1206904 - dom/ipc/tests/test_permission_for_nested_oop_app.html fails on Linux m-e10s with APZ enabled

Status: RESOLVED FIXED

(Core :: Panning and Zooming, defect)

Description

[Kartikaya Gupta (email:[e-mail])]

Attachment: Partial backtrace

With APZ enabled on desktop, the dom/ipc/tests/test_permission_for_nested_oop_app.html test fails on Linux m-e10s. According to the log it looks like the child process is crashing with the attached stack, which probably means we need to fix bug 1020199 in order to get this working.

Comment 1

[Kartikaya Gupta (email:[e-mail])]

This is the try push I was looking at that has the failure: https://treeherder.mozilla.org/#/jobs?repo=try&revision=41a2e496ef4b&group_state=expanded

Comment 2

[Kartikaya Gupta (email:[e-mail])]

Attachment: Patch

Simple workaround for the crash; this makes the test pass for me locally. This way we don't have to wait for the "proper" fix which peterv is working on in bug 1020199.

Comment 3

[Botond Ballo [:botond]]

Comment on attachment 8665047 [details] [diff] [review]
Patch

Review of attachment 8665047 [details] [diff] [review]:
-----------------------------------------------------------------

::: layout/ipc/RenderFrameParent.cpp
@@ +338,5 @@
>    // We can't get a ref to the APZCTreeManager until after the child is
>    // created and the static getter knows which CompositorParent is
>    // instantiated with this layers ID. That's why try to fetch it when
>    // we first need it and cache the result.
> +  if (!mApzcTreeManager && mAsyncPanZoomEnabled && XRE_IsParentProcess()) {

Please add a comment saying that the XRE_IsParentProcess() check is needed because in a nested-oop setup, RenderFrameParent can exist in a child process too.

Comment 4

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/c07f7e2d10fb

Comment 5

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/c07f7e2d10fb