1157476 - Remove synchronous dispatch in AndroidMediaResourceServer::Start

Status: RESOLVED FIXED

(Core :: Audio/Video, defect)

Description

[Bobby Holley (:bholley)]

This seems to run on the main thread anyway, and I think we can guarantee that we always initialize this thing on the main thread before using it off-main-thread. I'm hacking around it for now in bug 1144486, but we should fix it up properly.

Is there any reason not to do this at startup in MediaDecoder::InitStatics?

Comment 1

[Bobby Holley (:bholley)]

On IRC, cajbir and I decided we could do this in the MediaDecoder constructor.

Comment 2

[Bobby Holley (:bholley)]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=eede6af36b0c

Comment 3

[Bobby Holley (:bholley)]

(In reply to Bobby Holley (:bholley) from comment #2)
> https://treeherder.mozilla.org/#/jobs?repo=try&revision=eede6af36b0c

https://treeherder.mozilla.org/#/jobs?repo=try&revision=545486190bc7

Green.

Comment 4

[Bobby Holley (:bholley)]

Attachment: Remove synchronous dispatch in AndroidMediaResourceServer::Start. v1

Comment 5

[cajbir (:cajbir)]

Comment on attachment 8597583 [details] [diff] [review]
Remove synchronous dispatch in AndroidMediaResourceServer::Start. v1

Review of attachment 8597583 [details] [diff] [review]:
-----------------------------------------------------------------

r+ if you've tested that this works on physical hardware.

Comment 6

[Bobby Holley (:bholley)]

(In reply to cajbir (:cajbir) from comment #5)
> r+ if you've tested that this works on physical hardware.

I'm not in a position to do that easily. Can you suggest someone who's familiar with the feature who can do it?

This passes the robocop tests that exercise this codepath, FWIW.

Comment 7

[Bobby Holley (:bholley)]

And in particular, the failure mode here is going to be a very clear MOZ_CRASH (when the server is used without having been initialized), which is unlikely to go unnoticed.

Comment 8

[cajbir (:cajbir)]

(In reply to Bobby Holley (:bholley) from comment #6)
> 
> This passes the robocop tests that exercise this codepath, FWIW.


If the codepath is being tested via robocop that's fine.

Comment 9

[Bobby Holley (:bholley)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/3a1da9617e31

Comment 10

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/3a1da9617e31