1446832 - Hover state still set on element during mouseleave event.

Status: RESOLVED FIXED

(Core :: DOM: Events, defect)

Description

[Gawan Fiore]

User Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_11_4) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/64.0.3282.186 Safari/537.36

Steps to reproduce:

I can reproduce the error consistently by following these steps:
1) Create an HTML element and give it an id (let's say it's id is 'elem1').
2) Create a Javascript "mouseleave" event listener on that HTML element and corresponding function to be executed. In that function check which element(s) in the DOM has the :hover property (using any of the many ways to select elements with :hover). Also check which element(s) the cursor is currently over using document.elementFromPoint(event.clientX, event.clientY).


Actual results:

When the event is triggered and the above corresponding function is executed, the :hover selector returns that 'elem1' (and only 'elem1') is currently the element hovered over. However, document.elementFromPoint() returns that the element currently under the cursor is NOT 'elem1' - it is whatever element is next to (or behind) 'elem1'.


Expected results:

When a 'mouseleave' event is triggered for an element (in this case 'elem1') the cursor is no longer over that element and thus the element should no longer be selectable with :hover. Instead, the element that the cursor is currently over should be selectable with :hover. This is the behavior demonstrated by Chrome, Safari, Opera, etc.

Comment 1

[Emilio Cobos Álvarez [:emilio]]

Could you attach a test-case, using the "Attach file" link in bugzilla? It'd be much easier to see and debug.

Thanks!

Comment 2

[Emilio Cobos Álvarez [:emilio]]

(I suspect we're just updating the hover state after dispatching the element or something like that...)

Comment 3

[Emilio Cobos Álvarez [:emilio]]

Attachment: test-case.

Well, I just got curious myself, though it's late in Europe... :P

Should never print "element", but it does.

Comment 4

[Gawan Fiore]

Attachment: test.html

This is the simple test case I used.

Comment 5

[Gawan Fiore]

The test case I used shows effectively the same behavior as yours it looks like. Thanks for recategorizing - wasn't sure if this should land in CSS Object or Events.

Comment 6

[Emilio Cobos Álvarez [:emilio]]

So, the relevant code is here:

  https://searchfox.org/mozilla-central/rev/3abf6fa7e2a6d9a7bfb88796141b0f012e68c2db/dom/events/EventStateManager.cpp#4351

Note how we NotifyMouseOut (which triggers the mouseleave event), _then_ set the hover state on the new node (unsetting it later), _then_ dispatch mouseenter on the new one.

It may just be a matter of reordering those...

Comment 7

[Emilio Cobos Álvarez [:emilio]]

Are you interested in fixing this by any chance?

Happy to mentor if so, though probably someone more familiar with the events code may be a more suitable reviewer :).

It may be a matter of writing that patch as suggested in comment 6 and write an automated test (which can be a bit harder, you need to synthesize events). You could probably use:

  https://searchfox.org/mozilla-central/rev/3abf6fa7e2a6d9a7bfb88796141b0f012e68c2db/layout/style/test/test_hover_quirk.html#71

as a guide.

Otherwise I can try to fix it myself I guess :)

Comment 8

[Emilio Cobos Álvarez [:emilio]]

If so, https://developer.mozilla.org/en-US/docs/Mozilla/Developer_guide and https://developer.mozilla.org/en-US/docs/Mozilla/Developer_guide/How_to_Submit_a_Patch are the relevant bits :)

Comment 9

[Gawan Fiore]

I would be interested in fixing it, but I'm extremely busy right now and won't be able to look at it for another few days. If you (or anyone else) want to fix it in the meantime please feel free to.

Comment 10

[Emilio Cobos Álvarez [:emilio]]

(In reply to Gawan Fiore from comment #9)
> I would be interested in fixing it, but I'm extremely busy right now and
> won't be able to look at it for another few days. If you (or anyone else)
> want to fix it in the meantime please feel free to.

Ok, will do, thanks!

If you want any other bug to fix or something once you have more time feel free to reach out to me either over email or IRC :)

Comment 11

[Phabricator Automation]

Attachment: Bug 1446832: Update hover state before firing mouseleave, to match other browsers. r=smaug

Comment 12

[Phabricator Automation]

Comment on attachment 8960163 [details]
Bug 1446832: Update hover state before firing mouseleave, to match other browsers. r=smaug

Olli Pettay [:smaug] has approved the revision.

https://phabricator.services.mozilla.com/D772

Comment 13

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/62fbfdc5ac11
Update hover state before firing mouseleave, to match other browsers. r=smaug

Comment 14

[Stefan Hindli [:stefan_hindli]]

https://hg.mozilla.org/mozilla-central/rev/62fbfdc5ac11