990271 - L10n calls before l10n context is ready

Status: RESOLVED WONTFIX

(Firefox OS Graveyard :: Gaia::System::Lockscreen, defect)

Description

[Zibi Braniecki [:zbraniecki][:gandalf]]

LockScreen.refreshClock unconditionally calls L10n.get functions without waiting for l10n resources to be available.

The stack:
 
* LockScreen.refreshClock - https://github.com/mozilla-b2g/gaia/blob/master/apps/system/js/lockscreen.js#L848-L851

* Clock.start - https://github.com/mozilla-b2g/gaia/blob/master/apps/system/js/statusbar.js#L38

* LockScreen.handleEvent(screenchange) - https://github.com/mozilla-b2g/gaia/blob/master/apps/system/js/lockscreen.js#L170

* ScreenManager.fireScreenChangeEvent - https://github.com/mozilla-b2g/gaia/blob/master/apps/system/js/screen_manager.js#L575

* ScreenManager.turnScreenOn - https://github.com/mozilla-b2g/gaia/blob/master/apps/system/js/screen_manager.js#L428

* Startup - https://github.com/mozilla-b2g/gaia/blob/master/apps/system/js/bootstrap.js#L70

* window.onload

We either should block turning screen on until we have the resources available, or we delay refreshing LockScreen Clock until that.

Comment 1

[Zibi Braniecki [:zbraniecki][:gandalf]]

Greg, you seem to be the last editor of this code. Can you advise on the route here or point me to someone who can?

Comment 2

[Greg Weng [:snowmantw][:gweng][:λ]]

I have no clue now. One thing now I've found is that:

https://github.com/mozilla-b2g/gaia/blob/master/apps/system/js/lockscreen.js#L418

This would wait for L10n ready and then

https://github.com/mozilla-b2g/gaia/blob/master/apps/system/js/lockscreen.js#L449

Call the clock to refresh, and 

https://github.com/mozilla-b2g/gaia/blob/master/apps/system/js/lockscreen.js#L863

Get the L10n resource. So if this flow is actually followed, this situation should not happen.
Unless there is some error code which are out of this control.

However I must said I didn't modify L10n since the first refactoring (except re-arrange the code order). So I must take some time to investigate this.

Can you post some raw log or STR to help me reproduce this on my local machine?

Comment 3

[Zibi Braniecki [:zbraniecki][:gandalf]]

Yeah,

Just apply this patch https://pastebin.mozilla.org/4742076 and restart the phone with default locale.

Comment 4

[Zibi Braniecki [:zbraniecki][:gandalf]]

The patch I'm using to make this error not appear is https://pastebin.mozilla.org/4742097 - but I think that they may be a better solution.

Comment 5

[Greg Weng [:snowmantw][:gweng][:λ]]

I've found that the line:

https://github.com/mozilla-b2g/gaia/blob/master/apps/system/js/lockscreen.js#L170

may call refreshing the clock while the L10n is not ready. But I still need a investigation.

Comment 6

[Zibi Braniecki [:zbraniecki][:gandalf]]

Any progress on that?

Comment 7

[Greg Weng [:snowmantw][:gweng][:λ]]

Since I'm now resolving some tough regressions I've no change to solve this bug yet...

I can only (sincerely) guarantee I would take a look ASAP...

Comment 8

[Zibi Braniecki [:zbraniecki][:gandalf]]

ok. I can help you. I just need to know if we can delay the clock refresh until resources are loaded, or do you prefer to just skip the string updating in the clock_refresh function until resources are available?

Comment 9

[Greg Weng [:snowmantw][:gweng][:λ]]

Thanks. I think the later may be better.

Comment 10

[Zibi Braniecki [:zbraniecki][:gandalf]]

hmm.. it seems that when I skip the localization, the function is empty. Which means that it burns CPU cycles.

I can also just start the interval when l10n resources become available.

Does it make sense to you?

Comment 11

[Greg Weng [:snowmantw][:gweng][:λ]]

OK. We can discuss more detail when you done your patch. Thanks for your effort.

Comment 12

[BMO Automation]

Firefox OS is not being worked on