Allow passive mouse wheel event listeners to not force synchronous scrolling

[Simon Fraser (smfr)]

Since bug 149466 is fixed, we should avoid synchronous scrolling for pages with passive wheel event handlers.

[Radar WebKit Bug Importer]

<rdar://problem/28265360>

[Simon Fraser (smfr)]

Created attachment 411270 [details]
Patch

[Sam Weinig]

Comment on attachment 411270 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=411270&action=review

> Source/WebCore/ChangeLog:9
> +        Turn on ENABLE_WHEEL_EVENT_REGIONS for macOS, and hit-test the passive and active regions

This might totally not be feasible at this stage, but is this something we can do with a runtime switch rather than an #ifdef?

> Source/WebCore/page/Frame.cpp:319
>      // FIXME: This needs to look at wheel event handlers too.

Is this comment still needed?

[Simon Fraser (smfr)]

(In reply to Sam Weinig from comment #3)
> Comment on attachment 411270 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=411270&action=review
> 
> > Source/WebCore/ChangeLog:9
> > +        Turn on ENABLE_WHEEL_EVENT_REGIONS for macOS, and hit-test the passive and active regions
> 
> This might totally not be feasible at this stage, but is this something we
> can do with a runtime switch rather than an #ifdef?

I plan to remove the #ifdef. This could be behind a switch, but I'm not sure it needs to be?

> > Source/WebCore/page/Frame.cpp:319
> >      // FIXME: This needs to look at wheel event handlers too.
> 
> Is this comment still needed?

No, I resolved it.

[Sam Weinig]

(In reply to Simon Fraser (smfr) from comment #4)
> (In reply to Sam Weinig from comment #3)
> > Comment on attachment 411270 [details]
> > Patch
> > 
> > View in context:
> > https://bugs.webkit.org/attachment.cgi?id=411270&action=review
> > 
> > > Source/WebCore/ChangeLog:9
> > > +        Turn on ENABLE_WHEEL_EVENT_REGIONS for macOS, and hit-test the passive and active regions
> > 
> > This might totally not be feasible at this stage, but is this something we
> > can do with a runtime switch rather than an #ifdef?
> 
> I plan to remove the #ifdef. This could be behind a switch, but I'm not sure
> it needs to be?

No switch and no #ifdef is the best outcome, so I am happy.

> 
> > > Source/WebCore/page/Frame.cpp:319
> > >      // FIXME: This needs to look at wheel event handlers too.
> > 
> > Is this comment still needed?
> 
> No, I resolved it.

Remove it then!

[Simon Fraser (smfr)]

https://trac.webkit.org/changeset/268476/webkit

[Hector Lopez]

editing/editable-region/iframe.html

Test is a failing according to history on iOS. First occurrence of failure is at r268476.

History:

https://results.webkit.org/?suite=layout-tests&test=editing%2Feditable-region%2Fiframe.html

By the diff it looks as though a re-baseline for result to accommodate the difference in width of 10

Diff:
--- /Volumes/Data/slave/ipados-simulator-14-release-tests-wk2/build/layout-test-results/editing/editable-region/iframe-expected.txt
+++ /Volumes/Data/slave/ipados-simulator-14-release-tests-wk2/build/layout-test-results/editing/editable-region/iframe-actual.txt
@@ -33,7 +33,7 @@
                               (bounds 310.00 670.00)
                               (drawsContent 1)
                               (event region
-                                (rect (0,0) width=300 height=670)
+                                (rect (0,0) width=310 height=670)
                               (editable region
                                 (rect (8,286) width=284 height=24)
                                 (rect (9,431) width=285 height=24)

[Hector Lopez]

Re-baseline for test editing/editable-region/iframe.html

https://trac.webkit.org/changeset/268491/webkit

[Simon Fraser (smfr)]

*** Bug 217676 has been marked as a duplicate of this bug. ***