LineageOS 18.1 Official Release for Pro1

[marmistrz 93]

22 minutes ago, Hook said:

Again, as far as I can tell, this is controlled by the  the soft keyboard.  AOSP keyboard has sticky shift (but no setting for it) but it does have an autocapitilization toggle.  If you turn it off in the soft keyboard, the hardware keyboard no longer does it.

I switched to the AOSP keyboard, turned off the autocapitalization. the software keyboard doesn't autocapitalize, the hardware one does (tested by writing an e-mail in k9 mail)

[Rob. S. 1,660]

32 minutes ago, marmistrz said:

I switched to the AOSP keyboard, turned off the autocapitalization. the software keyboard doesn't autocapitalize, the hardware one does (tested by writing an e-mail in k9 mail)

Strange. To the best of my knowledge, there still must be some piece of software doing it. I've never seen a hardware keyboard doing autocapitalization...

[marmistrz 93]

17 minutes ago, Rob. S. said:

Strange. To the best of my knowledge, there still must be some piece of software doing it. I've never seen a hardware keyboard doing autocapitalization...

it's definitely done in software as it depends on the app used

Edited August 28, 2021 by marmistrz

[Hook 3,017]

1 hour ago, marmistrz said:

I switched to the AOSP keyboard, turned off the autocapitalization. the software keyboard doesn't autocapitalize, the hardware one does (tested by writing an e-mail in k9 mail)

This is just weird.  I tested it just before I posted before I a swear it was working.  But closer examination seems to show a highly varied response of different apps in terms of a) whether they autocapitalize,  b) whether they follow the soft keyboarsd setting and c) whether they offer a control of their own for Autocapitalization. This is all in terms of what the hardware keyboard does in the case of having the AOSP soft keyboard with the Autocapitalization setting.  Aquamail, based on K-9 , ignores the setting and always Autocapitalizes, as does Google messages.  The Browser Fulguris ignores the setting and never Autocapitalizes.  The text Editor Markor obeys the AOSP keyboard setting.  Textmaker (part of Softmaker office), ignores the setting but offers it's own Autocapitalization toggle.

My guess would be, although this is way beyond my noob knowledge, that Android doesn't give you guarantees that an app won't much with the keyboard settings, even if mucking in this case is just ignoring them.

 

 

[Rob. S. 1,660]

12 hours ago, marmistrz said:

the software keyboard doesn't autocapitalize, the hardware one does (tested by writing an e-mail in k9 mail)

10 hours ago, Hook said:

Aquamail, based on K-9 , ignores the setting and always Autocapitalizes, as does Google messages

Interesting. So it seems like there are application developers which bothered enough to implement at least a check for a physical keyboard, but didn't bother enough to really think about exactly how a physical keyboard should behave differently...

In theory, I would see two approaches to fix this – filing bug reports for every app behaving like this, or finding a way to (optionally, per-app) trick Android/LineageOS into not reporting the physical keyboard to apps, if there would be such a way...

(FWIW, I recently moved from K9 to FairEmail, which looks a bit like K9 in its brand new makeover, after ruling out Aquamail for some reason. I think it was because it didn't do PGP message encryption last time I looked... Seems like FairEmail doesn't autocapitalize.)

Edited August 29, 2021 by Rob. S.

[claude0001 1,340]

12 hours ago, marmistrz said:

I switched to the AOSP keyboard, turned off the autocapitalization. the software keyboard doesn't autocapitalize, the hardware one does (tested by writing an e-mail in k9 mail)

While my hardware (qwertz) in principle works well with AOSP keyboard (can disable autocapitalization in the settings, sticky shift works), I also observe some quirks in K9-Mail only. So maybe test also in other apps.

Most annoying for me is that whenever I type "im" (german for "in"), that gets auto-replaced by "I'm", although AOSP keybaord is set to german input language and all auto-replacements are disabled in the settings. K9 is my only app that does that. Still hoping that misbehaviour will disappear with some update of the app as it seems to be the only (open-source) one to support arbitray IMAP folder configurations ... 😐

Edited August 29, 2021 by claude0001

[EskeRahn 5,459]

(lineage-18.1-20210830-nightly-pro1-signed.zip on August 5 security patch installed smoothly using adb sideload and Open_Gapps-arm64-11-pico-20210712)

...But the reintroduced Accessibility / keyboard bug is not (yet) fixed, see above.

[Slion 1,201]

On 8/23/2021 at 6:29 PM, Slion said:

Anyway, it looks like that task switcher bug is not an issue anymore since there is an easy workaround for it.

However there is still that horizontal empty bar issue:
https://gitlab.com/LineageOS/issues/android/-/issues/3874

I thought that was caused by the nav bar however after switching from aosp keyboard to gboard it looks different, it's black instead of the current background color, therefore I'm thinking it could be caused by the current virtual keyboard while still being a system issue somehow.

I'm also surprised such a bug has not been spotted or tracked down earlier as it is such an obvious issue, so maybe there is also an easy workaround for it too.

As posted on GitLab there is a Magisk extension call "Fullscreen/Immersive Gestures (Q-S)" to workaround this issue. It's designed to hide the "hide keyboard" tool bar displayed below your onscreen keyboard. Use the volume +/- buttons to set it up. I used the following options:

• Vol+ Fullscreen (not sure though what's the difference with immersive)
• Vol+ Low sensitivity gesture (apparently also fixing the Task Switcher gesture being overly sensitive in landscape)
• Vol- Don't disable back gesture

It's also worth noting that this is so annoying that manufacturers like Samsung and OnePlus are adding settings options to disable it. Samsung calls it "Show button to hide keyboard", OnePlus calls it "Keyboard height adjustment".

See:

https://android.gadgethacks.com/how-to/remove-black-bar-under-your-galaxys-keyboard-0384624/

https://forums.oneplus.com/threads/removing-the-hide-keyboard-button.1289973/page-2#post-22839127

Edited September 2, 2021 by Slion

[Wheeljack 212]

The Aug/30 build introduced a weird sticky touchscreen issue for me.

It's very apparent in Mario Kart Tour. When I go for a power slide and then let go of the screen to go straight again, the driver keeps sliding as if I kept my finger on the screen.

In other situations tapping the screen seems delayed (maybe the release is registered too late).

I've switched to the other system slot to revert to build Aug/23 and the issue is gone.

Can anyone confirm?

[Hook 3,017]

3 hours ago, Wheeljack said:

The Aug/30 build introduced a weird sticky touchscreen issue for me.

It's very apparent in Mario Kart Tour. When I go for a power slide and then let go of the screen to go straight again, the driver keeps sliding as if I kept my finger on the screen.

In other situations tapping the screen seems delayed (maybe the release is registered too late).

I've switched to the other system slot to revert to build Aug/23 and the issue is gone.

Can anyone confirm?

I don't play Mario Kart, but I'm not seeing any sort of delay on my touchscreen.  Responsive and snappy as always. 

[aliceshard 1]

On 9/2/2021 at 5:57 PM, Wheeljack said:

The Aug/30 build introduced a weird sticky touchscreen issue for me.

It's very apparent in Mario Kart Tour. When I go for a power slide and then let go of the screen to go straight again, the driver keeps sliding as if I kept my finger on the screen.

In other situations tapping the screen seems delayed (maybe the release is registered too late).

I've switched to the other system slot to revert to build Aug/23 and the issue is gone.

Can anyone confirm?

I have same touchscreen issue. Touch screen does not react as usual, randomly freezes without any reason. (Aug/30 build)

https://www.youtube.com/watch?v=48H3H5kwVIE

https://www.youtube.com/watch?v=GLLCkAqtvvY

Edited September 3, 2021 by aliceshard

[Slion 1,201]

17 hours ago, aliceshard said:

I have same touchscreen issue. Touch screen does not react as usual, randomly freezes without any reason. (Aug/30 build)

https://www.youtube.com/watch?v=48H3H5kwVIE

https://www.youtube.com/watch?v=GLLCkAqtvvY

I think I have a similar issue. Looks like I'm getting lags or freeze from time to time. Definitely a regression in the latest release. Even though I'm not actually using the release but my own build from around the same time. It happened like 5 times while I was typing this message. That's 6 now.

[Slion 1,201]

I reverted back to a custom build from the 27th of August and I don't have that issue at all. Actually I believe my code checkout was front the 25th so at some point between the 25th and the 30st someone introduced that regression. Now go figure what it is exactly.

[Hook 3,017]

18 hours ago, aliceshard said:

I have same touchscreen issue. Touch screen does not react as usual, randomly freezes without any reason. (Aug/30 build)

https://www.youtube.com/watch?v=48H3H5kwVIE

https://www.youtube.com/watch?v=GLLCkAqtvvY

The first one looks like a weird variation of the phantom touch phenomenon.  I can't reproduce either on my Pro1 with a screen that was replaced in January by FxTec (not a Alieexpress Elephone U Pro LCD) using that latest nightly of LOS.  I tested the second video y swiping rapidly up and down on the settings menu.  Different speeds, different amounts of pressure.  Always rolled, never froze, no lag.

[Slion 1,201]

Those are the changes that occurred between the 25th and the 30st of August. There is obviously one Pro1 change that could be the culprit.

repo forall -c 'git log --since "AUG 25 2021" --until "AUG 30 2021" --pretty=format:"===============================%n%Cred$(git config --get remote.github.projectname)%n%Creset%H%n%an%n%ad%n%B%n"'
===============================
LineageOS/android_device_fxtec_pro1
15b9323d740da90ba3c00a86002a6b995c42d211
Luca Stefani
Wed Aug 18 22:07:34 2021 +0200
pro1: Adjust powerhint for QOS interface

Currently all writes to the dev interface are done
by WriteStringToFd that uses char internally.

The current values seem to expect to be written as s32,
this sadly won't happen and triggers a special case in
the write handling of pm_qos_power_write:
it expects char buffers to be encoded as base16 and
decodes them.
This means the current 44 is actually 0x44 -> 68 seen by
the kernel.

Luckily it seems like both accepted values for this node
don't hit the threshold to enter C2, so it was never
noticed in real usage and didn't effect the device
C-States handling during hints.

Signed-off-by: Luca Stefani <luca.stefani.ge1@gmail.com>
Change-Id: Ic544d4dcaa1edc3de913aed737baf1af88a45360

===============================
LineageOS/hudson
f3b5874ebc2c80fb4fae079e1558b35ffbac09c4
Sebastiano Barezzi
Fri Aug 27 02:53:04 2021 +0200
hudson: Welcome to POCO X3 Pro

Change-Id: I036d0dcbedfd3fa6d9aae2143022a180dd8df1af

===============================
LineageOS/mirror
1745da7441e70c5c3be5a322b4ddf0ec8e374d3c
Tim Schumacher
Fri Aug 27 14:02:01 2021 +0200
Updated aosp-minimal to 27-Aug-2021 12:02 UTC

Change-Id: Iede1aefc69adc87ed004479800734acdfa996032

===============================
LineageOS/lineage_wiki
2a96730cbee50d4617b626b89400a043eed633f4
LuK1337
Sat Aug 28 11:55:01 2021 +0200
wiki: Use correct closing tags in _includes/deviceinfo.html

Change-Id: Idc3130345c7488b2e6a0254369c1bafeaa566ce4

LATER:

I actually reduced further the list of suspected changes by looking at the source I used to build my custom build. 

Here is the suspicious change:
https://review.lineageos.org/c/LineageOS/android_device_fxtec_pro1/+/314887

Edited September 5, 2021 by Slion

[Slion 1,201]

I'm investigating that lag issue. As explained above I further reduced the list of potential culprit.

I switched back to my broken build and sure enough not long after boot I could reproduce the issue fairly easily.
I'm currently attempting a build reverting that change to see if that fixes it.

[Slion 1,201]

1 hour ago, Slion said:

I'm investigating that lag issue. As explained above I further reduced the list of potential culprit.

I switched back to my broken build and sure enough not long after boot I could reproduce the issue fairly easily.
I'm currently attempting a build reverting that change to see if that fixes it.

Looks like it's fixed. I'll test it some more but I already advised the guys at Lineage to revert it.

[Wheeljack 212]

Updated to the Sep/06 build today. Can confirm that the screen lag is fixed.

[Slion 1,201]

5 minutes ago, Wheeljack said:

Updated to the Sep/06 build today. Can confirm that the screen lag is fixed.

I did not checked that they actually reverted and somehow I doubt it... 😁

[Slion 1,201]

1 hour ago, Slion said:

I did not checked that they actually reverted and somehow I doubt it... 😁

It looks like it's still in there though somehow it was remerged? No sure what's going on TBH. Maybe they fixed it some other way too...

[Hook 3,017]

All the Pro1 labeled fixes seem to be the keyboard stuff.

 

https://download.lineageos.org/pro1/changes/

[EskeRahn 5,459]

(lineage-18.1-20210906-nightly-pro1-signed.zip on August 5 security patch installed smoothly using adb sideload and Open_Gapps-arm64-11-pico-20210712)

...But the reintroduced Accessibility / keyboard bug is not (yet) fixed, see above.

[claude0001 1,340]

@Slion I understand this was missing in the original patch. Could that explain what you observe?

Edited September 6, 2021 by claude0001

[Slion 1,201]

On 9/6/2021 at 3:57 PM, claude0001 said:

@Slion I understand this was missing in the original patch. Could that explain what you observe?

Unlikely

LATER: Actually you were right @claude0001see comment from @npjohnson below.

 

Edited September 8, 2021 by Slion

[Slion 1,201]

On 9/6/2021 at 1:52 PM, Wheeljack said:

Updated to the Sep/06 build today. Can confirm that the screen lag is fixed.

Seems like they indeed fixed it and somehow without reverting those changes. I wish someone could explain that to us. My latest build is real smooth.