[regression] OSK appears blank and/or flickers in the lock screen

[Nekhelesh Ramananthan]

After using OTA-14 a while, the OSK starts to exhibit a strange behaviour.

1. It appears blank after a while in the lock screen. You can still press randomly which then makes the OSK to appear fully. When you press a character again it starts to flicker.

2. OSK starts to flicker similar to bug #1590765 which was fixed in OTA-12. It seems to have resurfaced in OTA-14.

Systems Info:
OTA-14 Stable Channel
Devices: E4.5, E5

[Michal Predotka (mpredotka)]

I had this issue as well. bq e4.5 OTA-14

[Ante Smolcic (supersmola)]

I have this issue after OTA-14 on my BQ Aquarious E4.5.
It also flickers in messaging app. Sometimes it happens after the phone has not been used for a while.
I tried opening several apps: browser, settings, phone app, opening the keyboard in messaging, hiding it, and showing it, locking the phone etc... I couldn't reproduce it. However, after a while (2-3 hours) and a phone call I've got this flickering behaviour. Restart is needed.

There is a similar, if not identical, bug #1590765 but on the ubuntu-keyboard package. There is an exact video in comment #9.

[Anupam (anupam207)]

Same here, E5, OTA-14. Not only the lock screen OSK, keyboards everywhere gets blank and/or keeps on flickering.

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Daniel van Vugt (vanvugt)]

Did you try using OTA-13 for any period of time?

I ask this because it appears OTA-14 does not contain any changes to Mir. Only OTA-13 did...
http://people.canonical.com/~lzemczak/landing-team/ota/

So if OTA-13 was fine then this probably can't be related to Mir. The Mir binaries are unchanged(?)

[Daniel van Vugt (vanvugt)]

Verified that OTA-14 still contains the same Mir release as OTA-13 did. So Mir has not changed:
   https://launchpad.net/mir/0.24/0.24.0
   http://people.canonical.com/~lzemczak/landing-team/ota/ota-13.commitlog
   http://people.canonical.com/~lzemczak/landing-team/ota/ota-14.commitlog

Not a Mir bug unless you can reproduce the same problem in OTA-13.

[Nekhelesh Ramananthan (nik90)]

I do use BQ E4.5 as my daily phone and can confirm that this problem was not present in OTA-13. And since Mir hasn't changed between OTA-13 and OTA-14, it could be unity8 or ubuntu-keyboard causing the issue.

[Daniel van Vugt (vanvugt)]

Thanks for clarifying that. So the list of potential causes then isn't very big:

http://people.canonical.com/~lzemczak/landing-team/ota/ota-14.commitlog

[Michael Sheldon (michael-sheldon)]

There weren't any keyboard landings in OTA 14 either. While there weren't any Mir landings, there was a qtmir landing in OTA 14, so I've added that to the bug as a possible place to investigate as well.

[Ante Smolcic (supersmola)]

I have a suspicion this bug may manifest after a long period of inactivity related to the keyboard. I think I had not been writing any messages for more than 24 hours when I have had noticed it again.
HTH

[Daniel van Vugt (vanvugt)]

Actually this might be Mir's fault still. The first time the bug was noticed was when we introduced NBS and the fix back then was to back out NBS I think. But now I'm actually looking at the NBS code and it seems like it might be dropping the newest frame from the client without any guarantee the client (OSK) will ever produce a new one. Not totally sure, but this might be Mir's fault.

[Daniel van Vugt (vanvugt)]

My mistake. I now can't see how Mir can cause this.

[Launchpad Janitor (janitor)]

This bug was fixed in the package qtmir - 0.5.1+17.04.20170124-0ubuntu1

---------------
qtmir (0.5.1+17.04.20170124-0ubuntu1) zesty; urgency=medium

  [ Gerry Boland ]
  * Recalculate MirSurface.visible after first frame swap by client (LP:
    #1649077)

  [ Lukáš Tinkl ]
  * Some assorted cleanups and small optimizations
  * Apply default device keymap (LP: #1626435)

 -- Michał Sawicz <email address hidden> Tue, 24 Jan 2017 07:38:05 +0000