1216087 - Pointer is not unlocked when ancestor of the pointer lock element is removed

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[Xidorn Quan [:xidorn] UTC+10]

If we have tree like:
> div#outer
> |
> +- div#inner
and request pointer lock on #inner. If we remove the #outer element from the document, the pointer is not unlocked. However, if we only remove #inner, the pointer is unlocked.

Comment 1

[Xidorn Quan [:xidorn] UTC+10]

Attachment: MozReview Request: Bug 1216087 - Unlock pointer whenever the element is unbound from the tree.

Bug 1216087 - Unlock pointer whenever the element is unbound from the tree.

Comment 2

[Xidorn Quan [:xidorn] UTC+10]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=868fc71ea3ba

Comment 3

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

Comment on attachment 8675614 [details]
MozReview Request: Bug 1216087 - Unlock pointer whenever the element is unbound from the tree.

bah, silly mistake

Comment 4

[Xidorn Quan [:xidorn] UTC+10]

https://hg.mozilla.org/integration/mozilla-inbound/rev/272319a15188085adf710590289cf84b8192ca1a
Bug 1216087 - Unlock pointer whenever the element is unbound from the tree. r=smaug

Comment 5

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/272319a15188

Comment 6

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/272319a15188

Comment 7

[Xidorn Quan [:xidorn] UTC+10]

Comment on attachment 8675614 [details]
MozReview Request: Bug 1216087 - Unlock pointer whenever the element is unbound from the tree.

Approval Request Comment
[Feature/regressing bug #]: pointer lock
[User impact if declined]: pointer may not be unlocked as expected in some case
[Describe test coverage new/current, TreeHerder]: test included
[Risks and why]: low risk, the change is simple and straightforward
[String/UUID change made/needed]: n/a

Comment 8

[Liz Henry (:lizzard)]

Does this need to be in 43 for any particular reason?   Is it a regression (and if so, from when)
Can it ride with 44? I'd like for there to be more time for potential regressions to be uncovered if possible.

Comment 9

[Xidorn Quan [:xidorn] UTC+10]

This is not a regression. I believe this exists since we started supporting pointer lock. I consider this issue to be potentially harmful because we generate warnings for every mouse event when that happens.

Comment 10

[Liz Henry (:lizzard)]

Comment on attachment 8675614 [details]
MozReview Request: Bug 1216087 - Unlock pointer whenever the element is unbound from the tree.

Approved for uplift to aurora, includes tests, ok on m-c.

Comment 11

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/2a94d540599e