Closed Bug 934690 Opened 6 years ago Closed 6 years ago

[B2G][FTU] Final screen of FTU does not have image


(Firefox OS Graveyard :: Gaia::First Time Experience, defect)

Gonk (Firefox OS)
Not set


(blocking-b2g:koi+, firefox28 unaffected, b2g-v1.2 fixed, b2g-v1.3 unaffected)

1.2 C5(Nov22)
blocking-b2g koi+
Tracking Status
firefox28 --- unaffected
b2g-v1.2 --- fixed
b2g-v1.3 --- unaffected


(Reporter: gbennett, Assigned: eragonj)



(Keywords: regression, Whiteboard: [FT:System-Platform])


(4 files)

Attached image 2013-11-04-14-16-51.png
"Tap and hold the home button to browse and close recent apps" screen does not contain an image.

Repro Steps:
1) Updated Buri 1.2 comRIL to BuildID: 20131103004003
2) Restart phone.
3) Go through FTU to task manager page.

Image is not shown for task manager tutorial.

Proper image is shown with the text.

Environmental Variables:
Device: Buri 1.2 comRIL
BuildID: 20131103004003
Gaia: cb981e2f47bc644b4d178d54378c3676c946facf
Gecko: eec4da1b27eb
Version: 26.0
Firmware Version: 20131015

Repro frequency: 100%, 4/4
See attached: 2013-11-04-14-16-51.png
regression.  flagging for koi? triage.
blocking-b2g: --- → koi?
fwiw, this is working on 1.1 and 1.3, so something seems isolated to 1.2.  still need regression range
Issue did not happen on build 11/01 and was introduced on build 11/02. On build 11/01 all tutorial pictures are displayed properly, but on build 11/02 the tutorial is missing the Card View picture.

Issue did not occur -
Device: Buri v1.2 COM RIL
Build ID: 20131101004000
Gaia: e717aec947571f5daf923c040a82f9f0719bb526
Platform Version: 26.0
RIL Version: 
Base Build: 20131015

Issue occurs -
Device: Buri v1.2 COM RIL
Build ID: 20131102004003
Gaia: 2e5dc6b9600bf2bc2ff3fc359120e2883ef50e38
Platform Version: 26.0
RIL Version: 
Base Build: 20131015

Looks like the issue was introduced over the weekend.
Can we get a logcat and an adb shell dmesg please?

Chances are it may be related to OOM and Bug 923303/Bug 905882

It seems like we need to get 1.2 blob updated.  I find it odd that 1.3 is not affected.
Flags: needinfo?(gbennett)
Attached are the logcat and dmesg. Don't see any memory issues in the log, but still might be related.
Flags: needinfo?(gbennett)
What's before : 11-04 15:32:10.792: I/Gecko(140): ###!!! [Parent][AsyncChannel] Error: Channel error: cannot send/recv ?
Flags: needinfo?(gbennett)
I cleared the logcat right before I went into the tour area so that message showed up on the first tour page. I'm adding a full log of the entire FTU experience and I see some OOM messages, but that might be from the exporting contacts page I landed on after completing the dev forced FTU (which also repros the issue).
Flags: needinfo?(gbennett)
Not a smoketest blocker.
Keywords: smoketest
blocking-b2g: koi? → koi+
Ni on Lucas to help find an assignee here.
Flags: needinfo?(ladamski)
Flags: needinfo?(ladamski)
Whiteboard: [systemsfe]
Assignee: nobody → mhenretty
It looks like the changes from bug 911681 pulled in the ftu step text from bug 916262. However, bug 916262 was never approved for 1.2. This means that the images for ftu don't match their step instructions now.

We either will need to uplift bug 916262 to 1.2, or rework 911681 to not include the step in their tutorial steps.

Fred, you seemed to be in charge of bug 911681. Seems like we should rework this to exclude stuff. This also means making sure tablet images as well as phone images are correct. Can you comment here on what should be done?
Blocks: 911681
Flags: needinfo?(gasolin)
EJ is more involved in FTU dev, he's the proper person to answer that ni in detail
Flags: needinfo?(gasolin)
Hi Michael, 

I didn't know that its depended bug is not landed on 1.2. After asking Fred, it seems that has been moved out of 1.2 scope so I think I can add a 1.2 branch specific fix for this.

My proposal is to remove strings and change the step count in 1.2 branch. 

And by the way, I think it would be better to leave bug 911681 closed and track related changes in this bug.

What do you think ?
(In reply to EJ Chen [:eragonj] from comment #12)
> My proposal is to remove strings and change the step count in 1.2
> branch. 

Sounds good to me! I'll re-close the other bug. Thanks for taking this!
Assignee: mhenretty → ejchen
Attached file Bug 934690 patch
Borja, I just updated FTU slightly on v1.2 branch and need your review. 

Please help me when you have time !

Thanks :D
Attachment #832000 - Flags: review?(borja.bugzilla)

there is a related UX change on bug 937495, I will update images / strings when landing it.

You can review this one first and I will do following changes on that bug .

Thanks :)
Comment on attachment 832000 [details] [review]
Bug 934690 patch


After checking more details, 

I thinks bug 937495 is more complicated than this one and this bug may depend on that. 

I will ask for review later. 

Thanks :)
Attachment #832000 - Flags: review?(borja.bugzilla)
Comment on attachment 832000 [details] [review]
Bug 934690 patch

Hi Borja, 

I have updated the code for this bug. (Because we can't change l10n in this stage, I put some extra logics to detect and use the right strings.)

And by the way, because we (MoCo) decided to use v1.3 in flatfish, there is no need to consider the problems caused by bug 937495 (I will fix them on that bug). 

Please help me review this patch ! Thanks :D
Attachment #832000 - Flags: review?(borja.bugzilla)
Target Milestone: --- → 1.3 Sprint 5 - 11/22
Switching to System Platform per the assignee change
Whiteboard: [systemsfe] → [FT:System-Platform]
See Also: → 939310
Duplicate of this bug: 939310
See Also: 939310
koi+ is for 1.2. We should use 1.2 target milestone rather than 1.3 target milestone.
Target Milestone: 1.3 Sprint 5 - 11/22 → 1.2 C5(Nov22)
EJ, are you still working on this?
Flags: needinfo?(ejchen)
(In reply to Tim Guan-tin Chien [:timdream] (MoCo-TPE) (please ni?) from comment #21)
> EJ, are you still working on this?

I am waiting for Borja's review :D
Flags: needinfo?(ejchen)
Try to make the whole story clear, I will keep some notes about it below.

At first, we decided to use v1.2 as our basic version for flatfish. On bug 911681, UX needs a customized step on FTU asking users swipe up at the last step to leave FTU. In order to implement this feature, we added IAC_handler to make sure we can communicate between communications app and system app to achieve this, and this was landed on master and v1.2 finally.

But sadly, bug 911681 is depended on bug 916262. Originally related change will be landed on v1.2 so there is no problem on this, but finally, related bugs are not landed on v1.2 due to scope changes. It makes bug 916262 not landed on v1.2 while bug 911681 is landed, so this mistake caused bug 934690 happened ( there is a missing step on FTU ).

In order to fix bug 934690, I have to add some extra logics for it because l10n strings are fixed in this stage. When waiting for review of bug 934690, there is another bug - bug 934011 happened talking about "Force closing the Music app while music is playing results in notification media controls not working on subsequent launches". 

The reason why bug 934011 happened is because we cached IAC ports so that they are using old ports. At the same time, there is a version change for flatfish that we will use v1.3 instead of v1.2. In this way, we can back out bug 911681 to fix bug 934011 and leave it on v1.3. If bug 911681 is really backed out, bug 934690 will be invalid in the end. 

Hope this note can make the others know what's going on about these bugs.
Interesting.  on a side note, the logcat does indicate OOM.  I guess we'll wait for everything to land with the OEM 1.2 and these patches to see if it all gets fixed.
(In reply to EJ Chen [:eragonj] from comment #23)
> Try to make the whole story clear, I will keep some notes about it below.

After some follow-up discussions on bug 934011, it seems that the risk to back out bug 911681 is too high and after some tests, bug 934011 may not cause the problem as it is described on v1.2, but as Jim said, it would be better to land it on v1.2 to make IAC_handler solid. 

To be short, I think this bug can be discussed separately with the others because we won't back out bug 911681, so in this way, this bug has to be fixed later after review.
(In reply to Naoki Hirata :nhirata (please use needinfo instead of cc) from comment #24)
> Interesting.  on a side note, the logcat does indicate OOM.  I guess we'll
> wait for everything to land with the OEM 1.2 and these patches to see if it
> all gets fixed.

Interesting ... not sure whether there is any relationship between this bug and OOM or not. But on my side, this patch fixed the problem caused by not landing features. Maybe we can try again with all patches landed as you mentioned to see whether these problems still exist or not. 

Thanks for pointing out this :D
I would like to understand more the issue here. FTU is an app that is running *only* during the fist run of the device, so the 'task manager' can not be launched during the FTU. That's why I think that the STR (Steps To Reproduce) it's not accurate, or we need a complementary explanation. Ej, could you help me with this? Thanks!
Flags: needinfo?(ejchen)
Yeah Borja, 

I think STR is wrong. The proper story would be the one I told you just now. (check comment #23) 

If you have any problem about the patch or something, just let me know. 

Thanks :D
Flags: needinfo?(ejchen)
Comment on attachment 832000 [details] [review]
Bug 934690 patch

Code looks great, but we need to make Travis to be green! So let's fix the Gaia-ui-test! :)
Attachment #832000 - Flags: feedback+
Comment on attachment 832000 [details] [review]
Bug 934690 patch

Muchas gracias!!!! :)
Attachment #832000 - Flags: review?(borja.bugzilla) → review+
Closed: 6 years ago
Resolution: --- → FIXED
yeah ~

Thanks Borja, jajajajaja xD

(In reply to Borja Salguero [:borjasalguero] from comment #31)
> 0cfab8d01b4640f6182f51acf09947408e49f3f5
> c2dea53b36bb9d4331a94976344515f60dc5a3d4
> R+. Merged.
Verified fixed on keon device.
Does this need a fix on v1.3 or master branch?  c2dea53b36bb9d4331a94976344515f60dc5a3d4 is only on v1.2 and v1.2f
Flags: needinfo?(ejchen)
Hi John Ford,

there is no need to uplift this patch in v1.3 because the bug only exists in v1.2*.

Thanks for your help :)
Flags: needinfo?(ejchen)
You need to log in before you can comment on or make changes to this bug.