1052299 - [b2g][l10n] SIM PIN prompt and PUK prompt are not entirely localized on 2.0

Status: RESOLVED FIXED

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

Description

[Peipei Cheng (needinfo if you need my action)]

Attachment: 2014-08-12-15-58-31.png

1. Enable a SIM PIN and restart the phone.

2. When prompted to enter a SIM PIN, put in the incorrect pin 3 times.

3. Check all the strings on the "Enter PUK code" page.
   --> "Enter PUK code" page was not fully translated. But if you go back to SIm security page and try to disable PIN, the "Enter PUK code" page is fully translated.

Comment 1

[Peipei Cheng (needinfo if you need my action)]

(In reply to Peipei Cheng from comment #0)
> Created attachment 8471421 [details]
> 2014-08-12-15-58-31.png
> 
> 1. Enable a SIM PIN and restart the phone.
> 
> 2. When prompted to enter a SIM PIN, put in the incorrect pin 3 times.
> 
> 3. Check all the strings on the "Enter PUK code" page.
>    --> "Enter PUK code" page was not fully translated. 

But if you go back to
> SIm security page and try to disable PIN, the "Enter PUK code" page is fully
> translated.

if you lock screen and then unlock, the page become untranslated.

Comment 2

[Shaohua Wen]

I cannot find the untranslated strings, could it be a i18n issue?

Comment 3

[Francesco Lodolo [:flod]]

Théo, are you able to verify this? Not sure I have expendable SIMs around in case something goes wrong with PIN/PUK codes.

Comment 4

[Théo Chevalier]

Tried with latest 2.1 build on Flame and can't reproduce.

I suspect only 2.0 is affected (all locales), and the possible cause could be 2.0 patch from bug 1030094 not fixing the issue on 2.0 but working only on 2.1.
https://bugzilla.mozilla.org/show_bug.cgi?id=1030094#c40

Keeping needinfo, will check on 2.0 tomorrow at the office.

Comment 5

[Shaohua Wen]

@Theo any updates?

Comment 6

[Théo Chevalier]

Attachment: Flame 2.0, fr

This is confirmed, SIM PIN prompt (not the one on FTU, the one you get when you reboot) contains en-US strings (see screenshot) on 2.0 only.


Tested with:
Gaia      295c7f50707372e5af6d8e83148d2d970076dfd6
Gecko     https://hg.mozilla.org/releases/mozilla-b2g32_v2_0/rev/879c5208084f
BuildID   20140815000200
Version   32.0
ro.build.version.incremental=110
ro.build.date=Fri Jun 27 15:57:58 CST 2014
B1TC00011230

Comment 7

[Théo Chevalier]

[Blocking Requested - why for this release]:

Since labels to create PIN/PUK codes are involved and it's the first screen seen by users, I'd ask to block on this one :/
All locales are affected.

Comment 8

[Axel Hecht [:Pike]]

PS: We have those strings, it seems that the steps-to-reproduce show a code path that doesn't pick up the strings from l10n.

Thus, this doesn't smell like late-l10n at this point.

Comment 9

[Gregor Wagner [:gwagner]]

Aus, Ted, if I remember correctly one of you fixed a similar bug already. Can you take a look?

Comment 10

[Aus Lacroix [:aus]]

Ah yes, this is likely caused by misuse of the l10n API. I can take a look.

Comment 11

[Chris Kreinbring [:CKreinbring]]

Regression window unavailable.
Builds before the build noted below do not recognize the SIM card in either slot.

BuildID: 20140506163010
Gaia: 98ca8c55dbe2f21a8661d0eaa87f34d316c3bc98
Gecko: 4e4e0f502969
Platform Version: 32.0a1
Firmware Version: V123
User Agent: Mozilla/5.0 (Mobile; rv:32.0) Gecko/32.0 Firefox/32.0

Comment 12

[Aus Lacroix [:aus]]

Attachment: Pull Request - Ensure elements in SIM PIN and PUK dialog are translated.

This patch changes how we use the l10n api and also does an initial 'translate' of the elements. We're missing the l10n mutation observer in 2.0 and that dialog is injected dynamically, so it doesn't get translated when injected.

Comment 13

[Aus Lacroix [:aus]]

I also double checked on 2.1 and, as expected, it all works great!

Comment 14

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

Comment on attachment 8476294 [details] [review]
Pull Request - Ensure elements in SIM PIN and PUK dialog are translated.

That looks good! In 2.1 we can replace _l10n with setAttributes :) you have one unit test failing on tbpl, but except of that it's ready!

Comment 15

[Aus Lacroix [:aus]]

> That looks good! In 2.1 we can replace _l10n with setAttributes :) you have
> one unit test failing on tbpl, but except of that it's ready!

Indeed! And the code in 2.1 already uses setAttributes and relies on the mutation observer to do the rest. I'll look at that test and make sure we're all green before committing. Thanks! :)

Comment 16

[Aus Lacroix [:aus]]

Tests updated, green run locally, just waiting on tbpl before committing.

Comment 17

[Aus Lacroix [:aus]]

Test run is good (some orange factor, already known bugs).

Commit (v2.0 only): https://github.com/mozilla-b2g/gaia/commit/2e0da996317fd6bf3bb4a567ff72ce1a0428bf24

Fixed.