1116202 - Intermittent run-by-dir leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...) from test_AudioNodeDevtoolsAPI.html leaking AudioNodeStreams

Status: RESOLVED WORKSFORME

(Core :: Web Audio, defect)

Description

[Vaibhav (:vaibhav1994)]

in bug 1110982, we are looking to enable tests where we run a fresh browser instance per directory.  Usually what happens is that a few tests fail because they accidentally depend on the state of the browser from an earlier test.

In the try run:
https://treeherder.mozilla.org/ui/#/jobs?repo=try&revision=c1a0dea1e39c

This test causes a leak in mac debug jobs. The failure goes like this:

 TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)

Comment 1

[Joel Maher ( :jmaher ) (UTC -8)]

Ehsan, I see you are the original author of this test file (from bug 1015783), I am not sure why it leaks and only on osx, maybe it is when we instantiate the AudioContext:
http://dxr.mozilla.org/mozilla-central/source/dom/media/webaudio/test/test_AudioNodeDevtoolsAPI.html#16

I don't have a osx box to verify this is easy to reproduce locally.

Comment 2

[(no longer active)]

It would have been helpful if this bug mentioned that the failure is intermittent.

Anyway, this assertion explains the leak:

14:13:33     INFO -  [1653] ###!!! ASSERTION: Appshell already destroyed?: 'Error', file /builds/slave/try-osx64-d-000000000000000000/build/src/dom/media/MediaStreamGraph.cpp, line 1743
14:13:33     INFO -  #01: mozilla::MediaStreamGraphImpl::AppendMessage(mozilla::ControlMessage*) [dom/media/MediaStreamGraph.cpp:1798]
14:13:33     INFO -  #02: mozilla::MediaStream::Destroy() [dom/media/MediaStreamGraph.cpp:1983]
14:13:33     INFO -  #03: mozilla::dom::AudioNode::DestroyMediaStream() [xpcom/base/nsRefPtr.h:43]
14:13:33     INFO -  #04: mozilla::dom::AudioDestinationNode::DestroyMediaStream() [dom/media/webaudio/AudioDestinationNode.cpp:418]
14:13:33     INFO -  #05: mozilla::dom::AudioNode::DisconnectFromGraph() [dom/media/webaudio/AudioNode.cpp:188]
14:13:33     INFO -  #06: mozilla::dom::AudioNode::cycleCollection::Unlink(void*) [xpcom/base/nsRefPtr.h:208]
14:13:33     INFO -  #07: mozilla::dom::AudioDestinationNode::cycleCollection::Unlink(void*) [xpcom/glue/nsCOMPtr.h:371]
14:13:33     INFO -  #08: nsCycleCollector::CollectWhite() [xpcom/base/nsCycleCollector.cpp:3316]
14:13:33     INFO -  #09: nsCycleCollector::Collect(ccType, js::SliceBudget&, nsICycleCollectorListener*, bool) [xpcom/base/nsCycleCollector.cpp:3646]
14:13:33     INFO -  #10: nsCycleCollector::Shutdown() [xpcom/base/nsCycleCollector.cpp:3575]
14:13:33     INFO -  #11: nsCycleCollector_shutdown() [xpcom/base/nsRefPtr.h:43]
14:13:33     INFO -  #12: mozilla::ShutdownXPCOM(nsIServiceManager*) [gfx/layers/ipc/AsyncTransactionTracker.h:130]
14:13:33     INFO -  #13: ScopedXPCOMStartup::~ScopedXPCOMStartup() [toolkit/xre/nsAppRunner.cpp:1335]
14:13:33     INFO -  #14: XREMain::XRE_main(int, char**, nsXREAppData const*) [memory/mozalloc/mozalloc.h:234]
14:13:33     INFO -  #15: XRE_main [toolkit/xre/nsAppRunner.cpp:4446]
14:13:33     INFO -  #16: main [browser/app/nsBrowserApp.cpp:292]

The issue is that we destroy the AudioDestinationNode during the last CC in ShutdownXPCOM, at which point the appshell service is already gone away, so we cannot finish destroying the AudioNodeStreams, at which point those objects and anything else that it keeps alive will leak.

I never managed to reproduce this issue, but I wrote an experimental patch which might fix this.  Please test it.

Comment 3

[(no longer active)]

Attachment: Destroy the destination node's media stream when the AudioContext is being shut down because the window is getting destroyed

Vaibhav, can you please test this patch and let me know if it fixes the issue?  Thanks!

Comment 4

[Joel Maher ( :jmaher ) (UTC -8)]

pushed to try:
https://treeherder.mozilla.org/ui/#/jobs?repo=try&revision=8a354a6448a8

Comment 5

[Joel Maher ( :jmaher ) (UTC -8)]

I forgot to enable the test, repushing:
https://treeherder.mozilla.org/ui/#/jobs?repo=try&revision=580c64590d76

Comment 6

[Joel Maher ( :jmaher ) (UTC -8)]

ok, this one has tests:
https://treeherder.mozilla.org/ui/#/jobs?repo=try&revision=aed4d244040e

Comment 7

[Joel Maher ( :jmaher ) (UTC -8)]

7 / 10 10.8 debug m-oth jobs have the leak with this patch:
https://treeherder.mozilla.org/#/jobs?repo=try&revision=aed4d244040e

Comment 8

[Joel Maher ( :jmaher ) (UTC -8)]

Comment on attachment 8542265 [details] [diff] [review]
Destroy the destination node's media stream when the AudioContext is being shut down because the window is getting destroyed

Review of attachment 8542265 [details] [diff] [review]:
-----------------------------------------------------------------

while this is probably a good patch to have in general, it doesn't seem to be fixing the leak.

Comment 9

[(no longer active)]

Attachment: Destroy the destination node's media stream when XPCOM is being shut down

So it turns out that nsGlobalWindow::FreeInnerObjects is not guaranteed to be called for chrome window objects, which I think is the reason why the previous patch did not work.  Can you please try this one instead?

Comment 10

[Joel Maher ( :jmaher ) (UTC -8)]

https://tbpl.mozilla.org/?tree=Try&rev=2b5200c37d9d

Comment 11

[Joel Maher ( :jmaher ) (UTC -8)]

Comment on attachment 8542590 [details] [diff] [review]
Destroy the destination node's media stream when XPCOM is being shut down

Review of attachment 8542590 [details] [diff] [review]:
-----------------------------------------------------------------

we still have a lot of leaks:
https://treeherder.mozilla.org/#/jobs?repo=try&revision=5e08d8164d81

Comment 12

[(no longer active)]

(In reply to Joel Maher (:jmaher) from comment #11)
> we still have a lot of leaks:
> https://treeherder.mozilla.org/#/jobs?repo=try&revision=5e08d8164d81

Yes, I never intended to fix all possible leaks.  ;-)  But the AudioNodeStream leaks are gone now.  Please file the rest separately, I think there is another issue with the MediaStreamGraph shutdown that is not related to AudioNodeStreams.  Thanks!

Comment 13

[Paul Adenot (:padenot)]

Comment on attachment 8542590 [details] [diff] [review]
Destroy the destination node's media stream when XPCOM is being shut down

Review of attachment 8542590 [details] [diff] [review]:
-----------------------------------------------------------------

jmaher, I've filed bug 1117727 (with a log example and some notes), blocking 1110982 to track the rest of the leak, that I believe is unrelated, as ehsan said.

Comment 14

[(no longer active)]

This patch bounced because of test failures <https://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&author=eakhgari%40mozilla.com>.  Turns out that AudioDestinationNode::DestroyMediaStream() might early return when called during shutdown!

Comment 15

[(no longer active)]

Attachment: Destroy the destination node's media stream when XPCOM is being shut down

This will prevent us from needing to rely on the last CC to destroy
the stream for us.  The last CC is too late for destroying streams
properly.

Comment 16

[Ryan VanderMeulen [:RyanVM]]

This caused failures on other platforms as well (like OSX 10.8 debug - https://treeherder.mozilla.org/ui/logviewer.html#?job_id=5085703&repo=mozilla-inbound)

Please give this a full Try run before pushing again.

Comment 17

[(no longer active)]

Ah, all of these issues are caused because AudioContext is now being kept alive by the observer service for the duration of the browser's lifetime.  I am testing a fix on the try server right now.

Comment 18

[(no longer active)]

Attachment: Destroy the destination node's media stream when XPCOM is being shut down

This will prevent us from needing to rely on the last CC to destroy
the stream for us.  The last CC is too late for destroying streams
properly.

https://treeherder.mozilla.org/ui/#/jobs?repo=try&revision=c42aa732530c

Comment 19

[(no longer active)]

Comment on attachment 8544583 [details] [diff] [review]
Destroy the destination node's media stream when XPCOM is being shut down

Wrong patch...

Comment 20

[(no longer active)]

Attachment: Destroy the destination node's media stream when XPCOM is being shut down

This will prevent us from needing to rely on the last CC to destroy
the stream for us.  The last CC is too late for destroying streams
properly.

Comment 21

[Paul Adenot (:padenot)]

Comment on attachment 8544772 [details] [diff] [review]
Destroy the destination node's media stream when XPCOM is being shut down

Review of attachment 8544772 [details] [diff] [review]:
-----------------------------------------------------------------

This looks good, but what's with the android m-10 perma orange on the try push?

::: dom/media/webaudio/AudioDestinationNode.cpp
@@ +403,3 @@
>  
> +    if (target) {
> +      // GetOwner() may return null if we are destroying the media streami

nit: s/streami/stream/

Comment 22

[(no longer active)]

Hmm, so I tried reverting the AudioDestinationNode hunk in <https://treeherder.mozilla.org/ui/#/jobs?repo=try&revision=b49d78696b90> but that didn't fix the m-10 perma orange.

Any chance you can take a look at why that happens, please?

Comment 23

[Paul Adenot (:padenot)]

We can't really destroy the destination here, because the audio thread is still running, and the AudioDestinationNode is used all over the place to get the .currentTime. What happens here is that we destroy the AudioDestinationNode's MediaStream, and then it crashes some time later on a virtual call on a nullptr.

I think it would be better to force-shutdown the MSG on XPCOM shutdown or something. The issue is that that's also async (and can take some time, depending on the platform), because the audio callback needs to fire, notice it needs to exit immediately, and return. Is there an even that fires earlier than XPCOM shutdown, but also indicates shutdown we could hook up to?

Comment 24

[(no longer active)]

(In reply to Paul Adenot (:padenot) from comment #23)
> We can't really destroy the destination here, because the audio thread is
> still running, and the AudioDestinationNode is used all over the place to
> get the .currentTime. What happens here is that we destroy the
> AudioDestinationNode's MediaStream, and then it crashes some time later on a
> virtual call on a nullptr.

Ah, right.  Can't we just make that AudioNodeStream* an nsRefPtr<AudioNodeStream>?

> I think it would be better to force-shutdown the MSG on XPCOM shutdown or
> something. The issue is that that's also async (and can take some time,
> depending on the platform), because the audio callback needs to fire, notice
> it needs to exit immediately, and return. Is there an even that fires
> earlier than XPCOM shutdown, but also indicates shutdown we could hook up to?

There are several events that fire before shutdown but an async tear down of the MSG thread will not work with any of them since they're all fired sequentially one after another.

There is this new nsIAsyncShutdownBlocker interface which is supposed to let you spin the event loop before shutdown proceeds, but I can't find anywhere in the code where it's used (at least from C++).  Perhaps Yoric can help us on how to use it?

Comment 25

[David Teller [:Yoric] - still alive but not very active]

Due to lack of reviewer, I haven't been able to land C++ code using nsIAsyncShutdownBlocker yet. However, I'm quite willing to help.

Essentially, the use case is a service FooService and clients that need to complete some work before FooService shuts down. In many cases, said work even needs to be started during shutdown. nsIAsyncShutdown* lets you describe this kind of dependencies, in most cases without spinning the event loop. Does this match your need?

Comment 26

[(no longer active)]

Hmm, not sure if we need any kind of dependency tracking here.  We just need to wait for the MSG thread's shutdown sequence to be finished before letting the XPCOM services die.

roc, should we just spin the event loop here?

Comment 27

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Spinning the event loop during shutdown sounds very dangerous.

Is it not possible to cancel the callback, synchronize with the audio thread and have it shut down while we wait on a monitor?

Comment 28

[(no longer active)]

(In reply to Robert O'Callahan (:roc) (Mozilla Corporation) from comment #27)
> Spinning the event loop during shutdown sounds very dangerous.

That is my intuition as well, but FWIW it seems like common practice in other code such as IDB.  It seems to have worked in practice...  (and that surprises me!)

Comment 30

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

Spinning event loop during xpcom-will-shutdown is no-no and documentation is clear about that, but
other than that it should be ok (-ish at least).

Comment 31

[Ben Turner (not reading bugmail, use the needinfo flag!)]

(In reply to :Ehsan Akhgari (not reading bugmail, needinfo? me!) from comment #28)
> That is my intuition as well, but FWIW it seems like common practice in
> other code such as IDB.

Also workers!

Comment 32

[Ben Turner (not reading bugmail, use the needinfo flag!)]

(In reply to ben turner [:bent] (use the needinfo? flag!) from comment #31)
> Also workers!

And PBackground shutdown

Comment 33

[(no longer active)]

How about we spin in profile-before-change too, like all the k00l k1dz do these days? ;-)

Comment 34

[(no longer active)]

(In reply to Robert O'Callahan (:roc) (offline Jan 17-22) (Mozilla Corporation) from comment #27)
> Spinning the event loop during shutdown sounds very dangerous.
> 
> Is it not possible to cancel the callback, synchronize with the audio thread
> and have it shut down while we wait on a monitor?

Reassigning to Paul.  If the answer to this question is yes, he's a better person to finish this anyway.

Comment 35

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=6416326
repository: mozilla-inbound
start_time: 2015-02-09T09:05:31
who: rvandermeulen[at]mozilla[dot]com
machine: talos-mtnlion-r5-045
buildname: Rev5 MacOSX Mountain Lion 10.8 mozilla-inbound debug test mochitest-other
revision: f82c3ca3b6c8

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)

Comment 36

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=6416036
repository: mozilla-inbound
start_time: 2015-02-09T10:00:47
who: rvandermeulen[at]mozilla[dot]com
machine: t-w732-ix-015
buildname: Windows 7 32-bit mozilla-inbound debug test mochitest-other
revision: 994141100597

TEST-UNEXPECTED-FAIL | leakcheck | default process: 1771 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)
Return code: 1
Return code: 1
Return code: 1

Comment 37

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=6424257
repository: mozilla-inbound
start_time: 2015-02-09T11:50:51
who: rvandermeulen[at]mozilla[dot]com
machine: talos-mtnlion-r5-045
buildname: Rev5 MacOSX Mountain Lion 10.8 mozilla-inbound debug test mochitest-other
revision: 41e390a2cf6e

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)
2931 INFO TEST-UNEXPECTED-FAIL | toolkit/content/tests/widgets/test_contextmenu_nested.xul | open inner popup event type DOMMenuItemInactive fired - got DOMMenuItemInactive, expected popupshowing
2932 INFO TEST-UNEXPECTED-FAIL | toolkit/content/tests/widgets/test_contextmenu_nested.xul | open inner popup event target ID outermenu - got outermenu, expected innermain
2933 INFO TEST-UNEXPECTED-FAIL | toolkit/content/tests/widgets/test_contextmenu_nested.xul | Test timed out. - expected PASS

Comment 38

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=6425910
repository: mozilla-inbound
start_time: 2015-02-09T12:23:09
who: rvandermeulen[at]mozilla[dot]com
machine: talos-mtnlion-r5-047
buildname: Rev5 MacOSX Mountain Lion 10.8 mozilla-inbound debug test mochitest-other
revision: 8d1bb05dffa7

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)

Comment 39

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=6435936
repository: mozilla-inbound
start_time: 2015-02-09T15:46:48
who: wkocher[at]mozilla[dot]com
machine: talos-mtnlion-r5-083
buildname: Rev5 MacOSX Mountain Lion 10.8 mozilla-inbound debug test mochitest-other
revision: 3f53cf53a34e

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)

Comment 40

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=6437160
repository: mozilla-inbound
start_time: 2015-02-09T17:34:54
who: rvandermeulen[at]mozilla[dot]com
machine: talos-mtnlion-r5-064
buildname: Rev5 MacOSX Mountain Lion 10.8 mozilla-inbound debug test mochitest-other
revision: 84877a7e0333

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)

Comment 41

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=6441369
repository: mozilla-inbound
start_time: 2015-02-09T17:53:58
who: philringnalda[at]gmail[dot]com
machine: talos-mtnlion-r5-065
buildname: Rev5 MacOSX Mountain Lion 10.8 mozilla-inbound debug test mochitest-other
revision: 8e41d641fabb

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)

Comment 42

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=fx-team&job_id=1945105
repository: fx-team
start_time: 2015-02-09T17:28:59
who: nigelbabu[at]gmail[dot]com
machine: talos-mtnlion-r5-081
buildname: Rev5 MacOSX Mountain Lion 10.8 fx-team debug test mochitest-other
revision: 7c9cbe8b347d

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)
2931 INFO TEST-UNEXPECTED-FAIL | toolkit/content/tests/widgets/test_contextmenu_nested.xul | open inner popup event type DOMMenuItemInactive fired - got DOMMenuItemInactive, expected popupshowing
2932 INFO TEST-UNEXPECTED-FAIL | toolkit/content/tests/widgets/test_contextmenu_nested.xul | open inner popup event target ID outermenu - got outermenu, expected innermain
2933 INFO TEST-UNEXPECTED-FAIL | toolkit/content/tests/widgets/test_contextmenu_nested.xul | Test timed out. - expected PASS

Comment 43

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=6446637
repository: mozilla-inbound
start_time: 2015-02-09T20:40:27
who: philringnalda[at]gmail[dot]com
machine: talos-mtnlion-r5-066
buildname: Rev5 MacOSX Mountain Lion 10.8 mozilla-inbound debug test mochitest-other
revision: cb655f416527

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)

Comment 44

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=6458836
repository: mozilla-inbound
start_time: 2015-02-10T02:08:33
who: tomcat[at]mozilla[dot]com
machine: talos-mtnlion-r5-059
buildname: Rev5 MacOSX Mountain Lion 10.8 mozilla-inbound debug test mochitest-other
revision: 9af435afa2b6

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)
2931 INFO TEST-UNEXPECTED-FAIL | toolkit/content/tests/widgets/test_contextmenu_nested.xul | open inner popup event type DOMMenuItemInactive fired - got DOMMenuItemInactive, expected popupshowing
2932 INFO TEST-UNEXPECTED-FAIL | toolkit/content/tests/widgets/test_contextmenu_nested.xul | open inner popup event target ID outermenu - got outermenu, expected innermain
2933 INFO TEST-UNEXPECTED-FAIL | toolkit/content/tests/widgets/test_contextmenu_nested.xul | Test timed out. - expected PASS

Comment 45

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=6452791
repository: mozilla-inbound
start_time: 2015-02-10T00:28:29
who: tomcat[at]mozilla[dot]com
machine: talos-mtnlion-r5-053
buildname: Rev5 MacOSX Mountain Lion 10.8 mozilla-inbound debug test mochitest-other
revision: df1e73a2888c

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)

Comment 46

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=fx-team&job_id=1947435
repository: fx-team
start_time: 2015-02-10T02:22:25
who: tomcat[at]mozilla[dot]com
machine: talos-mtnlion-r5-048
buildname: Rev5 MacOSX Mountain Lion 10.8 fx-team debug test mochitest-other
revision: c077e79e1771

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)

Comment 47

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=fx-team&job_id=1948385
repository: fx-team
start_time: 2015-02-10T02:51:19
who: tomcat[at]mozilla[dot]com
machine: panda-0041
buildname: Android 4.0 armv7 API 11+ fx-team opt test plain-reftest-5
revision: 49be6ef4d093

REFTEST TEST-UNEXPECTED-FAIL | http://10.26.137.18:30041/tests/layout/reftests/image-rect/background-zoom-2.html | image comparison (==), max difference: 16, number of differing pixels: 11

Comment 48

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=6460802
repository: mozilla-inbound
start_time: 2015-02-10T02:53:39
who: tomcat[at]mozilla[dot]com
machine: talos-mtnlion-r5-023
buildname: Rev5 MacOSX Mountain Lion 10.8 mozilla-inbound debug test mochitest-other
revision: 4dcfb660c4f2

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)
2931 INFO TEST-UNEXPECTED-FAIL | toolkit/content/tests/widgets/test_contextmenu_nested.xul | open inner popup event type DOMMenuItemInactive fired - got DOMMenuItemInactive, expected popupshowing
2932 INFO TEST-UNEXPECTED-FAIL | toolkit/content/tests/widgets/test_contextmenu_nested.xul | open inner popup event target ID outermenu - got outermenu, expected innermain
2933 INFO TEST-UNEXPECTED-FAIL | toolkit/content/tests/widgets/test_contextmenu_nested.xul | Test timed out. - expected PASS

Comment 49

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=6459862
repository: mozilla-inbound
start_time: 2015-02-10T02:22:59
who: tomcat[at]mozilla[dot]com
machine: t-w732-ix-008
buildname: Windows 7 32-bit mozilla-inbound debug test mochitest-other
revision: ab8180c889da

TEST-UNEXPECTED-FAIL | leakcheck | default process: 1771 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)
Return code: 1
Return code: 1
Return code: 1

Comment 50

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=fx-team&job_id=1948183
repository: fx-team
start_time: 2015-02-10T02:50:46
who: tomcat[at]mozilla[dot]com
machine: talos-mtnlion-r5-061
buildname: Rev5 MacOSX Mountain Lion 10.8 fx-team debug test mochitest-other
revision: ac7e4337f0af

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)

Comment 51

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=fx-team&job_id=1949257
repository: fx-team
start_time: 2015-02-10T03:22:37
who: tomcat[at]mozilla[dot]com
machine: talos-mtnlion-r5-070
buildname: Rev5 MacOSX Mountain Lion 10.8 fx-team debug test mochitest-other
revision: dbe74171ca49

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)

Comment 52

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=6460568
repository: mozilla-inbound
start_time: 2015-02-10T03:58:45
who: tomcat[at]mozilla[dot]com
machine: talos-mtnlion-r5-061
buildname: Rev5 MacOSX Mountain Lion 10.8 mozilla-inbound debug test mochitest-other
revision: 87b7034309e2

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)

Comment 53

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=fx-team&job_id=1950682
repository: fx-team
start_time: 2015-02-10T04:13:43
who: tomcat[at]mozilla[dot]com
machine: talos-mtnlion-r5-062
buildname: Rev5 MacOSX Mountain Lion 10.8 fx-team debug test mochitest-other
revision: e57f94dd5cf3

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)

Comment 54

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=6464250
repository: mozilla-inbound
start_time: 2015-02-10T05:14:08
who: tomcat[at]mozilla[dot]com
machine: talos-mtnlion-r5-097
buildname: Rev5 MacOSX Mountain Lion 10.8 mozilla-inbound debug test mochitest-other
revision: 162e8e32b182

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)

Comment 55

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=fx-team&job_id=1952850
repository: fx-team
start_time: 2015-02-10T06:33:31
who: tomcat[at]mozilla[dot]com
machine: talos-mtnlion-r5-024
buildname: Rev5 MacOSX Mountain Lion 10.8 fx-team debug test mochitest-other
revision: 03733cfb9e9c

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)

Comment 56

[Legacy TBPL/Treeherder Robot]

log: https://treeherder.mozilla.org/logviewer.html#?repo=fx-team&job_id=1954542
repository: fx-team
start_time: 2015-02-10T08:23:37
who: rvandermeulen[at]mozilla[dot]com
machine: talos-mtnlion-r5-070
buildname: Rev5 MacOSX Mountain Lion 10.8 fx-team debug test mochitest-other
revision: 38764a9bae60

TEST-UNEXPECTED-FAIL | leakcheck | default process: 2331 bytes leaked (AsyncLatencyLogger, AudioNodeEngine, AudioNodeStream, CondVar, ControlMessage, ...)

Comment 57

[Joel Maher ( :jmaher ) (UTC -8)]

:padenot, can you comment on this, otherwise I am inclined to disable this test for osx debug.

Comment 58

[Andrew McCreight [:mccr8]]

This seems to have mysteriously gone away.

Comment 59

[Paul Adenot (:padenot)]

yeah, not sure what's up, although since we now ::Close the context during FreeInnerObject, that certainly have changed the timing.

Comment 60

[Paul Adenot (:padenot)]

Closing, feel free to reopen if it starts again.