1443482 - "contextmenu" event emitted before "mouseup"

Status: NEW

(Remote Protocol :: Marionette, defect, P3)

Description

[Andreas Tolfsen ❲:ato❳]

As report in https://github.com/mozilla/geckodriver/issues/1201 the
DOM events emitted for Control/Meta + Right + Click sequence are
in the wrong order.  Marionette incorrectly emits the contextmenu
event before the mouseup event.

Expected order: mousemove, mousedown, mouseup, contextmenu
Actual order:   mousemove, mousedown, contextmenu, mouseup

Comment 1

[Andreas Tolfsen ❲:ato❳]

contextmenu is fired in [1] and mouseup is fired either on the next
pointerUp action or implicitly when the sequence ends.  For this
to work as expected we need to maintain some state in [1] and fire
it after the next pointerUp has been processed.

Comment 2

[Henrik Skupin [:whimboo][⌚️UTC+2]]

(In reply to Andreas Tolfsen 「:ato」 from comment #0)
> Expected order: mousemove, mousedown, mouseup, contextmenu
> Actual order:   mousemove, mousedown, contextmenu, mouseup

I don't know what's the correct order of firing the events, and why Chrome/Edge is handling that differently than us. Also MDN doesn't mention it and I cannot find something in the HTML spec.

Olli, can you please help us? Thanks.

Comment 3

[phlonsing]

(In reply to Henrik Skupin (:whimboo) from comment #2)
> (In reply to Andreas Tolfsen 「:ato」 from comment #0)
> > Expected order: mousemove, mousedown, mouseup, contextmenu
> > Actual order:   mousemove, mousedown, contextmenu, mouseup
> 
> I don't know what's the correct order of firing the events, and why
> Chrome/Edge is handling that differently than us. Also MDN doesn't mention
> it and I cannot find something in the HTML spec.
> 
> Olli, can you please help us? Thanks.

I just wanted to add that there is not only a difference between Chrome/Edge but also between Firefox with Selenium and Firefox without Selenium.
In my opinion the difference betweed Firefox with/without Selenium is more important because Selenium/WebDriver should simulate real user behavior as accurate as possible. The order of the events may make a difference in the behavior of web frameworks.

Regards,
Philip

Comment 4

[Olli Pettay [:smaug][[e-mail]]]

This is OS dependent.

Comment 5

[Henrik Skupin [:whimboo][⌚️UTC+2]]

(In reply to Olli Pettay [:smaug] from comment #4)

This is OS dependent.

Do we have some code or comments in-tree that you could refer to?

Comment 6

[Olli Pettay [:smaug][[e-mail]]]

https://github.com/w3c/pointerevents/issues/191#issuecomment-346300383

https://searchfox.org/mozilla-central/rev/f8a41209af503016e78278774052d48d8c52b91c/widget/windows/nsWindow.cpp#5700,5723 is where we get the event from Windows.

https://searchfox.org/mozilla-central/rev/f8a41209af503016e78278774052d48d8c52b91c/modules/libpref/init/StaticPrefList.yaml#9950-9956

Comment 7

[Henrik Skupin [:whimboo][⌚️UTC+2]]

We haven't had the time to work on this bug yet. Lets discuss about it's priority. It doesn't look like a P2.

Comment 8

[Henrik Skupin [:whimboo][⌚️UTC+2]]

Maybe my work on bug 1773393 could fix that. Otherwise this is a P3 as discussed today in the triage meeting.