1917867 - Test changes from bug 1917456 non-fatally assert with "event for content that isn't in a document"

Status: ASSIGNED

(Core :: DOM: Events, defect)

Description

[Emilio Cobos Álvarez (:emilio)]

That is, this NS_ASSERTION: https://searchfox.org/mozilla-central/rev/6e6265bd607cbe4c96e714f86d3d9e36620f63d6/layout/base/PresShell.cpp#8422

Masayuki, do you know if this is worrisome? I can change the test to not trigger it if it's just a test issue.

Comment 1

[Emilio Cobos Álvarez (:emilio)]

Attachment: Bug 1917867 - Simplify PresShell event dispatching asserts. r=smaug,masayuki

Comment 2

[Emilio Cobos Álvarez (:emilio)]

Attachment: Bug 1917867 - Make implicit pointer capture not capture anonymous content. r=masayuki,smaug

Comment 3

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)(Away until 6/30 except 24)]

In theory yes because PresShell::EventHandler basically handles user input. Therefore, the event target should be connected to the document in most cases. So, it detects a bug of dispatching event on correct target. From user's point of view, it looks like an input is ignored by web apps. So, I think that it's not serious, but may make users inconvenient.

On the other hand, in this case, the event must be gotpointercapture or lostpointercapture event. According to the spec, they must be fired on the capturing element. So, it may be disconnected. It seems that they should be dispatched without PresShell::EventHandler, but ePointeGotCapture event needs to be handled by EventStateManager to dispatch pointer boundary events. So, for avoiding to duplicate the path, it's reasonable to use PresShell::EventHandler to dispatch them.

So, I think that the assertion should not be tested if the event message is ePointerGotCapture or ePointerLostCapture.

Comment 4

[Emilio Cobos Álvarez (:emilio)]

So for PointerLostCapture we already do something different here: https://searchfox.org/mozilla-central/rev/4a8bd8efdfaa43dd14a16d3cb15bf86796fd1def/dom/events/PointerEventHandler.cpp#914

So probably only PointerGotCapture?

Comment 5

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/a5780f3e66d4
Simplify PresShell event dispatching asserts. r=smaug

Comment 6

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)(Away until 6/30 except 24)]

Oh, yeah, so, only ePointerGotCapture may be dispatched to uncomposed element even dispatched via PresShell:EventHandler. So, it should be added to the condition of NS_ASSERTION. (However, I'm not sure whether the event should be fired because the capture should've been being released at least.)

Comment 7

[pstanciu]

https://hg.mozilla.org/mozilla-central/rev/a5780f3e66d4

Comment 8

[BugBot [:suhaib / :marco/ :calixte]]

The leave-open keyword is there and there is no activity for 6 months.
:emilio, maybe it's time to close this bug?
For more information, please visit BugBot documentation.