Closed Bug 1009151 Opened 8 years ago Closed 8 years ago

[Geeksphone] please separate geeksphone crash reports until they resolve their driver issues

Categories

(Socorro :: Webapp, task)

ARM
Gonk (Firefox OS)
task
Not set
normal

Tracking

(Not tracked)

RESOLVED WORKSFORME

People

(Reporter: nhirata, Unassigned)

Details

(Whiteboard: [b2g-crash])

Geeksphone has some old driver issues which is causing a lot of crashing to occur.

Please separate the crashes so we can look at them separately.
What is the best way for us to identify geeksphone, does Android Manufacturer/Model have enough info? Could you please link to an example crash on crash-stats? Thanks!
Flags: needinfo?(nhirata.bugzilla)
I actually do not agree we should exclude them anywhere. Apart from that, Socorro doesn't even include them usefully anywhere yet, so no need to exclude. ;-)

I also heavily doubt the "old driver" scenario given that the last huge issue we had that you called out as one was actually a debugging mechanism turned on and blocked by seccomp.
That seccomp issue wasn't completely fixed.  It was closed off as "low repro rate".

As a discussion point, should we examine all of the geeksphone crashes?  I can file bugs for all of them if you like, at the same time I am not sure how many dev will look at those bugs.

Also to note, 1.4+ uses JB, 2.0 is suppose to go to Kitkat; I believe Geeksphone is still on ICS.  I have to double check this; having said that if they are at ICS, their gonk level is going to be out of date for some of the APIs to work correctly.  Esp. things such as NFC.
Flags: needinfo?(nhirata.bugzilla) → needinfo?(kairo)
(In reply to Naoki Hirata :nhirata (please use needinfo instead of cc) from comment #3)
> That seccomp issue wasn't completely fixed.  It was closed off as "low repro
> rate".

No, it was fixed, as discussed in the bug.

> As a discussion point, should we examine all of the geeksphone crashes?  I
> can file bugs for all of them if you like, at the same time I am not sure
> how many dev will look at those bugs.

Right now, the Geeksphone are the ones that actually produce data for 1.4 and 2.0 in at least some reasonable rate, while we do not have the Flame in many people's hands yet. So, I think we should at least take a look there.

> Also to note, 1.4+ uses JB, 2.0 is suppose to go to Kitkat; I believe
> Geeksphone is still on ICS.

Yes, they are in ICS because the closed-source-loving company that produced the chipset abandoned enabling new Android versions for that chipset before the Geeksphone and other first-wave phones were even shipped. Don't ask me about my personal opinion on that.
That said, I believe we are actually working against our principles if we do not support ICS-based builds at least to the end of 2014 - again, that's a personal opinion.


All that said, there is basically nothing to do in this bug - as I said, Socorro already ignores GP anyhow (as it does almost all our data) because of both the wrongly set channel and because Socorro hasn't adopted fully to the B2G landscape yet.
As for my custom reports, they already clearly sign which phones are affected so it's easy to see what is GP stuff, and those reports are only meant as a temporary stopgap until Socorro eventually will catch up.
Flags: needinfo?(kairo)
Oh.  you're right.  I can just filter by channel in the Super search.
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → WORKSFORME
(In reply to Naoki Hirata :nhirata (please use needinfo instead of cc) from comment #5)
> Oh.  you're right.  I can just filter by channel in the Super search.

Yes, that should work fine there as well (and hopefully GP will have channels some time that even label clearly it's theirs, so we'll be able to filter even better).
You need to log in before you can comment on or make changes to this bug.