Connecting to Bluetooth classic keyboards

[kaspar]

Originally Posted by dhdurgee

Would a powered hub solve that problem?

Dave

It does for my Pomera DM250 running Debian. I run this USB C OTG cable with a battery pack and the USB A keyboard dongle for the ThinkPad TrackPoint Keyboard II, which has the red mouse dot, so it is keyboard and mouse in one device.

[Frogm4n]

Ah, ha. I've figured out some of my keyboard issues! This BT keyboard of mine has multiple modes for different OSs, such as Android, Windows, and iOS. The default keys are different across the modes. Turns out it was in iOS mode likely due to being used with an iPad earlier, and that means it wasn't out putting PgUp and PgDn keycodes, on top of using media keys instead of the F-keys. Putting it into Android mode will send PgUp/PgDN keycodes, but still defaults to media keys. Windows mode defaults to F-keys and all other expected keycodes.

Still haven't gotten any page turn response in the stock Kindle reader, though. I'm curious what evtest output codes people get on keyboards that do work to turn pages in the stock Kindle UI.

[clint32]

I unfortunately missed a small but important step in the first post

Once reloading the udev rules, you'll also want udevadm trigger to put the new one into effect.

Originally Posted by kbarni

We tried to debug my issue with Clint and Sighery, it's a problem with the authentication using the pairing process. The keyboard asks for a pairing confirmation that doesn't show up in the ace_bt_cli interface.

I do think there's still a possibility of pairing devices that require a PIN -- I don't personally have any to test with, though.

If anyone else wants to test -- once you run pair <addr> you can then run pin 0000 while the pairing is still in progress and it hasn't timed out. (The output is confusing because it writes over the prompt, but running the pin is only possible after a pairing address has been set internally.)

[clint32]

Originally Posted by Frogm4n

Ah, ha. I've figured out some of my keyboard issues! This BT keyboard of mine has multiple modes for different OSs, such as Android, Windows, and iOS. The default keys are different across the modes. Turns out it was in iOS mode likely due to being used with an iPad earlier, and that means it wasn't out putting PgUp and PgDn keycodes, on top of using media keys instead of the F-keys. Putting it into Android mode will send PgUp/PgDN keycodes, but still defaults to media keys. Windows mode defaults to F-keys and all other expected keycodes.

Still haven't gotten any page turn response in the stock Kindle reader, though. I'm curious what evtest output codes people get on keyboards that do work to turn pages in the stock Kindle UI.

There's an unfortunate snag with the page up/page down to turn the page: it depends on your Kindle model

These keys are handled specially by the window manager (in /etc/xdg/awesome/lab126_whisper_touch.lua). These keys are used by Kindles with physical page turn buttons, and the downside is the window manager checks device capabilities to decide whether these will be permitted key presses.

If you don't have a device with these buttons, you'll see an error in /var/log/messages along the lines of:

E WindowManager:whisperTouchButton Handled By WinMgr:reason:featureNotEnabled

Similarly, newer models with an accelerometer and the double-tap-to-turn feature use this same system. For these, F7 will turn the page. Pressing it on a device without the accelerometer will also get blocked by the window manager.

If you have xdotool (which you can grab from another Linux distro), you can work around this by sending to the X window of the reader directly. (Using the default root window will cause it to get blocked by the WM.)

xdotool key --window $(xdotool getactivewindow) Page_Down

I have a small program in the works (useful with my 8bitdo micro, functioning as a gamepad or normal keyboard) to forward along/remap these key presses (and a bit more, like changing brightness and orientation). It's not quite ready for release yet, but once I write a bit of documentation and put together some KUAL helpers, I'll be sure to share it here too

[matthew_h]

I was able to pair a magic keyboard that wanted me to confirm the pin by running the pair command through ssh while my kindle was in Bluetooth settings and having a conformation pop up on the Kindle.

[Frogm4n]

Originally Posted by Frogm4n

I've discovered unexpected functionality on my Scribe running 5.18.1: Pressing F1 - Force screen rotation normal, F2 - Force screen rotation inverted. Will be over-ridden by sensor when the Scribe is moved. Seems to work in all situations in the Kindle UI, but still no response at all in KOReader. If I do get this updated to 5.18.5, I'll see how the new rotation lock responds vs this current behavior.

I did the 5.18.5 update today so that I could test the rotation lock. First I updated to the 2.2.3 Hot Fix, and then did the 5.18.5 update and jailbreak recovery. The update wiped the script and the rules file from this mod. I put them back, and luckily found out that it kept the BT pairing, at least! Keyboard works fine again after restoring the files.

The Scribe screen rotation lock seems to disable the sensor. F1 and F2 still force rotation, and the screen stays that way when locked. Like I noticed earlier, the sensor outputs rotation event ESC codes when running ace_bt_cli but when the rotation lock is enabled it does not.

[esverde]

Is it can be paired with game controller? I just have a single-sided controller, I think it is very suitable for turning page!

[Ruinemacil]

Hi everyone! ace_bt_cli gets stuck here:

[root@kindle ]# ace_bt_cli
ACEBTCLI BT Client :: start
p_data (size:27) = 1B 00 00 00 00 00 00 00 00 00 00 61 63 65 5F 62
74 5F 63 6C 69 00 00 00 00 00 00
ACEBTCLI BT Client :: opened session with 0x2afc0

>: pair D8:EB:FD:86:4C:FB
ACEBTCLI str: D8EBFD864CFB
ACEBTCLI Initiate Pairing Success

>: ACEBTCLI ACEBT_DISCOVERY_STATE_STOPPED
ACEBTCLI CLI callback : bond state changed() status: 0 addr: D8:EB:FD:86:4C:FB state: BONDING
ACEBTCLI CLI callback : bond state changed() status: 0 addr: D8:EB:FD:86:4C:FB state: BONDING
ACEBTCLI CLI callback : bond state changed() status: -206 addr: D8:EB:FD:86:4C:FB state: NONE
ACEBTCLI CLI callback : bond state changed() status: -206 addr: D8:EB:FD:86:4C:FB state: NONE

Any ideas?