Bug 233252 – Updating grid gap value does not recalculate styles

• Component: Layout and Rendering
• Status: RESOLVED FIXED

[Brad Frost]

When updating the grid-gap (or gap, column-gap, row-gap, etc.) by editing the style attribute of the element, the styles are not recalculated and element does not change.
This also happens when editing the grid-gap value in Safari Inspector.

In other browsers, the styles are recalculated and the layout is updated.

Here is an example showing the behavior: https://jsbin.com/guhufuk/1/edit?css,js,output

[Radar WebKit Bug Importer]

<rdar://problem/85727334>

In case the example above doesn't load, this is another repro: https://codesandbox.io/s/eager-lederberg-vx8pg?file=/src/App.js

[Simon Fraser (smfr)]

cc'd the Igalia grid layout experts.

[zalan]

Created attachment 451120 [details]
Test reduction

[zalan]

This looks like a styleAdjuster bug. We get the setStyle call right after the gap value change but we compute the diff to be Equal -> no layout.
If I force the diff to Layout, the jsbin example starts working.

[zalan]

This affects flex gaps too.

[zalan]

@@ -664,6 +664,9 @@ static bool rareNonInheritedDataChangeRequiresLayout(const StyleRareNonInherited
     if (first.flexibleBox != second.flexibleBox)
         return true;
 
+    if (first.columnGap != second.columnGap || first.rowGap != second.rowGap)
+        return true;
+
     if (first.order != second.order
         || first.alignContent != second.alignContent
         || first.alignItems != second.alignItems

fixes both grid and flex. patch is coming up.

[zalan]

Created attachment 451126 [details]
Patch

[zalan]

will fix the "!=" older compiler vs. auto generating now issue

[zalan]

Created attachment 451129 [details]
[fast-cq]Patch

[EWS]

Committed r289241 (246925@main): <https://commits.webkit.org/246925@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 451129 [details].

[Brad Frost]

Thank you all for addressing this! Looking forward to the release.

[zalan]

(In reply to Brad Frost from comment #12)
> Thank you all for addressing this! Looking forward to the release.
Thank you for filing it!