Closed Bug 1024470 Opened 10 years ago Closed 10 years ago

[Flame][v2.1] The country and city changes not visible in the date & time FTU menu

Categories

(Core :: General, defect)

ARM
Gonk (Firefox OS)
defect
Not set
normal

Tracking

()

VERIFIED FIXED
mozilla33

People

(Reporter: RobertC, Assigned: kanru)

References

Details

(Keywords: qablocker, regression, smoketest, Whiteboard: [fromAutomation][xfail])

Attachments

(1 file)

Changing the country and the city in the time zone menu has no visible effect.

This issue was reproduced with both manual and automated tests.

Build info(flame):
Gaia      41db6954a67efc55016744bc8f6591ae9e07a285
Gecko     https://hg.mozilla.org/mozilla-central/rev/9e8e3e903484
BuildID   20140612040203
Version   33.0a1
ro.build.version.incremental=94
ro.build.date=Tue May 20 09:29:20 CST 2014


STR:
1. open FTU app
2. tap next until you reach Date & Time menu
3. change country and city

Expected results:
The values for country and city are updated

Actual results:
No changes are visible.

Note: Changing the values for date and time will also update the country/city values to the correct ones. 
This issue is not reproduced on v1.4
We need to know if this happens on 2.0 & what the last working & first broken build is in automation.
blocking-b2g: --- → 2.1?
Keywords: qawanted
QA Contact: rpribble
I am unable to repro this issue on the Flame v2.0 MOZ ril.

v2.0 Environmental Variables:
Device: Flame v2.0 MOZ
BuildID: 20140612000201
Gaia: 2bb66630315299ca947e40fcec23c9f7ea012111
Gecko: 670d69879f0e
Version: 32.0a2
Firmware Version: v10G-2

Changes made to the date & time screen occur as expected and are saved if the user checks them in the settings after completing the FTE.

Leaving qawanted tag for the automation regression window.
QA Whiteboard: [QAnalyst-Triage?]
Flags: needinfo?(jmitchell)
Flame 2.1 = Repro
Flame 2.0 = NO repro
Flame 1.4 = NO repro

QA-Wanted to Check the Buri branch
Flags: needinfo?(jmitchell)
QA Contact: rpribble → jmercado
This issue occurs in Mozilla-Inbound during a period where there was no builds for over 26 hours.  As such the Pushlog for that window is quite large, more so than the Central Window.  I am also including both windows in case Central is easier to work with.

Central Regression Window:

Last working
Environmental Variables:
Device: Flame
BuildID: 20140611072409
Gaia: 0750f66a0004870773c9a743fa6bdbe124379336
Gecko: 429f55154e83
Version: 33.0a1

First Broken
Environmental Variables:
Device: Flame
BuildID: 20140611081515
Gaia: 0750f66a0004870773c9a743fa6bdbe124379336
Gecko: c7391b84d9c2
Version: 33.0a1

Last working gaia / First broken gecko - Issue DOES occur
Gaia: 0750f66a0004870773c9a743fa6bdbe124379336
Gecko: c7391b84d9c2

First broken gaia / Last working gekko - Issue does NOT occur
Gaia: 0750f66a0004870773c9a743fa6bdbe124379336
Gecko: 429f55154e83

Gecko Pushlog: http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=429f55154e83&tochange=c7391b84d9c2




Mozilla-inbound Regression Window

Last working
Environmental Variables:
Device: Flame
BuildID: 20140610081416
Gaia: f42ebc93554979501d3ac52bcf9e69cb4b310a4f
Gecko: 00d5164a4049
Version: 33.0a1

First Broken
Environmental Variables:
Device: Flame
BuildID: 20140611101104
Gaia: 508c85713fdf35eb2b3c8e8340e1d2ee306e6aec
Gecko: e1562c119643
Version: 33.0a1

Last working gaia / First broken gecko - Issue DOES occur
Gaia: f42ebc93554979501d3ac52bcf9e69cb4b310a4f
Gecko: e1562c119643

First broken gaia / Last working gecko  - Issue does NOT occur
Gaia: 508c85713fdf35eb2b3c8e8340e1d2ee306e6aec
Gecko: 00d5164a4049

Gecko Pushlog:  http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=00d5164a4049&tochange=e1562c119643
Flags: needinfo?(jmitchell)
Can we try bisecting this using Buri builds instead?
Flags: needinfo?(jmitchell)
Mozilla-inbound Regression Window

Last working 
Environmental Variables:
Device: Buri
BuildID: 20140610222008
Gaia: 0750f66a0004870773c9a743fa6bdbe124379336
Gecko: b35a6797354b
Version: 33.0a1

First Broken 
Environmental Variables:
Device: Buri
BuildID: 20140610224608
Gaia: 0750f66a0004870773c9a743fa6bdbe124379336
Gecko: a72228f459ec
Version: 33.0a1

Last working gaia / First broken gecko - Issue DOES occur
Gaia: 0750f66a0004870773c9a743fa6bdbe124379336
Gecko: a72228f459ec

First broken gaia / Last working gecko - Issue does NOT occur
Gaia: 0750f66a0004870773c9a743fa6bdbe124379336
Gecko: b35a6797354b

Gecko Pushlog:  http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=b35a6797354b&tochange=a72228f459ec
Flags: needinfo?(jmitchell)
Broken by bug 879475.
Blocks: 879475
Component: Gaia::First Time Experience → General
Product: Firefox OS → Core
Version: unspecified → Trunk
Kan-Ru - Can you either fix this quickly or backout?
Flags: needinfo?(kchen)
QA Whiteboard: [QAnalyst-Triage?] → [QAnalyst-Triage+]
Flags: needinfo?(jmitchell)
(In reply to Jason Smith [:jsmith] from comment #8)
> Kan-Ru - Can you either fix this quickly or backout?

We can't backout this. I'll fix this asap.
Flags: needinfo?(kchen)
QA Whiteboard: [QAnalyst-Triage+] → [QAnalyst-Triage+][lead-review+]
Make sure that we run mozilla::dom::time::InitializeDateCacheCleaner only in the chrome process and run InitOnContentProcessCreated even prelaunch process is disabled.
Assignee: nobody → kchen
Status: NEW → ASSIGNED
Attachment #8441956 - Flags: review?(khuey)
After the introduction of the Nuwa process, it's easy to break such cross-process observers. In bug 965912 we fail to run mozilla::dom::time::InitializeDateCacheCleaner() to register the observer for system time changes. In this bug, we call mozilla::dom::time::InitializeDateCacheCleaner() too early and make the second call fail because it is already initialized and just skip the whole stuff in it.

We may consider cleaning up the old instance of mozilla::dom::time::DateCacheCleaner() if mozilla::dom::time::InitializeDateCacheCleaner() is called the second time. It looks safe to do it except that the parent side will notice that some unknown observer needs to be unregistered and spits out a warning in http://mxr.mozilla.org/mozilla-central/source/hal/Hal.cpp#204
Comment on attachment 8441956 [details] [diff] [review]
Make sure we run InitOnContentProcessCreated in all cases

Review of attachment 8441956 [details] [diff] [review]:
-----------------------------------------------------------------

::: dom/ipc/ContentChild.cpp
@@ +674,5 @@
>      nsRefPtr<SystemMessageHandledObserver> sysMsgObserver =
>          new SystemMessageHandledObserver();
>      sysMsgObserver->Init();
>  
> +    InitOnContentProcessCreated(/* AfterNuwaFork = */false);

aAfterNuwaFork

@@ +1840,5 @@
>              toplevel = toplevel->getNext();
>          }
>  
>          // Perform other after-fork initializations.
> +        InitOnContentProcessCreated(/* AfterNuwaFork = */true);

and here
Attachment #8441956 - Flags: review?(khuey) → review+
Keywords: qablocker
Whiteboard: [fromAutomation] → [fromAutomation][xfail]
https://hg.mozilla.org/mozilla-central/rev/df16d540f143
Status: ASSIGNED → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla33
Issue verified fixed on 2.1 Flame

Result: The values for country and city are updated

Environmental Variables:
Device: Flame 
Build ID: 20140620040204
Gaia: ccd70903544486bea04e85d8a4aacf63f1de2a72
Gecko: bdac18bd6c74
Version: 33.0a1  
Firmware Version: v121
Status: RESOLVED → VERIFIED
Already on 2.1
blocking-b2g: 2.1? → ---
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: