824136 - No backward button causes STK GCF test case to fail

Status: RESOLVED FIXED

(Firefox OS Graveyard :: Gaia::Settings, defect, P1)

Description

[Carol Yang [:cyang]]

In GCF TC 27.22.4.1.1.7, the test is ran as follows:

- Start GCF test which prompts for device power up
- After power up, hitting Continue from GCF test will send a proactive command for Display Text of "<GO-BACKWARDS>"

At this point, in order for the test to pass, the user must press the back button on the device. However, pressing the back button does nothing.

Given that the back button is not supported as all navigation comes from the UI, this GCF test case will never pass. Are there any plans to support the back button?

Comment 1

[Lawrence Mandel [:lmandel] (use needinfo)]

From the description, the test case reads as invalid. There is no back button on the target device. The target device will have a single button. I'm resolving this bug as invalid due to the fact that the device has no back button.

Comment 2

[Michael Vines [:m1] [:evilmachines]]

The idea here would be to display a "back" button on the display, like the 2nd level of the settings app for example.

Comment 3

[Lawrence Mandel [:lmandel] (use needinfo)]

Is the suggestion to add a soft key back button to specific apps or a generic soft key back button that is always available on the phone?

Comment 4

[Michael Vines [:m1] [:evilmachines]]

I think we'd just be looking at a UX change to the STK app when the display text proactive command is presented.   If partners are ok with this GCF test case failing then that's a fine resolution as well.

Comment 5

[Lawrence Mandel [:lmandel] (use needinfo)]

Thanks for the clarification Michael. 

Reassigning to Gaia based on comment 4.

Comment 6

[Joe Cheng [:jcheng] (please needinfo)]

Daniel, is this a cert concern for you with this test case failing? due to not having a back key in STK?

Comment 7

[Joe Cheng [:jcheng] (please needinfo)]

Triage: BB+, P1, C3 - certification blocker

Comment 8

[David Palomino [:dpalomino]]

Adding dependency: v1 certification blocker

Comment 9

[Etienne Segonzac (:etienne)]

Is there any way for a mere mortal to reproduce the bug?

Comment 10

[Daniel Coloma:dcoloma]

frsela, can you have a look at this?

Comment 11

[Fernando R. Sela (no CC, needinfo please) [:frsela]]

(In reply to Daniel Coloma:dcoloma from comment #10)
> frsela, can you have a look at this?

Sure.

Comment 12

[Fernando R. Sela (no CC, needinfo please) [:frsela]]

I need more info since a back button is showed on the top-left corner of the screen and it's working well.

Comment 13

[Yoshi Cheng-Hao Huang [:allstars.chh][:allstarschh][:yoshi]]

Backward button is on the top left and is working well

Comment 14

[Fernando R. Sela (no CC, needinfo please) [:frsela]]

The back button is available and is working fine.
Closing the bug as INVALID.

Comment 15

[Carol Yang [:cyang]]

Pressing the top-left corner arrow does take me to the Settings main menu. However, it does not issue the proper terminal response, causing this GCF TC to fail.

The terminal response "Backward move in the proactive SIM session requested by the user" should be sent.

Comment 16

[Fernando R. Sela (no CC, needinfo please) [:frsela]]

(In reply to Carol Yang from comment #15)
> Pressing the top-left corner arrow does take me to the Settings main menu.
> However, it does not issue the proper terminal response, causing this GCF TC
> to fail.
> 
> The terminal response "Backward move in the proactive SIM session requested
> by the user" should be sent.

Thanks for the clarification Carol.

This button has different behavior depending in which screen are you.

1.- If you are in the main menu (list of STK apps or STK menu) this button exits the STK app and returns to the settings list

2.- If you are inside any app (after user sendStkMenuSelection), it sends the proactive command "STK_RESULT_BACKWARD_MOVE_BY_USER"

For your explanation, the first behavior should send another proactive command before close STK app.

Yoshi, STK_RESULT_OK is enough if no menu selection was used? or use again STK_RESULT_BACKWARD_MOVE_BY_USER for the first case

Comment 17

[Yoshi Cheng-Hao Huang [:allstars.chh][:allstarschh][:yoshi]]

I don't have GCF at hand, might need to check again tomorrow.

But as I read again in comment 1,
maybe Carol is talking about 
"When the DISPLAY TEXT dialog shows, user could press a back key"

not when user is browsing the menu, user could press a back key.

Comment 18

[Carol Yang [:cyang]]

Yoshi, some more info that may be of help:

In this particular TC, the user does not navigate to the SIM Toolkit menu. The DISPLAY_TEXT will popup once the test issues the proactive command.

TS 51.010 says this for step 5:

USER -> ME	Indicate the need to go backwards in the proactive SIM application session

From an Android device, pressing either the soft key back button or the physical back button on the device would issue the terminal response.

Comment 19

[Fernando R. Sela (no CC, needinfo please) [:frsela]]

Attachment: Updates the STK header exit/back buttons with any proactive command (PR Link)

Ouch. I see... when a any command is sent before STK_CMD_SELECT_ITEM, the header is not updated, so the default EXIT button is used. This patch fix it.

Carol, I don't have any SIM card here which sends a proactive command on boot up, can you test with this patch? Thanks

Comment 20

[Anshul]

Fernando, Siddartha knows a way to test all the STK changes on a simulator by simply sending a known payload. Would you be able to test on your end and Carol can validate it later next week?

Comment 21

[Fernando R. Sela (no CC, needinfo please) [:frsela]]

(In reply to Anshul from comment #20)
> Fernando, Siddartha knows a way to test all the STK changes on a simulator
> by simply sending a known payload. Would you be able to test on your end and
> Carol can validate it later next week?

Oh this is awesome and great.

I normally test it with a modified version of gaia to "simulate" the STK commands. It's usefull to test UI but fails navigation since the SIM card receives strange responses :)

Can you explain me how to simulate it? Thanks in advance.

Comment 22

[Anshul]

[e-mail] can point you the instructions.

Comment 23

[Vivien Nicolas (:vingtetun) (:21) - (NOT reading bugmails, needinfo? please)]

Comment on attachment 700218 [details]
Updates the STK header exit/back buttons with any proactive command (PR Link)

Just a small change to do and this is fine.

Comment 24

[Yoshi Cheng-Hao Huang [:allstars.chh][:allstarschh][:yoshi]]

CCing Siddartha.

Siddartha, can you check the GCF as Carol mentioned in comment 1?
By Carol's comments, maybe Frsela and I got some misunderstanding about the requirements.

Comment 25

[Fernando R. Sela (no CC, needinfo please) [:frsela]]

Comment on attachment 700218 [details]
Updates the STK header exit/back buttons with any proactive command (PR Link)

NIT changed

Comment 26

[Yoshi Cheng-Hao Huang [:allstars.chh][:allstarschh][:yoshi]]

Attachment: Press Back Key

Hi, Fernando
As Siddartha and I think the bug is when DISPLAY_TEXT, 
there is no where to press the 'back' button.

Maybe we need to opinion from UX first.

Comment 27

[Fernando R. Sela (no CC, needinfo please) [:frsela]]

(In reply to Yoshi Huang[:yoshi][:allstars.chh] from comment #26)
> Created attachment 700248 [details]
> Press Back Key
> 
> Hi, Fernando
> As Siddartha and I think the bug is when DISPLAY_TEXT, 
> there is no where to press the 'back' button.
> 
> Maybe we need to opinion from UX first.

Yeap, that's I misunderstand anyway the proposed patch will fix the back button if a proactive command opens another kind of screens (like INPUT one)

I'll ask UX in order to define this, but I suggest a second button like in a Ok/Cancel dialog but with Ok/Back

Comment 28

[Fernando R. Sela (no CC, needinfo please) [:frsela]]

Attachment: Adding a back key

I've another WIP patch (https://github.com/frsela/gaia/tree/STK/Bug824136_b) which adds the back button.

So now the Back button send STK_RESULT_BACKWARD_MOVE_BY_USER and the Ok one  STK_CMD_DISPLAY_TEXT.

Anyway we need to define how it will work on the cases of "responseNeeded" which will be responded BEFORE show the DISPLAY box, in that case, the STK_RESULT_BACKWARD_MOVE_BY_USER won't be sent since the command was previously processed.

Comment 29

[Fernando R. Sela (no CC, needinfo please) [:frsela]]

Attachment: Adding a back key on DISPLAY boxes

Patch to add the back key showed in https://bugzilla.mozilla.org/attachment.cgi?id=700262&action=edit

UX: You agree with this proposal?

Comment 30

[Rafael Rebolleda]

With so little information, it's pretty difficult –if possible at all– to know where the back or the ok button should take the user, thus this solution might completely leave the user somewhere unexpected. For the record, this is a bad solution.

In any case, OK should be the rightmost button.

Comment 31

[Vivien Nicolas (:vingtetun) (:21) - (NOT reading bugmails, needinfo? please)]

Comment on attachment 700218 [details]
Updates the STK header exit/back buttons with any proactive command (PR Link)

r+ on the code but I have no idea if this is going to fix the testcase.

Comment 32

[Vivien Nicolas (:vingtetun) (:21) - (NOT reading bugmails, needinfo? please)]

Comment on attachment 700291 [details]
Adding a back key on DISPLAY boxes

Same comment as the above.

Comment 33

[Daniel Coloma:dcoloma]

Carol, can you test if the suggested patches are enough to meet the GCF test case?

Comment 34

[Carol Yang [:cyang]]

(In reply to Daniel Coloma:dcoloma from comment #33)
> Carol, can you test if the suggested patches are enough to meet the GCF test
> case?

Daniel, we've tested this locally and so far so good. I'll need to get into the lab tomorrow morning to report final results on whether GCF test passes or not though. Will update asap. Just to make sure we're on the same page, only attachment 700218 [details] and 700291 are needed right?

Comment 35

[Fernando R. Sela (no CC, needinfo please) [:frsela]]

Hi,

Regarding both patches.

700218 fix the back button on all STK screens when they are call by a direct async proactive command (I mean, the user didn't ask for it but the SIM card sent it -> Open a STK app, Display a message, ask for INPUT data, ...)

Without this patch, the behavior of the back button is not updated until you surf through the STK menus so can be set to exit or back (depending on the last update of it).

----

Regarding the other patch is ONLY for the DISPLAY_MSG command which is showed in a modal window and isn't possible to push the STK back button, so I added the back button to the modal dialog.

Anyway, talking with Rafael Rebolleda, we should try to mix both buttons in one, I mean, if the dialog is opened after a user ask for it, act as Ok one but for async msg, act as back one. But I'm not really convinced since the meaning for the STK app are different responses ...

The problem here is that the user could be confused. Which button to press? Ok? Back? ... but this is the STK standard ... no to much to do here :s

Comment 36

[Fernando R. Sela (no CC, needinfo please) [:frsela]]

(In reply to Rafael Rebolleda [:rafaelrebolleda] from comment #30)

> In any case, OK should be the rightmost button.

Changed: https://github.com/frsela/gaia/commit/4f0771a0c500ce78e636f5774da3295a075d9a99

Comment 37

[Fernando R. Sela (no CC, needinfo please) [:frsela]]

(In reply to Carol Yang from comment #34)
> only attachment 700218 [details] and 700291 are needed right?

Yes, both of them since each one fixes different use cases.

Thank you.

Comment 38

[Fernando R. Sela (no CC, needinfo please) [:frsela]]

Both patches with r+ merged:

https://github.com/mozilla-b2g/gaia/commit/61e7b7cf7910c63dfe7bf0cc24e664c94d8de9fd
https://github.com/mozilla-b2g/gaia/commit/c69d46215195d53c20b777512357982b0b23ac3f

Comment 39

[Carol Yang [:cyang]]

Hi Fernando, I just pulled in the patches you mentioned in Comment 38. This allows the GCF TC 27.22.4.1.1.7 (which this bug was opened on) to pass. But just thought you'd like to know that there does seem like a problem in the following scenario:

In the case where a backwards move is expected for a Get Inkey (i.e. 27.22.4.2.1.3), there is no Back button and the option to move back is via the top-left corner arrow. When pressing the arrow, a backward move terminal response does get sent, however, on the UI, nothing happens. I'm not sure if this is intentional or not since you wouldn't know where to go backwards to from a proactive command. Also not sure if you want to track this in a diff bug but thought I'd bring this to your attention.

Comment 40

[Fernando R. Sela (no CC, needinfo please) [:frsela]]

(In reply to Carol Yang from comment #39)
> Hi Fernando, I just pulled in the patches you mentioned in Comment 38. This
> allows the GCF TC 27.22.4.1.1.7 (which this bug was opened on) to pass. But
> just thought you'd like to know that there does seem like a problem in the
> following scenario:
> 

\o/ nice to know !

> In the case where a backwards move is expected for a Get Inkey (i.e.
> 27.22.4.2.1.3), there is no Back button and the option to move back is via
> the top-left corner arrow. When pressing the arrow, a backward move terminal
> response does get sent, however, on the UI, nothing happens. I'm not sure if
> this is intentional or not since you wouldn't know where to go backwards to
> from a proactive command. Also not sure if you want to track this in a diff
> bug but thought I'd bring this to your attention.

Yes, in all STK screens (except popup ones like DISPLAY messages) the back button is the top-left one in order to use the same experience in all the system, for the user this button is intuitive to go back, to the previous screen.

When the arrow is pressed, the response icc.STK_RESUT_BALCKWARD_MOVE_BY_USER is sent to the SIM card. (https://github.com/mozilla-b2g/gaia/blob/master/apps/settings/js/icc.js#L114)

So the next screen to show is responsible of the STK app since from the UI I cann't know which one is.

If it's needed to send a different response let me know or also we can trace is st. bad is happening. Could you sent me a logcat with debug enabled?

change the following lines to enable it:

-> https://github.com/mozilla-b2g/gaia/blob/master/apps/settings/js/utils.js#L9
-> https://github.com/mozilla-b2g/gaia/blob/master/apps/system/js/icc_cache.js#L10

Check it and if you consider this is a bug, we shall track it in a new one.

Comment 41

[Carol Yang [:cyang]]

(In reply to Fernando R. Sela [:frsela] from comment #40)
> (In reply to Carol Yang from comment #39)

> So the next screen to show is responsible of the STK app since from the UI I
> cann't know which one is.
> 

It doesn't seem like the STK app always knows where to go back to, at least not for the case of 27.22.4.2.1.3. Android behavior takes user back to the homescreen. Would this be something safe to do?

> If it's needed to send a different response let me know or also we can trace
> is st. bad is happening. Could you sent me a logcat with debug enabled?

STK_RESULT_BACKWARD_MOVE_BY_USER is the right response to send so no need to change anything here :)

Comment 42

[Carol Yang [:cyang]]

From section 6.11 of the 3GPP TS 11.14:

Backward move is a possible value returned with terminal response for the following:

- DISPLAY_TEXT
- GET_INKEY
- GET_INPUT
- SELECT_ITEM

Curious if it makes sense that for SELECT_ITEM, it goes back to the SIM Toolkit menu and for the other cases, sense there was no menu it came from, to return to the homescreen?