1967878 - Make `PointerEventHandler` manage synthesizing `ePointerMove` per `pointerId`

Status: RESOLVED FIXED

(Core :: DOM: Events, defect)

Description

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)(sick)]

Bug 1966551 will start synthesizing ePointerMove when eMouseMove is synthesized. However, that means that only the last pointer which caused the synthesized eMouseMove will work as Pointer Events spec defined. For supporting multiple hoverable pointers, like pen or WebDriver's synthetic multiple mice, we should store the last position of all pointers and synthesize ePointerMove for all pointers.

Comment 1

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)(sick)]

Attachment: Bug 1967878 - part 1: Clean up `PointerInfo` r=edgar!,#dom-core

PointeInfo will have more members so that the bool should be
bit field to keep its size small as far as possible.

And also the constructor takes bool arguments too many. Therefore,
it's hard to read the callers. Therefore, this changes them to
enum classes.

Finally, some callers copies WidgetPointerEvent data to PointerInfo.
So, taking const WidgetPointerEvent& makes the caller side simpler.

Comment 2

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)(sick)]

Attachment: Bug 1967878 - part 2: Make `PresShell` synthesize `ePointerMove` separately from `eMouseMove` r=edgar!,smaug!,#dom-core

Pointer Events should be handled per pointerId. However, currently,
we dispatch pointer boundary events at a layout change only for the
last pointer which updated the last mouse location because we dispatch
a synthesized ePointerMove as a preceding pointer event of a synthetic
eMouseMove.

This patch makes PointerInfo store the last pointer state for each
pointerId. Then, the synthesizer can dispatch synthetic
ePointerMove for any active pointerIds. Finally, PresShell can
stop handling a preceding ePointerMove for a synthetic eMouseMove.

Comment 3

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)(sick)]

Attachment: Bug 1967878 - part 3: Make `PointerEventHandler` store the last mouse state instead of root `PresShell`s r=edgar!

It does not make sense that PresShell has members to store all data
which are required for dispatching a synthetic eMouseMove.

This patch makes PointerEventHandler store the most last mouse event
data except pointerId. Then, PresShell starts storing only the
last pointerId to skip more expensive checks whether the PresShell
actually needs to dispatch a synthesized eMouseMove.

This makes it easier to fix bug 749553.

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://github.com/mozilla-firefox/firefox/commit/f7dce0b1c623
https://hg.mozilla.org/integration/autoland/rev/0cef7d5b75f8
part 1: Clean up `PointerInfo` r=edgar,dom-core
https://github.com/mozilla-firefox/firefox/commit/eb51dedf6da7
https://hg.mozilla.org/integration/autoland/rev/7cca02201479
part 2: Make `PresShell` synthesize `ePointerMove` separately from `eMouseMove` r=edgar,smaug,dom-core
https://github.com/mozilla-firefox/firefox/commit/13f2c9c94a1a
https://hg.mozilla.org/integration/autoland/rev/dd7642eeb99d
part 3: Make `PointerEventHandler` store the last mouse state instead of root `PresShell`s r=edgar

Comment 5

[Cristian Tuns]

https://hg.mozilla.org/mozilla-central/rev/0cef7d5b75f8
https://hg.mozilla.org/mozilla-central/rev/7cca02201479
https://hg.mozilla.org/mozilla-central/rev/dd7642eeb99d

Comment 6

[Web Platform Test Sync Bot [:wpt-sync] (Matrix: #interop:mozilla.org)]

Created web-platform-tests PR https://github.com/web-platform-tests/wpt/pull/53215 for changes under testing/web-platform/tests

Comment 7

[Web Platform Test Sync Bot [:wpt-sync] (Matrix: #interop:mozilla.org)]

Upstream PR merged by moz-wptsync-bot