Delete WebCore::DisplayList::SetInlineFillGradient

• Component: WebCore Misc.
• Status: RESOLVED FIXED

[Wenson Hsieh]

This display list item was originally introduced to allow Canvas Lines to run to fully run without using any out-of-line display list items, when using Concurrent Display Lists.

However, the way this item is currently being encoded is not safe, since Gradient::Data is a Variant.

I did not measure use of this display list item to affect performance of Canvas Lines in the streaming display list prototype, so instead of making it work with the new architecture, we should just delete it altogether.

[Wenson Hsieh]

Created attachment 440426 [details]
For EWS

[Simon Fraser (smfr)]

Comment on attachment 440426 [details]
For EWS

View in context: https://bugs.webkit.org/attachment.cgi?id=440426&action=review

> Source/WebCore/ChangeLog:10
> +        -       It should never have been encoded and decoded as POD data in the first place, due to the fact that it

weird flex but ok

[Wenson Hsieh]

Comment on attachment 440426 [details]
For EWS

View in context: https://bugs.webkit.org/attachment.cgi?id=440426&action=review

Oops, looks like I missed a spot when splitting this patch out, where we still look for `GraphicsContextState::FillGradientChange` in `containsOnlyInlineStateChanges`. This should fix the failing tests in mac-wk2

>> Source/WebCore/ChangeLog:10
>> +        -       It should never have been encoded and decoded as POD data in the first place, due to the fact that it
> 
> weird flex but ok

😅

[Wenson Hsieh]

Created attachment 440445 [details]
For EWS

[EWS]

Committed r283697 (242623@main): <https://commits.webkit.org/242623@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 440445 [details].

[Radar WebKit Bug Importer]

<rdar://problem/83963203>