796827 - [b2g-bluetooth] Destroy BluetoothHfpManager instance while in shutdown

Status: RESOLVED FIXED

(Core :: DOM: Device Interfaces, defect)

Comment 1

[Gina Yeh [:gyeh] [:ginayeh]]

Follow up of Bug 791197

Comment 2

[Gina Yeh [:gyeh] [:ginayeh]]

Attachment: v1 patch

Comment 3

[Kyle Machulis [:qdot] [:kmachulis] (INACTIVE)]

Comment on attachment 666874 [details] [diff] [review]
v1 patch

Review of attachment 666874 [details] [diff] [review]:
-----------------------------------------------------------------

The fact that the diff isn't lining up worries me. Looks good otherwise, but would like to make sure the diff is what is expected before r+'ing. Re-r? after either fixing or telling me I'm crazy. :)

::: dom/bluetooth/BluetoothHfpManager.cpp
@@ +36,5 @@
> +
> +namespace {
> +
> +StaticRefPtr<BluetoothHfpManager> gBluetoothHfpManager;
> +

Nit: Don't need extra newlines between this and other statics.

@@ +106,5 @@
>    : mCurrentVgs(-1)
>    , mCurrentCallIndex(0)
>    , mCurrentCallState(nsIRadioInterfaceLayer::CALL_STATE_DISCONNECTED)
>  {
> +bool

So I think your diff may be off here. Splinter is showing Init() to be /in/ the constructor?

@@ +342,5 @@
> +{
> +  MOZ_ASSERT(NS_IsMainThread());
> +
> +  gInShutdown = true;
> +

Nit: We don't need all of these extra newlines.

Comment 4

[Gina Yeh [:gyeh] [:ginayeh]]

Attachment: v2 patch

Oops, you got me! I found it after attaching the patch :p

Comment 5

[Gina Yeh [:gyeh] [:ginayeh]]

Attachment: v2 patch

register NS_XPCOM_SHUTDOWN_OBSERVER_ID in Init()

Comment 6

[Gina Yeh [:gyeh] [:ginayeh]]

Attachment: v2 patch

remove observers of mozsettings-changed in Cleanup()

Comment 7

[Gina Yeh [:gyeh] [:ginayeh]]

Attachment: Final version, r=qdot

try server: https://tbpl.mozilla.org/?tree=Try&rev=7fc431b62908

Comment 8

[Eric Chou [:ericchou] [:echou]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/fd238b14ba35

Comment 9

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/fd238b14ba35