820414 - [COST CONTROL] Data Usage UI crashes when changing date to a previous date

Status: RESOLVED FIXED

(Firefox OS Graveyard :: Gaia::Cost Control, defect)

Description

[Marcelino Veiga Tuimil [:sonmarce]]

STR:
1-Pull down notification tray
2-Tap in the data usage module
3-Tap in the wheel to enter settings
4-Stablish yesterday as the beginning of the billing cycle 
5-Change date to the day before yesterday (with adb shell date -s yyyymmdd.hhmmss)

Expected result --> Graphic and data usage module changes accordingly to the date, as you have change of billing cycle, data from previous billing cycle should be shown if available, or nothing

Actual result --> Data usage module shows data consumption from the billing cycle corresponding to the date before any change (real today), and graphic in data usage details shows no data (and no graphic in fact)

Comment 1

[Salvador de la Puente González [:salva]]

The problem is actually part of a more general affecting any change of the date to a previous one so rewording.

Nominating to tef? because if not, the applications is not able to return to a stable state. But if it lands, despite the information could be inaccurate, at least the UI does not crash and it's able to return to a stable situation after fixing the date.

Comment 2

[Michael Vines [:m1] [:evilmachines]]

(Just trying to assess the tef? here, does this only occur when the device date is moved backwards either by |adb shell| or perhaps NITZ while the cost control app is active?)

Comment 3

[Salvador de la Puente González [:salva]]

I was trying and it happen in both situations.

Comment 4

[Salvador de la Puente González [:salva]]

Attachment: Avoiding UI crash when changing the date to a previous date

This patch does not define a correct behavior for these situations, only prevent UI from crashing.

Comment 5

[Michael Vines [:m1] [:evilmachines]]

(patch looks low risk and of course would be nice to avoid the UI crashing in this case, so assuming the patch it doesn't grow in scope/risk on review then tef+.  If needed, let's raise another bug for doing more than avoiding the UI crash so we can potentially get a better fix in v1.1 or beyond)

Comment 6

[Salvador de la Puente González [:salva]]

(In reply to Michael Vines [:m1] [:evilmachines] from comment #5)
> If needed, let's raise another bug for doing more than avoiding
> the UI crash so we can potentially get a better fix in v1.1 or beyond)

Hi, it is just what we did in bug 838556

Comment 7

[Francisco Jordano [:arcturus] [:francisco]]

Comment on attachment 724595 [details]
Avoiding UI crash when changing the date to a previous date

r+ working well, great job as always Salva.

Comment 8

[Salvador de la Puente González [:salva]]

Master: 00f2a15d8995a12284f7038462249f819a8966cd

Comment 9

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

v1-train: 3b4626d5e2c58b9c5647cbd76264ae787cc2a38c

There are fairly large merge conflicts on v1.0.1.  Try doing this:

  git checkout v1.0.1
  git cherry-pick 3b4626d5e2c58b9c5647cbd76264ae787cc2a38c
  <resolve merge conflicts>
  git commit

Comment 10

[Salvador de la Puente González [:salva]]

This bug has a dependency on bug 841321. I got apply this one without the dependency but what do you want me to do? Consider including bug 841321 as well, please.

Comment 11

[Alex Keybl [:akeybl]]

(In reply to Salvador de la Puente González [:salva] from comment #10)
> This bug has a dependency on bug 841321. I got apply this one without the
> dependency but what do you want me to do? Consider including bug 841321 as
> well, please.

If bug 841321 is truly a blocker (I'm very skeptical), then you need to set blocking-b2g:tef? on that bug with an explanation. If it's not, please prepare a patch that will apply cleanly to v1.0.1 specifically.

Comment 12

[Salvador de la Puente González [:salva]]

I'm nominating the bug 841321 as tef?

If not passing the triage, I have a patch ready to be merged cleanly with 1.0.1

Comment 13

[Salvador de la Puente González [:salva]]

Bug 841321 has been marked as tef+, once it is merged, this will apply cleanly.

Comment 14

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

v1.0.1: 8c9339d5e0420132811b7e7d42424a1f64979bdc

Comment 15

[Deepa Subramanian]

The issue is still found on v1.0.1 but not fixed for v1.1. 

The data usage for wifi shows a huge negative number when we change the date to 3/30/2013 (later date). 

Issues repros on 
Unagi Build ID: 20130329070203
Kernel Date: Dec 5
Gecko: http://hg.mozilla.org/releases/mozilla-b2g18_v1_0_1/rev/56c922308fd1
Gaia: 0a9f78bffafda93a159c1f502e8b110c2f49a500

Comment 16

[Darren]

The issue should be covered by test case:
https://moztrap.mozilla.org/manage/cases/?filter-id=6910

Comment 17

[Deepa Subramanian]

Sorry, I wanted comment saying the issue is NOT found on v1.0.1 but Found in v1.1 when I tried to repro this issue on build 20130329070203.