[Webkit-unassigned] [Bug 134663] [EFL] Fix crash caused by invalid cursor image.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=134663





--- Comment #11 from Gyuyoung Kim <gyuyoung.kim at webkit.org>  2014-07-06 19:32:45 PST ---
(From update of attachment 234461)
View in context: https://bugs.webkit.org/attachment.cgi?id=234461&action=review

>>>> Source/WebKit2/UIProcess/API/efl/EwkView.cpp:-225
>>>> -    self->updateCursor();
>>> 
>>> This callback was added by r165730 in order to restore previous cursor before going out of webview.
>>> http://trac.webkit.org/changeset/165730
>>> 
>>> Can we restore previous cursor when coming back to inside WebView without this cursor update ?
>> 
>> Re-entrance point can be always changed randomly according to user's intention.
>> In addition, cursor image is newly updated whenever mouse re-enter the webview again.
>> So, I can' sure whether restoration is good idea.
>> 
>> However, currently custom cursor seems to be destroyed as soon as it was rendered.
>> So, at least, updateCursor on mouse in callback seems to be removed.
> 
> I got your point, but it just mean that we move back to previous state.
> 
> IMO, One other solution is removing Evas Object based cursor.
> Recently, I realized that evas object cursor is not useful to collaborate with multiple windows.
> 
> After removed evas object cursor, I think that we can implement custom cursor as xpixmap which EwkView maintains.
> Although it tightly depends on X, ewebkit now only support X backend and we can make general function like applyFallbackCursor having several backend until EFL implement common APIs for the cursor.

How about keeping void EwkViewEventHandler<EVAS_CALLBACK_MOUSE_IN>::handleEvent(void* data, Evas*, Evas_Object*, void*) with "notImplemented", then add "FIXME: bla bla" ?

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.