798451 - Provide alternative to wifi crash submission

Status: RESOLVED WONTFIX

(Firefox OS Graveyard :: General, defect)

Description

[David Bolter [:davidb] (NeedInfo me for attention)]

Kairo mentioned that Brazil doesn't have much wifi coverage at all. We should probably figure out how we are going to get telemetry and crash data in the wild. Filing this bug so we can make a conscious decision whether we can do anything here.

Comment 1

[Hubert Figuiere [:hub]]

It should be noted that currently we will submit over cellular network if it is the only one available.

And there was mention of disabling this. I wonder if the option shouldn't be in the "carrier profiles" (I don't know how we manage these)

Comment 2

[Robert Kaiser]

(In reply to Hub Figuiere [:hub] from comment #1)
> It should be noted that currently we will submit over cellular network if it
> is the only one available.

Right now, we don't implement the final UX flow at all yet, because we're still waiting on a final decision on some detail there, but in that UX there will be a requirement to send only over wifi (same is true for telemetry, btw).

This bug was filed as a reaction to that requirement and our UX researchers apparently saying that there's almost zero wifi coverage in Brazil (I'm paraphrasing here, I haven't seen their session because it was too late at night for me).

Comment 3

[David Bolter [:davidb] (NeedInfo me for attention)]

Morphing bug, thanks Hub!

Comment 4

[David Bolter [:davidb] (NeedInfo me for attention)]

Kairo, sorry I mid-air collided, please feel free to further morph the summary.

Comment 5

[Robert Kaiser]

(In reply to David Bolter [:davidb] from comment #4)
> Kairo, sorry I mid-air collided, please feel free to further morph the
> summary.

I think the original is fine once we implement what's in the upcoming UX wireframes. See page 4 in http://people.mozilla.com/~lco/Settings_B2G/Crash%20Reports%20concepts%20v2.pdf (those are the WIP concepts and have been posted to dev-gaia, which is public, so I guess we can link here as well).

Comment 6

[Dietrich Ayala (:dietrich)]

Is this optional for V1, or required? If crash submissions could be sent over paid connections, please nominate this for blocking-basecamp.

Comment 7

[Henrik Skupin [:whimboo][⌚️UTC+1]]

(In reply to Dietrich Ayala (:dietrich) from comment #6)
> Is this optional for V1, or required? If crash submissions could be sent
> over paid connections, please nominate this for blocking-basecamp.

Not sure what the correct next action is here and who can answer that. But nominating for blocking so it can be triaged correctly or send to the proper person.

Comment 8

[Robert Kaiser]

I don't think we can or need to get anything done here for v1. We will not have much crash data from the field in Brazil, but we'll need to live with that. It's even more important to get data from our smoketesters and they should have wifi available at least in some places.

Comment 9

[Henrik Skupin [:whimboo][⌚️UTC+1]]

It highly depends on where you are testing the device. It's a mobile phone so I wouldn't assume that much of the testing will happen at home where wifi is available. All the crashes I have faced so far happened while I was out of home and no wifi network around.

But hey, I found a nice workaround at least for our dogfooding folks. Simply 'adb pull' all the files under '/data/b2g/mozilla/Crash Reports/pending' to your local disk and stick the files into the Firefox crash reports folder. You can submit the crashes with Firefox. Here an example which shows that it works: bp-9bbea992-da20-4cc0-b224-aa5aa2121213

Comment 10

[Robert Kaiser]

(In reply to Henrik Skupin (:whimboo) from comment #9)
> It highly depends on where you are testing the device. It's a mobile phone
> so I wouldn't assume that much of the testing will happen at home where wifi
> is available. All the crashes I have faced so far happened while I was out
> of home and no wifi network around.

People don't need to test the phone where wifi is available, we queue up the crash reports and send them once they connect to a wifi network (at least that's the design, we haven't yet done full QA on this working as it should) - if you use a build before bug 814078, we only send the recent crash when connecting, but after that, we should send all queued reports.

Comment 11

[Henrik Skupin [:whimboo][⌚️UTC+1]]

(In reply to Robert Kaiser (:[e-mail]) from comment #10)
> People don't need to test the phone where wifi is available, we queue up the
> crash reports and send them once they connect to a wifi network (at least
> that's the design, we haven't yet done full QA on this working as it should)
> - if you use a build before bug 814078, we only send the recent crash when
> connecting, but after that, we should send all queued reports.

Oh, that changes things totally. Yes I can see that bug 814078 works for me after the update to a newer build. Thanks a lot. So I will reset the basecamp flag for now given that it doesn't sound like a blocker anymore.

Comment 12

[BMO Automation]

Firefox OS is not being worked on