On Apr 16, 2013, at 5:08 AM, Allan Sandfeld Jensen <k...@carewolf.com> wrote:
> Hi webkit-dev > > I have recently uploaded a new patch to > https://bugs.webkit.org/show_bug.cgi?id=18930 to implement mouseenter and > mouseleave events. > > These events are part of DOM3 > http://www.w3.org/TR/DOM-Level-3-Events/#events-mouseevents and supported by > MSIE, Mozilla and pre-blink Opera. They are also > listed as issues jQuery needs to work around in WebKit > https://bugs.webkit.org/show_bug.cgi?id=110007 > > Previously they have not been implemented because they were not considered > important, and could cause performance regression by issuing more events on > every mouse-move that may not even be used by many websites. > > After I refactored hit-testing in the fall and especially moved the update of > hover/active state out of hit-testing, it is now possible to dispatch these > events where the hover state is updated, which makes the logic of dispatching > mouseenter and mouseleave events correctly simple. It is also possible to > only > do one check for possible capturing listeners before issuing any mouseenter > or > mouseleave events for a mousemove. > > This is what the new patch does. It does incur an additional overhead by > checking for the existence of capturing listeners, but the check is of lower > magnitude than existing iterations over the tree, and should keep performance > comparable, and is in all circumstances faster than issuing new events that > fall for deaf ears. > > Are there more objections for the support of mouseenter and mouseleave events > in WebKit? Seems like a reasonable feature to add. - Maciej
_______________________________________________ webkit-dev mailing list webkit-dev@lists.webkit.org https://lists.webkit.org/mailman/listinfo/webkit-dev
