marmistrz

I must've missed that, thanks!

Out of curiosity, I guess that IGG perks have no warranty either?

tdm, could we possibly have both drivers in the lineage kernel and be able to choose between them without rebuilding the kernel (possibly requiring root)? I'm thinking about a way to adapt the keyboard layout to fit the Polish language and I'll almost certainly will need to somehow remap some fn+something combo. Probably I'd want to use fn+a as altgr+a but keep fn+l as it is. There might be other languages for which a dead fn key causes trouble.

What's the purpose of the changes in the "Add keylogger" commit? https://github.com/tdm/android_kernel_fxtec_msm8998/commit/b03a10e872bc451e7300a1d5ed93204f3c3468b1

@tdm is there a technical reason for that or is it just a bug?

No, but I assumed internal storage should work equally well (when soneone says sdcard in the context of Android, I often think /sdcard, which is often mapped to the internal storage)

There's nothing like this for me...

Does LineageOS recovery allow flashing zips from sdcard? I didn't see such an option anywhere and I thought only sideload is possible.

I tried OTA with 20200914 and again a softbrick, using the `adb sideload` method unbricked the device. I'm wondering about the root cause. One hypothesis is that I wrote /system. I followed this guide of mine to install the F-Droid Privileged Extension. Touching /system in any way will break OTAs on stock, and I'm wondering if the same may be the case for LineageOS. Saying "just don't write system" is not a solution, because it's impossible to modify /etc/hosts without writing /system 🙂

Is the 20200831 OTA safe to install? there was a lot of trouble with 20200824 (system not booting), wondering if anyone had similar issues with 20200831 Or should I just flash the update with adb sideload?

As for GPS, yesterday I tried to get a cold fix while traveling in a car and it took about 20 minutes, even though there were roughly 12 satellites visible all the time. It was in Poland. Today it was much better and it the fix was almost instant. Looks like under some circumstances the GPS fix is terribly slow.

I tried switching in fastboot and it didn't work. I ended up wiping data 😞

There is certainly a hardware issue with my device. After long time of usage I noticed spontaneous key events for keys that I didn't even touch. Looks like the keyboard is short-circuiting. I'd send it back for warranty, but since people have waited 7 months for the replacement, I'm a little afraid to send it back for warranty until they finish the pre-orders. Btw. I also have issues with the USB-C port disconnecting while moving the cable slightly and suspicions of hardware issues with the camera sensor (remaining glue or other dirt in the camera unit)

Same here, no gapps at all. I tried downgrading (unsupported, I know), but the recovery doesn't accept sideloading in either case (both 20200803 and 20200824). I'm simply getting adb: sideload connection failed: no devices/emulators found adb: trying pre-KitKat sideload method... adb: pre-KitKat sideload connection failed: no devices/emulators found

As for autocapitalization, it seems to be some high-level mechanism. I noticed that when I'm writing im in the Messaging app, it's automatically corrected to I'm, even when I'm using Polish for both hardware and software keyboard. This doesn't happen in some other apps, such as Conversations. Btw. it appears that on my old OnePlus 3 running LineageOS 14.1 there's an option for autocapitalization when I plug a USB keyboard, however, this option does nothing. There's no such option on LineageOS 16.0 on the Pro1.

Could you please provide instructions for updates from test22 to official in the OP? It was discussed somewhere, but I don't remember what the conclusion was

I managed to fix the problems with GPS by clearing all caches. Now I'm getting a fix within 5-10s, GPS, GLONASS and Beidou is used. # rm -rf /data/cache # rm -rf /data/dalvik-cache

I'm not sure if it's just the configuration: some time ago, on test16 GPS worked and I used to get a fix within less than a minute. Now I don't get any fix. I don't know if GPS broke directly after updating to test22 or later on. I updated from test22 directly from test16, without any intermediate releases. Maybe I could try wiping all the caches? I know how to do it with TWRP, but the LOS recovery doesn't provide such an option.

I'm always using SatStat and OsmAnd.

I have recently updated from test16 to test22 and today I noticed I can no longer get a GPS fix. SatStat sees a couple of satellites, but marks them red. I attached logcat. I tried clearing the data for com.qualcomm.location and FusedLocation, but it doesn't help. I haven't updated to the official releases yet, but since test22 and current official are mostly the same, this shouldn't matter. Let me know if I should try it. I'd prefer to avoid wiping everything for obvious reasons. gpslogcat.txt

Well, the point is that even while building in-tree, I won't be able to build the kernel unless I apply this patch. It looks like `make bootimage` (and in general the kernel build system) uses the host compiler at some early stage to compile dtc instead of the Lineage DT compiler. This makes sense, because the Lineage DT compiler is an arm64 cross-compiler and dtc will be run on the host and needs to be an x86 binary. SHIPPED scripts/dtc/dtc-lexer.lex.c HOSTCC scripts/dtc/dtc-lexer.lex.o HOSTLD scripts/dtc/dtc

@tdmwe need to cherry-pick this commit to enable build of gcc 10 hosts and regenerate the "shipped" files. (dtc is built using the host compiler and upstream doesn't contain the "shipped" files) https://github.com/torvalds/linux/commit/e33a814e772cdc36436c8c188d8c42d019fda639 I would've done it myself, but I'm not sure how to properly cherry-pick with gerrit to preserve authorship.

Thanks for your reply. I'll try building it within the Lineage tree. You mean an entry in the settings app, so that people can toggle it without command-line tweaking? I can't join the Telegram group without providing my phone number, which I'm unwilling to do. Is there any IRC/Matrix/XMPP/... room/bridge?

Okay, so I'd like to contribute to LineageOS for Pro1, and there are two things I think of 1. Add a possibility to turn off autocapitalization 2. Implement sticky fn. The idea here is to store a boolean flag `fn_sticky_pressed` and set it `aw9523b_check_keys` while checking the state of the keys So I have questions 1. In which part of the software stack is the autocapitalization for hw keyboard handled? This doesn't look like in the kernel driver 2. What's the correct way of building the kernel? I did the following on Arch Linux, gcc 10.1.0 export ARCH=arm64 export SUBAR

Ok, I see what's the reason for the unexpected behavior of alt/fn+alt. fn is handled using these arrays: https://github.com/LineageOS/android_kernel_fxtec_msm8998/blob/2120e09d09b3a94c4de9395523a587d9909224d4/drivers/input/keyboard/qx1000.c#L257-L335 It appears that KP_LEFTALT is handled in this part of the file: https://github.com/LineageOS/android_kernel_fxtec_msm8998/blob/2120e09d09b3a94c4de9395523a587d9909224d4/drivers/input/keyboard/qx1000.c#L649-L654 and no check is made for the state of the fn key. What kind of key event is emitted on stock for fn+alt? (I can't find the kernel