When the column size of a grid area changes, we need to set prefer width dirty for the child that has width constraints to the grid area.
Created attachment 442148 [details] Patch
Comment on attachment 442148 [details] Patch Nice!
Comment on attachment 442148 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=442148&action=review > Source/WebCore/ChangeLog:9 > + reset the OverrideContainingBlockLogicalWidth value when we are executing a definite strategy. This is only performed when the Direction() is the InlineAxis (kForColumns if it's not an orthogonal item). However, the updateOverridingContainingBlockContentSizeForChild() call is executed when we are processing the BlockAxis direction (kForRows under the alerady mentioned assumption of a parallel item). > Source/WebCore/ChangeLog:12 > + I'm not sure this is correct. The grid item's preferred width shouldn't depend on its containing block (the grid area in this case) size, should it ?
Created attachment 442357 [details] Patch
(In reply to Javier Fernandez from comment #3) > Comment on attachment 442148 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=442148&action=review > > > Source/WebCore/ChangeLog:9 > > + reset the OverrideContainingBlockLogicalWidth value when we are executing a definite strategy. > > This is only performed when the Direction() is the InlineAxis (kForColumns > if it's not an orthogonal item). However, the > updateOverridingContainingBlockContentSizeForChild() call is executed when > we are processing the BlockAxis direction (kForRows under the alerady > mentioned assumption of a parallel item). > > > Source/WebCore/ChangeLog:12 > > + > > I'm not sure this is correct. The grid item's preferred width shouldn't > depend on its containing block (the grid area in this case) size, should it ? Thanks very much for the comments and the offline discussions. The patch has now been updated.
Committed r284793 (243502@main): <https://commits.webkit.org/243502@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 442357 [details].
<rdar://problem/84621191>