webkit-tests on the ios-sim-ews.
Bot: ews124  Port: ios-simulator-wk2  Platform: Mac OS X 10.11.6

[Javier Fernandez]

Created attachment 291805 [details]
Patch

Fixed layout tests.

[Manuel Rego Casasnovas]

Comment on attachment 291805 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=291805&action=review

> Source/WebCore/ChangeLog:18
> +        algorithm depend on these override sizes.

Typo: s/depend/depends/

> Source/WebCore/ChangeLog:22
> +        clear these override sizes and performs a layout of the affected grid

Typo: s/performs/perform/

> Source/WebCore/rendering/RenderGrid.cpp:465
> +    // We need to clear both own and containingBlock override sizes to
> +    // ensure we get the same result when grid's intrinsic size is
> +    // computed again in the updateLogicalWidth call bellow.

We do this only for orthogonal items, shouldn't we say that on the comment.

> Source/WebCore/rendering/RenderGrid.cpp:471
> +    // We do cache orthogonal items during the placeItemsOnGrid call, which is
> +    // executed later. However, we are
> +    // only interested on running this logic when we are performing a
> +    // relayout, hence we have already cached
> +    // the orthogonal items.

Mmm, you don't use cached orthogonal items here.
Is this comment accurate?

> LayoutTests/fast/css-grid-layout/repeating-layout-must-produce-the-same-results.html:19
> +.i1 {
> +  color: magenta;
> +  background: cyan;
> +}

Nit: Maybe you can use firstRowFirstColumn class which is more descriptive
and already has a background color.
And you just add here the color for the text.

> LayoutTests/fast/css-grid-layout/repeating-layout-must-produce-the-same-results.html:20
> +.i2 {

Ditto but using firstRowSecondColumn.

> LayoutTests/fast/css-grid-layout/repeating-layout-must-produce-the-same-results.html:21
> +  color: red;

It's weird to use "red" if there's not a fail in a test.

> LayoutTests/fast/css-grid-layout/repeating-layout-must-produce-the-same-results.html:30
> +        document.body.offsetLeft;
> +        document.body.offsetLeft;

Do you need to repeat twice this line?

[Manuel Rego Casasnovas]

For future reference this has been already solved in Blink:
https://codereview.chromium.org/2361373002/

[Javier Fernandez]

Comment on attachment 291805 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=291805&action=review

>> LayoutTests/fast/css-grid-layout/repeating-layout-must-produce-the-same-results.html:30
>> +        document.body.offsetLeft;
> 
> Do you need to repeat twice this line?

Well, perhaps it's too much, but I wanted to be really sure that many layouts doesn't produce different results. I guess we can live with just one extra layout, but I guess we can keep as it is, so we performs the same test than Blink.

[Javier Fernandez]

Created attachment 291810 [details]
Patch

Applied suggsted changes.

[Javier Fernandez]

Created attachment 291814 [details]
Patch

Properly forcing relayout in the test.

[Javier Fernandez]

Created attachment 291829 [details]
Patch

Patch for landing.

[WebKit Commit Bot]

Comment on attachment 291829 [details]
Patch

Clearing flags on attachment: 291829

Committed r207460: <http://trac.webkit.org/changeset/207460>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.