[Flame] B2G brick after flashing master Gecko/Gaia to Flame with v123 roms

RESOLVED WONTFIX

Status

RESOLVED WONTFIX
4 years ago
5 months ago

People

(Reporter: vicamo, Unassigned)

Tracking

unspecified
ARM
Gonk (Firefox OS)

Firefox Tracking Flags

(b2g-v2.1 affected, b2g-v2.2 affected)

Details

Attachments

(1 attachment)

(Reporter)

Description

4 years ago
Created attachment 8495248 [details]
setringtone-no-manifest.log

+++ This bug was initially created as a clone of Bug #937485 +++

STR:
1) Flash Flame with ROMs v123
2) Build B2G Flame from master branch
3) Flash Gecko and Gaia

Adb logcat contains:

  Full message: TypeError: invalid 'in' operand aManifest
  Full stack: this.DOMApplicationRegistry.updateDataStore@resource://gre/modules/Webapps.jsm:701:1

With additional debug prints added, that nullized aManifest is returned with aId = 'setringtone'. It seems master branch Gaia has no such app any more, but it does exist in Flame v123 gonk images. Somehow /system/b2g/webapps/setringtone has been removed during flashing Gaia, but the apps metadata are not updated correctly.

See bug 937485 comment 34 - 36.
(Reporter)

Updated

4 years ago
status-b2g-v2.1: --- → affected
status-b2g-v2.2: --- → affected
(Reporter)

Updated

4 years ago
Flags: needinfo?(squibblyflabbetydoo)
Flags: needinfo?(kyle)
Flags: needinfo?(fabrice)
Flags: needinfo?(dflanagan)
I've never had issues with this, even when flashing various ROMs, but yes, setringtone doesn't exist anymore as an app.

(Also, aren't we on v180 for the ROMs now?)
Flags: needinfo?(squibblyflabbetydoo)
(Reporter)

Comment 2

4 years ago
(In reply to Jim Porter (:squib) from comment #1)
> I've never had issues with this, even when flashing various ROMs, but yes,
> setringtone doesn't exist anymore as an app.

This only happens when you flash only Gecko & Gaia. It can be worked around by either reset Gaia, which means data lose, or place an additional return statement in Gecko WebApps.jsm as attachment 8495242 [details] [diff] [review] does.

> (Also, aren't we on v180 for the ROMs now?)

That's for KitKat. General users don't have flame-kitkat build yet. v123 is the latest JB Gonk.
[Blocking Requested - why for this release]: We need this fixed asap for existing users.
blocking-b2g: --- → 2.1?
Get your Flame KK base image (v180) from https://developer.mozilla.org/en-US/Firefox_OS/Developer_phone_guide/Flame, it's available to all.  All should update as the last real JB product was based on B2G v1.3
Vicamo, when you update gecko and gaia, is that an atomic update? Or are you 1) updating gecko 2) restart 3) updating gaia?

I agree in general that we should not fail to startup but non atomic updates are known to be "at your own risk". So clearly not a release blocker imho.

If I can repro I'll work on a fix (the proposed patch is more a band aid than a proper fix).
Flags: needinfo?(fabrice)
Removing ni? since I'm not sure how I'm involved here. Re ni? if I'm missing something.
Flags: needinfo?(kyle)
(Reporter)

Comment 7

4 years ago
(In reply to Fabrice Desré [:fabrice] from comment #5)
> Vicamo, when you update gecko and gaia, is that an atomic update? Or are you
> 1) updating gecko 2) restart 3) updating gaia?

Flash Gecko, wait Gecko runs for a while to stablize, then flash Gaia. Actually either case bricks.

One may either `make -C gaia reset-gaia` or flash all (`./flash.sh`) instead to get riddle of the forever blank screen.

> I agree in general that we should not fail to startup but non atomic updates
> are known to be "at your own risk". So clearly not a release blocker imho.

For me, only flash whole system is atomic. Neither does `./flash.sh gecko && ./flash.sh gaia` count. But that means user data lose. :(

> If I can repro I'll work on a fix (the proposed patch is more a band aid
> than a proper fix).
(Reporter)

Comment 8

4 years ago
(In reply to Kyle Machulis [:kmachulis] [:qdot] (USE NEEDINFO?) from comment #6)
> Removing ni? since I'm not sure how I'm involved here. Re ni? if I'm missing
> something.

I found your name appears in the history several times recently. And it's qDot. It's quite reasonable that he's involved. :P
(Reporter)

Updated

4 years ago
Flags: needinfo?(dflanagan)
(In reply to Vicamo Yang [:vicamo][:vyang] from comment #7)
> (In reply to Fabrice Desré [:fabrice] from comment #5)
> > Vicamo, when you update gecko and gaia, is that an atomic update? Or are you
> > 1) updating gecko 2) restart 3) updating gaia?
> 
> Flash Gecko, wait Gecko runs for a while to stablize, then flash Gaia.
> Actually either case bricks.

Do you mean that just flashing gecko bricks also?


> For me, only flash whole system is atomic. Neither does `./flash.sh gecko &&
> ./flash.sh gaia` count. But that means user data lose. :(

Not "in the wild" users. Only people that do updates themselves with their own builds.
(Reporter)

Comment 10

4 years ago
(In reply to Fabrice Desré [:fabrice] from comment #9)
> (In reply to Vicamo Yang [:vicamo][:vyang] from comment #7)
> > (In reply to Fabrice Desré [:fabrice] from comment #5)
> > > Vicamo, when you update gecko and gaia, is that an atomic update? Or are you
> > > 1) updating gecko 2) restart 3) updating gaia?
> > 
> > Flash Gecko, wait Gecko runs for a while to stablize, then flash Gaia.
> > Actually either case bricks.
> 
> Do you mean that just flashing gecko bricks also?

1) Flash gecko only => brick
2) Flash gecko, wait, flash gaia => brick
3) Flash gecko, wait, flash gaia, wait, reboot => still brick
4) Flash gecko, wait, flash gaia, wait, reset-gaia => no brick
5) `./flash.sh` => no brick
Seems this affects the Open C as well, and the attachment in comment #2 fixes it. Seems like a pretty reasonable and trivial patch, can we not just get it reviewed and in and fix this?

Updated

4 years ago
No longer blocks: 937485
Its unclear this is an issue in engg builds or something else. As chris said in comment #11, I am happy to land the fix here if it helps people.

TOny, has QA come across this or do you have any input here?
Flags: needinfo?(tchung)
(In reply to bhavana bajaj [:bajaj] from comment #12)
> Its unclear this is an issue in engg builds or something else. As chris said
> in comment #11, I am happy to land the fix here if it helps people.
> 
> TOny, has QA come across this or do you have any input here?

QA was not doing full flash against v123 since that is a JB product.   We have stopped support on v123, and since moved to v180 and up, only focusing on full flashing master, 2.1, and 2.0 builds (if needed).   Please shift users off of v123, and have them update the base image on v180 going forward.   The v180 base is publically shared on MDN.
Flags: needinfo?(tchung)
Per comment #13, this v123 is old, removing blocking request.
blocking-b2g: 2.1? → ---

Updated

a year ago
Product: Core → Core Graveyard
Closing as we are not working on Firefox OS anymore.
Status: NEW → RESOLVED
Last Resolved: 5 months ago
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.