Summary: | Bringing up the context menu on a link might also follow the link | ||||||
---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Brady Eidson <beidson> | ||||
Component: | WebKit2 | Assignee: | Brady Eidson <beidson> | ||||
Status: | RESOLVED FIXED | ||||||
Severity: | Normal | CC: | hugo.lima | ||||
Priority: | P2 | Keywords: | InRadar | ||||
Version: | 528+ (Nightly build) | ||||||
Hardware: | All | ||||||
OS: | All | ||||||
Attachments: |
|
Description
Brady Eidson
2011-04-12 17:59:55 PDT
Created attachment 89325 [details]
Patch v1
Comment on attachment 89325 [details]
Patch v1
r=me
Is there a way to test this? EventSender.mouseDown exists...
The problem with layout testing it is the asynchronous nature of the race between events and the context menu. You need to queue up an extra mouse event in the WebProcess (the mouse up) before the UIProcess actually shows the context menu. I can't think of a way we could do this reliably - at least certainly not with the existing eventSender API in LayoutTestController. I'm open to suggestions, but think it's prudent to land as-is in the meantime. This patch causes any layout test using mouse right click to fail on wk2, e.g.: fast/events/mouse-click-events.html Isn't the port responsible for send or not mouse events to webkit2 when the context menu is showing? Sorry if I misunderstood something, I'm new to webkit and started to hack on qt-wk2 click events layout tests. if this bug happen only when using async events, i.e. not using webkittestrunner, what about removing the "if (m_isShowingCon...)" only on WebPage::mouseEventSyncForTesting ? hmmm, I think I got the problem, forget my last comment. |