Duplicated keys

[adam.c.r.roman 25]

Anyone else getting repeated keys AND keys out of order?  I can type something like APPLE and it may come out APPELL.  Most certainly software, and not just the debouncing issue I described that can repeat keys.

And I also have experienced severe lagg (that extra "g" was a repeated key for real) where the keyboard input is several seconds behind.  It seems to happen a lot faster when I Copy and Paste web text.  I have to wonder if this is due to a keylog file that we don't know about.  Because, unlike the repeat issue, that problem is always resolved on a reboot (so maybe a new log file is created from scratch and there is no delay for a while????)  Just a thought of what else to explore in endless amounts of flawed code...

[Rob. S. 1,660]

5 hours ago, adam.c.r.roman said:

Anyone else getting repeated keys AND keys out of order?  I can type something like APPLE and it may come out APPELL.  Most certainly software, and not just the debouncing issue I described that can repeat keys.

Indeed, but wasn't that something that's been reported very early and also fixed in the first stock Android update? 

[EskeRahn 5,460]

7 hours ago, adam.c.r.roman said:

....the keyboard input is several seconds behind....

That sounds like something eats up the processing power, so it is unable to process the keystrokes in real time.
But obviously what ever does the processing should have had sufficient priority for this not to happen normally.

Anything you are aware of running in the background that might be what is disturbing. E.g. some complex sound decoding, playing high-resolution sound?
I would try to install something like GSam or the like, to find out what is using the resources. (GSam registers the battery pull over time, but any substantial CPU activity should also leave a mark in the power consumption). It MIGHT be something that it is OK got a higher priority than the keyboard.

To make it do it's best, make sure to give att to see the battery statistics. (for the version paid with money/ads respectively)

adb -d shell pm grant com.gsamlabs.bbm.pro android.permission.BATTERY_STATS
adb -d shell pm grant com.gsamlabs.bbm     android.permission.BATTERY_STATS

[Laska 41]

I think it is temperature related: on a hot, sunny day I get almost every key doubled. And it never happen with software keyboard input which shoud have same processing priority.

[EskeRahn 5,460]

21 minutes ago, Laska said:

I think it is temperature related: on a hot, sunny day I get almost every key doubled. And it never happen with software keyboard input which shoud have same processing priority.

I would be surprised if the handling of the physical keys and touch are not substantially different, so do not think we can deduce from differences between the two.

[adam.c.r.roman 25]

On 8/30/2021 at 1:17 AM, Rud said:

Same issue here. double "e", sometime no "d" or "a" and now and then the keyboard lags (words displaying on screen 3 to 4 words behind what I'm typing).

...

Both issues described above not happening constantly, so this makes me think they are both software issues.

I monitored CPU usage for a bit, but as of now, I couldn't pinpoint an app that bottlenecks the CPU.

I have found that a reboot always fixes the severe lags, where it outputs like 1 character a secondd, but nothing helps those repeated keys like you just saw, because...  I'll say it again...  there's something in the Software where it's not sampling and debouncing the keyboard fast enough, regardless of moderate physical degradation to the keyboard.

The lag issue is confusing, however.  I have to wonder if there's a keylog being created every time the device reboots, and it gets to be so huge, that writing keystrokes to the log file takes priority to them actually being output to the application layer, and we experience this huge lag.  I notice this lag occur more quickly when I copy and paste web content into a text editor.  Perhaps because there is just such a huge amount of text going into a log file, and then typing and adding one more character at a time takes a lot of time to open, modify, and save the file ever time a key is pressed???

And no, I wouldn't expect any bottlenecks - it has more to do with How the data is being processed and passed around, not CPU speed.  After all, it's only a few keystrokes a second (other than the time proper debouncing takes).  We need someone to dissect the f#@& out of parts of these OS's that Nobody seems to be familiar with, and unfortunately, I just don't have the time to explore that.

[Slion 1,201]

Are you guys still on stock? Try switch to Lineage OS. I've have had micro lags issues on stock but not on LOS.

[EskeRahn 5,460]

On 10/20/2021 at 6:08 AM, adam.c.r.roman said:

there's something in the Software where it's not sampling and debouncing the keyboard fast enough, regardless of moderate physical degradation to the keyboard.

I agree that it sounds like a software priority issue. But rather than lack of priority I would guess something else having too high priority thus interfering with the keyboard operation, as not everyone reports it. But could still be a system isssue, where some other system stuff gets activated by some actions or apps not all use or use the same way.

Starting with the ones power consumption points at as posssible culprits, I  would try to stop (or even deactive/uninstall) stuff and see if whatever triggers the phenomenon can be narrowed down.

Ultimately trying a factory reset, and test for the issue before reinstalling/restoring.

[auvo.salmi 135]

On 10/24/2021 at 11:59 AM, Slion said:

Are you guys still on stock? Try switch to Lineage OS. I've have had micro lags issues on stock but not on LOS.

I have had issues on both LOS and AICP. Haven't tried stock for ages though. But I have to say that incidence of those strange issues when keyboard suddenly provides some strange characters, usually including a number inside a word, did decrease greatly when I switched to AICP.

[EskeRahn 5,460]

7 hours ago, auvo.salmi said:

I have had issues on both LOS and AICP. Haven't tried stock for ages though. But I have to say that incidence of those strange issues when keyboard suddenly provides some strange characters, usually including a number inside a word, did decrease greatly when I switched to AICP.

When typing very fast, and perhaps having more than one key down at a time, there are N-key-rollover issues. Could that be the issue? And if I remember right, it was @TDM that made a new variant of the keyboard-driver that improved this for Lineage 16 (that I assume is still in use on 18 and AICP)

[VaZso 1,998]

13 hours ago, EskeRahn said:

When typing very fast, and perhaps having more than one key down at a time, there are N-key-rollover issues. Could that be the issue? And if I remember right, it was @TDM that made a new variant of the keyboard-driver that improved this for Lineage 16 (that I assume is still in use on 18 and AICP)

Right, the new driver should has less issue.

Anyway, I also experience this under LineageOS 18.1, maybe mostly 'n' key what appears for me as "ghost" key on my QWERTZ layout, I think that is 'b' on shifted QWERTY.

[EskeRahn 5,460]

58 minutes ago, VaZso said:

...., maybe mostly 'n' key what appears for me as "ghost" key ....

Ah so that is why it is called N-key-rollover. ... Sorry, could not help it. *LOL*

[Wheeljack 212]

I've just installed the LOS 18.1 Nov 1st update. And OMG... what a difference that new keyboard driver makes.

[Slion 1,201]

2 hours ago, Wheeljack said:

I've just installed the LOS 18.1 Nov 1st update. And OMG... what a difference that new keyboard driver makes.

What was changed?

[Wheeljack 212]

2 hours ago, Slion said:

What was changed?

https://review.lineageos.org/c/LineageOS/android_kernel_fxtec_msm8998/+/318274/

[Sean McCreary 328]

8 hours ago, Wheeljack said:

I've just installed the LOS 18.1 Nov 1st update. And OMG... what a difference that new keyboard driver makes.

We're really glad it helps :-)

[Laska 41]

Could this new driver be installed on stock?

[auvo.salmi 135]

On 10/28/2021 at 10:23 AM, EskeRahn said:

When typing very fast, and perhaps having more than one key down at a time, there are N-key-rollover issues. Could that be the issue? And if I remember right, it was @TDM that made a new variant of the keyboard-driver that improved this for Lineage 16 (that I assume is still in use on 18 and AICP)

Not in my case at least. I usually get double a, l, o etc.

edit: my device is QWERTY

Edited November 4, 2021 by auvo.salmi

[PatBue 4]

I have had the same problem for a few days. Sometimes it seems to me that the leather / fabric cover is statically charging the display and causing the effect. When you slide it out of the case, the typing starts. Strange... 😕

Edited November 22, 2021 by PatBue

[EskeRahn 5,460]

4 hours ago, PatBue said:

I have had the same problem for a few days. Sometimes it seems to me that the leather / fabric cover is statically charging the display and causing the effect. When you slide it out of the case, the typing starts. Strange... 😕

This could make a LOT of sense. If I try to operate the Pro1 with a cheap slow charger that is not grounded, I see the issue too, but never unplugged nor plugged to a grounded device like the PC. And surface currents and static electricity is sort of related, being (relatively) high voltage and low current.

[PatBue 4]

34 minutes ago, EskeRahn said:

This could make a LOT of sense. If I try to operate the Pro1 with a cheap slow charger that is not grounded, I see the issue too, but never unplugged nor plugged to a grounded device like the PC. And surface currents and static electricity is sort of related, being (relatively) high voltage and low current.

Aah. Okay ... well, I'll stay relaxed and take it as a lovable character trait of the device and still change the case 🙂

[adam.c.r.roman 25]

On 11/1/2021 at 11:47 AM, Sean McCreary said:

We're really glad it helps 🙂

Sorry, haven't seen this thread in a while.  (I don't get alerts for some reason.)  I'm delighted that my ranting about debouncing and key-scan priority has seemingly made some significant difference.  In a proper design, keyboard debounccing (see that?) wouldn't be the responsibility of the host processor, but rather be the sole responsibility of a microcontroller in charge of the keyboard.  Just look into the original Intel 8048 microcontroller in Every IBM PC keyboard in the 80's.  Oh, and a similar dedicated microcontroller in Every Other Computer Keyboard Since!  There's a reason for this!!!!!  They don't need to worry about other tasks having higher priority, they're just doing their fucking job properly, on the millisecond!!  They simply always scan, ddebounce, andd report proper data to the host... done, and done!  (Again, actual key repeats there.)  But in this case, brilliant people like Sean McCreary and his collaborators have had to find workarounds to a flawed system design!

Now, I know the hardware isn't going to change for the better any time soon, and this seems to be a very effective fix...

So I have to repeat the question posed by Laska on 11/2/21 - Can we get this driver in the stock OS from f(x)tec?  Can we be sent updates soon??  And they should Certainly be shipping the Pro1-X with ANYTHING that can help the keyboard issues.  Because we've proven that this issue only gets worse over time.  I say all this with such enthusiasm because I really do love this phone and just want it to live up to its potential.  It's changedd my life for the better, but is still constantly annoying!

[claude0001 1,341]

4 hours ago, adam.c.r.roman said:

Can we get this driver in the stock OS from f(x)tec?

I guess, the simple answer to this is (and has been for a while): No.

The official Android 9 for the Pro1 is dead. I do not expect that there will be any more fixes from F(x)tec.

Although the stock keyboard driver is open-sourced, producing a fixed driver and installing it in the stock ROM in binary form is probably not possible because of code signing. I am no expert on this though.

The pragmatic way of getting better keyboard support would be to simply install a recent LineageOS release, which improved on the duplicate-stroke issue (besides many more enhancements). The relevant patches are here and here. Note that the keyboard driver (the qx1000.c source against which the patches are applied) was completely rewritten by @tdm as part of the initial port to Lineage OS 16.0, so these patches would probably not apply directly to the Stock sources.

Edited February 25, 2022 by claude0001

[Sean McCreary 328]

On 2/24/2022 at 11:51 PM, adam.c.r.roman said:

I'm delighted that my ranting about debouncing and key-scan priority has seemingly made some significant difference.

You really should have opened a bug report, as I never saw your complaints about duplicated keys.