Bug 210778 – EventHandler::selectCursor() has broken resize over coordinate conversion code

• Component: UI Events
• Status: RESOLVED FIXED

[Simon Fraser (smfr)]

This looks really wrong, and is broken:

                if (FrameView* view = m_frame.view())
                    inResizer = layer->isPointInResizeControl(view->windowToContents(roundedIntPoint(result.localPoint())));

Why windowToContents() on the localPoint from a HitTestResult?

[Simon Fraser (smfr)]

This is apparent in this very textfield, where you get the iBeam cursor over the resizer.

[Simon Fraser (smfr)]

Weirdly result.localPoint() is not local when set by RenderBlock::updateHitTestResult()

[Radar WebKit Bug Importer]

<rdar://problem/62261183>

[Simon Fraser (smfr)]

Created attachment 397390 [details]
Patch

[Simon Fraser (smfr)]

*** Bug 156234 has been marked as a duplicate of this bug. ***

[alan]

Comment on attachment 397390 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=397390&action=review

> Source/WebCore/page/EventHandler.cpp:1546
> +            // FIXME: With right right-aligned text in a box, the renderer here is usually a RenderText, which prevents showing the resize cursor: webkit.org/b/210935.

too many rights.

> Source/WebCore/page/EventHandler.cpp:1547
> +            if (auto layer = downcast<RenderLayerModelObject>(*renderer).layer()) {

not sure what the WebKit style at this point, but I'd prefer auto* layer =

[alan]

> not sure what the WebKit style at this point, but I'd prefer auto* layer =
what the WebKit style is even

[Simon Fraser (smfr)]

Created attachment 397392 [details]
Patch

[EWS]

Committed r260615: <https://trac.webkit.org/changeset/260615>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 397392 [details].