848006 - Alarm does not sound if phone is plugged in

Status: VERIFIED FIXED

(Firefox OS Graveyard :: Gaia::Clock, defect)

Description

[David Flanagan [:djf]]

Steps to reproduce:

1) Plug phone into computer via USB cable. (I have remote debugging on, but don't have the sdcard mounted, I don't know if that is relevant).

2) Set an alarm for a couple of minutes in the future, and set it to recur on all weekdays

3) Wait for alarm

4) At the specified time, the screen displays the alarm, but there is no sound or vibration

5) Hit snooze and wait 5 minutes

6) Same as step 4: alarm screen shows but not sound is played.

7) Hit snooze, unplug phone and wait 5 minutes

8) Alarm plays.

Comment 1

[David Flanagan [:djf]]

This might be related to bug 845347.  Cc'ing some people from that bug and nominating.

Comment 2

[Naoki Hirata :nhirata (please use needinfo instead of cc)]

djf, I'm not sure.  I was able to replicate bug 845347 without having the device plugged in.  I'll see if I can find the regression window for this at the same time.

Comment 3

[David Flanagan [:djf]]

I've tried the STR above twice and experienced the bug both times.  With some further testing, I think I've found that the phone has to go to sleep (or you have to push the sleep button) after setting the alarm. If you just set it for 1 minute in the future and the phone does not sleep, then you get the sound.

This occurs for me even without setting any recurrances for the alarm in step 2.

I've now turned off remote debugging and the console in the settings app, and I still see the bug when plugged in, so it isn't just affecting developers. If reproducable, this would affect anyone charging their phone overnight.

After unplugging the phone I got one silent alarm and one noisy alarm.  I've got no idea what is going on.  But having the phone plugged in does seem to be a factor.

Comment 4

[David Flanagan [:djf]]

I'm continuing to see this with further testing. If the phone is plugged in and asleep, then I don't get sound. If it is unplugged or if it is awake the I do get sound.

Comment 5

[Naoki Hirata :nhirata (please use needinfo instead of cc)]

Does it matter if you sleep it with an app being in the foreground?

I'm currently trying to play with the notion that maybe it might also have to do with an app in the foreground or the homescreen being in the foreground as well...

Comment 6

[David Flanagan [:djf]]

Naoki,

I think when I was testing this, I would normally set the alarm, then hit the home button to return to the homescreen and then sleep the phone.  I'm not sure if I always did that though. Can you reproduce it?

I started seeing this on a nightly build from late last week.  Today I made a custom build of b2g18 and gaia master and am still seeing it.

Comment 7

[(inactive after 6/18) Alive Kuo [:alive]]

I could repro this one!
Investigating.

Comment 8

[[e-mail], [e-mail]]

According Description step 4) "At the specified time, the screen displays the alarm, but there is no sound or vibration", I guess that the mozHidden value is not reliable as before in the same checking time. Keep to investigate.

Comment 9

[(inactive after 6/18) Alive Kuo [:alive]]

E/GeckoConsole( 1761): Content JS LOG at app://clock.gaiamobile.org/js/onring.js:43 in rv_init: [ Wed Mar  6 15:52:00 2013 ][clock] First checkin is HIDDEN.
E/GeckoConsole( 1761): Content JS LOG at app://clock.gaiamobile.org/js/onring.js:57 in rv_checkMozHidden: [ Wed Mar  6 15:52:01 2013 ][clock] Second checkin is HIDDEN.
E/GeckoConsole( 1761): Content JS LOG at app://clock.gaiamobile.org/js/visibility.js:3 in visibilityChange: [ Wed Mar  6 15:52:01 2013 ] [clock] I am shown.
E/GeckoConsole( 1761): Content JS LOG at app://clock.gaiamobile.org/js/visibility.js:5 in visibilityChange: [ Wed Mar  6 15:52:10 2013 ] [clock] I am hidden.

Seems alarm is up before visibilitychange event..

Comment 10

[(inactive after 6/18) Alive Kuo [:alive]]

I add a 300ms*10 count timer this time:

:42 in anonymous: Upgrading db
:47 in anonymous: Upgrading db done
43 in rv_init: [ Wed Mar  6 17:21:01 2013 ][clock] First checkin is HIDDEN.
57 in rv_checkMozHidden: [ Wed Mar  6 17:21:01 2013 ][clock] Second checkin is HIDDEN.
222 in timer: [ Wed Mar  6 17:21:01 2013 ][clock]  1  cont checkin is . HIDDEN
Content JS LOG at app://clock.gaiamobile.org/js/visibility.js:3 in visibilityChange: [ Wed Mar  6 17:21:01 2013 ] [clock] I am shown.
222 in timer: [ Wed Mar  6 17:21:01 2013 ][clock]  2  cont checkin is . SHOWN
222 in timer: [ Wed Mar  6 17:21:01 2013 ][clock]  3  cont checkin is . SHOWN
222 in timer: [ Wed Mar  6 17:21:01 2013 ][clock]  4  cont checkin is . SHOWN
222 in timer: [ Wed Mar  6 17:21:02 2013 ][clock]  5  cont checkin is . SHOWN
222 in timer: [ Wed Mar  6 17:21:02 2013 ][clock]  6  cont checkin is . SHOWN
222 in timer: [ Wed Mar  6 17:21:02 2013 ][clock]  7  cont checkin is . SHOWN
222 in timer: [ Wed Mar  6 17:21:03 2013 ][clock]  8  cont checkin is . SHOWN
222 in timer: [ Wed Mar  6 17:21:03 2013 ][clock]  9  cont checkin is . SHOWN
222 in timer: [ Wed Mar  6 17:21:03 2013 ][clock]  10  cont checkin is . SHOWN

Comment 11

[(inactive after 6/18) Alive Kuo [:alive]]

Tested with https://github.com/alivedise/gaia/tree/muted-alarm_v1-train


app://system.gaiamobile.org/js/screen_manager.js:368 in scm_turnScreenOn: [ Wed Mar  6 18:58:00 2013 ][non clock] screen is actually on.  =====> power.screenEnabled = true
app://system.gaiamobile.org/js/attention_screen.js:132 in anonymous: Wed Mar  6 18:58:00 2013 [non clock] set true!  =====> onring iframe setVisible to true
app://system.gaiamobile.org/js/visibility.js:3 in visibilityChange: [ Wed Mar  6 18:58:01 2013 ] [non clock:system] I am shown.   ======> system app(content) get visibilitychange event
app://clock.gaiamobile.org/js/onring.js:43 in rv_init: [ Wed Mar  6 18:58:01 2013 ][clock] First checkin is HIDDEN.   ======> onring.js check document.mozHidden
app://clock.gaiamobile.org/js/onring.js:57 in rv_checkMozHidden: [ Wed Mar  6 18:58:01 2013 ][clock] Second checkin is HIDDEN.   ======> onring.js check document.mozHidden w/ setTimeout(0)
app://clock.gaiamobile.org/js/onring.js:223 in timer: [ Wed Mar  6 18:58:01 2013 ][clock]  1  cont checkin is . HIDDEN   ======> onring.js check document.mozHidden w/ setTimeout(300ms)
app://clock.gaiamobile.org/js/visibility.js:3 in visibilityChange: [ Wed Mar  6 18:58:01 2013 ] [clock] I am shown.  ======> onring.js get visibilitychange event

Comment 12

[(inactive after 6/18) Alive Kuo [:alive]]

So there's latency between setVisible to onring.html and when it's up to check mozHidden, the state is still invisible.

Comment 13

[Justin Lebar (not reading bugmail)]

(In reply to Alive Kuo [:alive] (Life is a struggle.) from comment #12)
> So there's latency between setVisible to onring.html and when it's up to
> check mozHidden, the state is still invisible.

So can you just listen to the mozvisibilitychange event, and if the visibility changes away from hidden, start the alarm?

We need to fix this bug.

Comment 14

[(inactive after 6/18) Alive Kuo [:alive]]

(In reply to Justin Lebar [:jlebar] from comment #13)
> (In reply to Alive Kuo [:alive] (Life is a struggle.) from comment #12)
> > So there's latency between setVisible to onring.html and when it's up to
> > check mozHidden, the state is still invisible.
> 
> So can you just listen to the mozvisibilitychange event, and if the
> visibility changes away from hidden, start the alarm?
> 
> We need to fix this bug.

Yes.

The only thing I wonder this breaks the fix that alarm needs to be dismissed when it goes off during a call. But that's not a big feature, more like an edge case.

Comment 15

[(inactive after 6/18) Alive Kuo [:alive]]

Attachment: https://github.com/mozilla-b2g/gaia/pull/8514

Patch v1:
Add the last guard to ensure alarm really sounds/vibrates if the timer to check visibility doesn't come in time.

Comment 16

[Tim Guan-tin Chien [:timdream] (please needinfo)]

Comment on attachment 722139 [details]
https://github.com/mozilla-b2g/gaia/pull/8514

It's very sad that those OOP timing/racing issue in Gecko force us to do all these checks, but I guess it's necessary.

Comment 17

[(inactive after 6/18) Alive Kuo [:alive]]

https://github.com/mozilla-b2g/gaia/commit/0bee8fb8d6f7c321e5cc3d79b26e539ac9c4f04e master

Comment 19

[Naoki Hirata :nhirata (please use needinfo instead of cc)]

It's a bit late, but I could not reproduce the bug on 2/27 build and I could repro it on the 2/28 build...  Odd.  I guess it very much has to do with the timing/racing.

Comment 20

[John Ford [:jhford] CET/CEST Berlin Time]

Uplifted commit 0bee8fb8d6f7c321e5cc3d79b26e539ac9c4f04e as:
v1-train: 87cc896b7a2016df090169dd1ba21bba1e879722
v1.0.1: 24718e0f3fac9ea98b35a772531431adc7553f67

Comment 21

[Angela Hubenya]

The alarm sounds when the phone is plugged in to usb now. Issue is verified as fixed. 

Unagi, Build ID: 20130322070203
Gecko: http://hg.mozilla.org/releases/mozilla-b2g18_v1_0_1/rev/4931ec89ebbe
Gaia: 85fd164691bb852f1cfaf82405df4380629ced6e

Comment 22

[cbarker]

Created 2 test cases in Moztrap, one for the Alarm Triggering whilst charging, and one for the Alarm Triggering whilst connected via USB.

See cases here:
https://moztrap.mozilla.org/manage/cases/?pagenumber=1&pagesize=20&sortfield=created_on&sortdirection=desc&filter-id=6834&filter-id=6835