1013739 - [Dolphin][Flame][FFOS 1.4]Long touch the build-in keyboard and income a call or alarm,the key touching is invalid.

Status: RESOLVED FIXED

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

Description

[孟宪茂]

steps to reproduce:

1.Open SMS and edit messages or add contacts;

2.Long press any of the characters on the virtual keyboard (such as the English state character u, o);

3.Incoming a call or alarm;

4.Hang up the phone, return SMS interface.

Test result:the keys of virtual keyboard become touching  invalid.After restarting devices,the virtual keyboard becomes normal.

Also see the issue in FFOS 1.3.

Comment 1

[bhavana bajaj [:bajaj]]

Since this is reproducible on 1,3 don't think we would block on this

Comment 2

[Peter Dolanjski [:pdol]]

(In reply to bhavana bajaj [:bajaj] from comment #1)
> Since this is reproducible on 1,3 don't think we would block on this

Sounds like we should still fix it at some point, though, so backlog may make sense.

Comment 3

[Ivan Tsay (:ITsay)]

Hi Xianmao,

When this happens, is it the only way to recover from this issue by rebooting the phone? How about kill the app from cardview?

Comment 4

[孟宪茂]

(In reply to Ivan Tsay (:ITsay) from comment #3)
> Hi Xianmao,
> 
> When this happens, is it the only way to recover from this issue by
> rebooting the phone? How about kill the app from cardview?
Dear  Ivan Tsay,
It is the only way to recover from this issue by rebooting the phone;

Kill the app and restart it,the key touching is still invalid;

Besides,other apps can not use the input method ,either.

Thanks.

Comment 5

[Jason Smith [:jsmith]]

Per the above comments, moving to backlog.

Comment 6

[Keven Kuo [:kkuo]]

Hi! Viral,

Looks like we have similar case before. Please comment.

--
Keven

Comment 7

[viral [:viralwang]]

we met the issue before but I remember it's fixed already.
https://bugzilla.mozilla.org/show_bug.cgi?id=946164#c50
Shall this issue happened with APZC or not?

(In reply to 孟宪茂 from comment #4)
> (In reply to Ivan Tsay (:ITsay) from comment #3)
> > Hi Xianmao,
> > 
> > When this happens, is it the only way to recover from this issue by
> > rebooting the phone? How about kill the app from cardview?
> Dear  Ivan Tsay,
> It is the only way to recover from this issue by rebooting the phone;
> 
> Kill the app and restart it,the key touching is still invalid;
> 
> Besides,other apps can not use the input method ,either.
> 
> Thanks.

I'm thinking it will be recovery by killing keyboard, not apps.

Comment 8

[Ivan Tsay (:ITsay)]

Hi Pei-Pei,

Please help to check if this one can be reproducible on Flame. Thanks.

Comment 9

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

(In reply to Ivan Tsay (:ITsay) from comment #8)
> Hi Pei-Pei,
> 
> Please help to check if this one can be reproducible on Flame. Thanks.

This is reproducible on Flame. Test build:

Gaia      5527784b263755f55b620441fc95a66496fa18cf                         
Gecko    https://hg.mozilla.org/releases/mozilla-b2g30_v1_4/rev/334e9712f079        
BuildID   20140528160208                                                   
Version   30.0                                                             
ro.build.version.incremental=94                                            
ro.build.date=Tue May 20 09:29:20 CST 2014

Comment 10

[James Zhang (Spreadtrum)]

(In reply to pcheng from comment #9)
> (In reply to Ivan Tsay (:ITsay) from comment #8)
> > Hi Pei-Pei,
> > 
> > Please help to check if this one can be reproducible on Flame. Thanks.
> 
> This is reproducible on Flame. Test build:
> 
> Gaia      5527784b263755f55b620441fc95a66496fa18cf                         
> Gecko    https://hg.mozilla.org/releases/mozilla-b2g30_v1_4/rev/334e9712f079
> 
> BuildID   20140528160208                                                   
> Version   30.0                                                             
> ro.build.version.incremental=94                                            
> ro.build.date=Tue May 20 09:29:20 CST 2014

100% reproduce rate.

Comment 11

[Jason Smith [:jsmith]]

Product in comment 2 already concluded this wasn't a blocker.

Comment 12

[Ivan Tsay (:ITsay)]

Renominate it because the only way to recover it is to reboot the phone. I think it is the bad user experience when this issue is hit.

Comment 13

[viral [:viralwang]]

I can not reproduce it in my side, not sure if I make any of wrong reproduce step.
Here's my video: https://copy.com/VCQZ2sV9UJpU

Gaia      5527784b263755f55b620441fc95a66496fa18cf
Gecko     https://hg.mozilla.org/releases/mozilla-b2g30_v1_4/rev/334e9712f079
BuildID   20140528160208
Version   30.0
ro.build.version.incremental=94
ro.build.date=Tue May 20 09:29:20 CST 2014

Comment 14

[Preeti Raghunath(:Preeti)]

Tim

Please weigh in for technical risk. We'd like to block on this but will negate the blocking if risk is high.

Comment 15

[Tim Guan-tin Chien [:timdream] (please needinfo)]

I cannot reproduce this on master/2.1. qawanted to verify 2.0/2.1.

It's likely this is accidentally fixed between 1.4 and 2.0, we can make the decision about this bug when we find the patch that fix this. (should I tag "regressionwindow-wanted" in this case?)

Comment 16

[Rudy Lu [:rudyl] (inactive after 2015/7/3)]

I cannot reproduce this issue either on Flame with the following 2.0 build,
Gaia      2e5636e852a9354a5f8072b179cf16f72647cfd6
Gecko     https://hg.mozilla.org/mozilla-central/rev/8bd92dc9ef59
BuildID   20140608160201
Version   32.0a1
ro.build.version.incremental=eng.cltbld.20140603.191906
ro.build.date=Tue Jun  3 19:19:16 EDT 2014

--
I'll try to investigate this on v1.4.

Comment 17

[Lionel Mauritson]

(In reply to Tim Guan-tin Chien [:timdream] (MoCo-TPE) (please ni?) from comment #15)
> I cannot reproduce this on master/2.1. qawanted to verify 2.0/2.1.

Unable to reproduce on 2.0/2.1.

Device: Flame 2.1 - Master
Build ID: 20140610040208
Gaia: f42ebc93554979501d3ac52bcf9e69cb4b310a4f
Gecko: 9dc0ffca10f4
Version: 33.0a1 (2.1 - Master) 
Firmware Version: v10G-2
User Agent: Mozilla/5.0 (Mobile; rv:33.0) Gecko/33.0 Firefox/33.0

Device: Flame 2.0
Build ID: 20140610000223
Gaia: 8d865839d932bfbd5e157f376f74d8cb12bfdd51
Gecko: 1d4046a8cb6c
Version: 32.0a2 (2.0) 
Firmware Version: v10G-2
User Agent: Mozilla/5.0 (Mobile; rv:32.0) Gecko/32.0 Firefox/32.0

Comment 18

[Joshua Mitchell (Inactive)]

QA-Wanted to check Flame 1.4 - and if it does repro there then please check Flame 1.3

Comment 19

[Rudy Lu [:rudyl] (inactive after 2015/7/3)]

2 findings when investigating,

  1. This issue would happen because we could not receive the touchend event to reset the state that alternative menu is showed, and in this state, we would ignore all new touch events.

  2. Cannot reproduce this issue if I simply turn on the "keyboard oop" option on v1.4.

--
Will try to workaround this issue by reset the state when the keyboard is hidden.

Comment 20

[Rudy Lu [:rudyl] (inactive after 2015/7/3)]

Attachment: Patch for v1.4

This is a v1.4-specific patch to workaround this issue, since this could not reproduced on master.

Tim could you please review this?
Thank you.
(I didn't add unit tests here since we don't have that for keyboard.js.)

Comment 21

[Rudy Lu [:rudyl] (inactive after 2015/7/3)]

Set the status for 2.0/2.1 according to Comment 17 and our own test result.

Comment 22

[Lionel Mauritson]

(In reply to Joshua Mitchell from comment #18)
> QA-Wanted to check Flame 1.4 - and if it does repro there then please check
> Flame 1.3

DOES occur on both Flame 1.4 and the v10G-2 1.3 base image.

Device: Flame 1.4
Build ID: 20140612000202
Gaia: 7fc73d4cb1bece31f50e8ccf6fb98af3984a9ebf
Gecko: bcd308fbbf38
Version: 30.0 (1.4) 
Firmware Version: v10G-2

Device: Flame 1.3
Build ID: 20140520094859
Gaia: a73235d23685e9898f40647cebd83b3fcbfd0117
Gecko: b637b0677e15318dcce703f0358b397e09b018af
Version: 28.0 (1.3) 
Firmware Version: v10G-2

Comment 23

[Tim Guan-tin Chien [:timdream] (please needinfo)]

Comment on attachment 8439164 [details] [review]
Patch for v1.4

Let's land this to master/2.0 too so keyboard could still work even if we run it inproc on these branches?

Comment 24

[Rudy Lu [:rudyl] (inactive after 2015/7/3)]

Attachment: Patch for master

Patch for master, and carry forward the r+.

Comment 25

[Rudy Lu [:rudyl] (inactive after 2015/7/3)]

Landed to master first,
https://github.com/mozilla-b2g/gaia/commit/f95fa5082acf858835ed5e77e133869cf44cf259

--
Modify the state of 2.0/2.1 to affected, since this should occur when keyboard is run in-process.

Comment 26

[Jan Jongboom [:janjongboom] (inactive)]

Why are we landing this in 2.x and not in 1.4 yet?

Comment 27

[Rudy Lu [:rudyl] (inactive after 2015/7/3)]

For patches that are not branch-specific, we would land them to master first, and release engineer team would help to uplift these patches if the patch stays on master without causing regression.

Comment 28

[Ryan VanderMeulen [:RyanVM]]

v2.0: https://github.com/mozilla-b2g/gaia/commit/cc354660b5b9b597adfe87eaaadd579c5ef75f39
v1.4: https://github.com/mozilla-b2g/gaia/commit/164644d91290708a71436dfdf4301e33b92e2c77

Comment 29

[[:SalvadorR]]

This needs to be covered by a new test case.

Comment 30

[KTucker [:KTucker][Inactive 3/4/2016]]

Test case has been added in moztrap:

https://moztrap.mozilla.org/manage/case/14322/

Comment 31

[Coler]

This issue has been verified successfully on lame2.1&2.0.
Reproducing rate: 0/5
See attachment: Verify_Flame_Keyboard.mp4

Flame2.0 build version:
Gaia-Rev        8d1e868864c8a8f1e037685f0656d1da70d08c06
Gecko-Rev       https://hg.mozilla.org/releases/mozilla-b2g32_v2_0/rev/c756bd8bf3c3
Build-ID        20141130000204
Version         32.0

Flame2.1 build version:
Gaia-Rev        ccb49abe412c978a4045f0c75abff534372716c4
Gecko-Rev       https://hg.mozilla.org/releases/mozilla-b2g34_v2_1/rev/18fb67530b22
Build-ID        20141130001203
Version         34.0

Comment 32

[Coler]

Attachment: Verify_Flame_Keyboard.MP4