ChromeClient::needsImmediateRenderingUpdate() only exists to work around a WebKit1 bug

[Simon Fraser (smfr)]

ChromeClient::needsImmediateRenderingUpdate() only existing to work around a WebKit1 bug

[Simon Fraser (smfr)]

Created attachment 410563 [details]
Patch

[Said Abou-Hallawa]

Comment on attachment 410563 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=410563&action=review

> Source/WebCore/ChangeLog:9
> +        because the RunLoopObserver was always invalidate at the end of the callback.

invalidating? or calling invalidate()?

> Source/WebKitLegacy/mac/WebView/WebViewData.h:125
> +    bool m_insideCallback { false };
> +    bool m_rescheduledInsideCallback { false };

I think using two booleans can be a little bit difficult to understand. Can't we use enum for the states of these two members?

    enum class SchedulingState { None, InsideCallback, AlreadyRescheduled } // Maybe the names need to be changed.

> Source/WebKitLegacy/mac/WebView/WebViewData.mm:129
> +        SetForScope<bool> insideCallbackScope(m_insideCallback, true);
> +        m_rescheduledInsideCallback = false;

This can be changed to 
    SetForScope<SchedulingState> change(m_schedulingState, SchedulingState::InsideCallback);

[Simon Fraser (smfr)]

https://trac.webkit.org/changeset/268022/webkit

[Radar WebKit Bug Importer]

<rdar://problem/69979228>