I know this would be possible to use to swap the modifier keys, but in terms of the way those keys change when interacting with certain applications (i.e, terminals), is that something that's within the scope of this project? Or is that something that's going to be limited to display servers? Basically can this project do something similar to Kinto?

As I said in the title, the same key is mapped to different keys depending on the application.

For example, mapping C-a to Home in an application like firefox
In emacs C-a is still C-a

Man pages don't need to be compressed the work the same as long as they are installed to somewhere in $MANPATH.

Most software in my experience always deal with plaintext uncompressed man pages. The man page compression is always done by distro packagers.

It will also be better because git won't need to deal with binary files.

This should be simple to fix by amending the Makefile. I can make a PR if you agree.

FreeBSD is one of the few (only?) BSDs that implement evdev and uinput. It may be possible to port keyd to it with minimal (no?) changes. If anyone has success compiling and running it please post here. I might this give this a go later.

Having some trouble binding these two keys. My config is

pageup = home
pagedown = prtsc
rightshift = \

The Pageup key is changed correctly, but Pagedown has no effect. Curiously, it does type a Tilde if I press it while in a terminal. I couldn't see PrtSc mentioned anywhere in the keyd -l command, is it not usable?

And as for the Right Shift bind, my keyboard is set to UK Layout in my system settings, which seems to make the backslash key register as the Hash/Tilde key instead. I tried binding the key to Hash to see how that would affect things, and then it would only print a pound sign.
The binding works on Windows using SharpKeys, and it names the key I have it bound to as "ISO Extra Key 00_56". Any way I could bind that key properly as well?

When I create a layer triggered on leftalt, with a parent layout as modifier 'M', with key mappings that trigger leftalt-* combinations, I get a strange effect where the original parent layer ('M' in this case) is triggered in addition to the A-1 or A-2 keys:

leftalt = layer(macmeta)

[macmeta:M]
1 = A-1
2 = A-2

When I run keyd -m I see that there are indeed multiple keydown and keyup events--the meta key goes down/up, then alt goes down and 1 goes down, then alt goes up. (Note that only Alt-1 is being pressed physically):

keyd virtual keyboard: meta down
keyd virtual keyboard: meta up
keyd virtual keyboard: alt down
keyd virtual keyboard: 1 down
^[1keyd virtual keyboard: alt up
keyd virtual keyboard: 1 up

I'm going to reboot and see if anything changes, but I wanted to leave this here in case the data has value in replication for anyone else.

Hi, I was using keyd successfully for a long time now, but it started disabling mouse buttons.

Config

# Turns capslock into an escape key when pressed and a control key when held.
capslock = overload(M, esc)

Env

OS: Arch Linux - rolling
keyd version:

pacman -Q keyd-git                                                                                     
keyd-git 1.1.2.r14.95e1d89-1

Mouse HW: Logitech MX Master 3 (doesn't work with neither bluetooth nor unifying receiver)

sudo keyd -m and clicking with the mouse

sudo keyd -m  
keyd: src/main.c:764: evdev_monitor_loop: Assertion `keycode_table[ev.code].name' failed.
[1]    8299 abort      sudo keyd -m

First of all: as a user with a QMK keyboard that also has to regularly type on a laptop your keyd is a blessing.

As a diehard vim guy I quickly was able to translate the most important stuff of my qmk setup to keyd:

# Turns capslock into an escape key when pressed and a control key when held.
capslock = overload(C, esc)

# Remaps the escape key to capslock
# esc = capslock

# space when tapped, spacefn when held
space = overload(spacefn, space)
tab = overload(tabfn, tab)

[spacefn]
h = left
j = down
k = up
l = right

[tabfn]
h = home
j = pagedown
k = pageup
l = end

• caps as ctrl/esc works like a charm, bye xcape
• initial tests with SpaceFn and TabFn work: nice, bye xmodmap+setxkbmap

Then I started to write an email and realizedthatthespace keyddoesnt work asreliable :/
My typing speed is non-spectacular (75-80 wpm according to monkeytype.com) but a space doesnt come through two thirds of the time.

Is there anything I can do to make this more accurate? Did I configure something wrong?

Just installed keyd and used it to remap home to delete, because my keyboard manufacturer thinks that one is more important than the other. And it only took me 2 minutes. -- ImJustPassinBy

I use keyd for everything and love it! -- that1communist

I have been having a play with keyd today and it's really neat... It's making the capslock key, which has been mapped to esc as a default for vim, into a much more versatile key triggering the different modes and using esc to return to the main mapping. Super simple configuration and potential to be open the keyboard up a bit without having to buy a Planck just yet. -- textandmetal

When I try to reassign < and >, keyd only considers the buttons located at , and . and not the buttons located at the left of my z key.

keyd -m reports AT Translated Set 2 keyboard: 102nd down/up for this button.

Is there a way to use the numeric code for assignment?

Hi

The keys remapped to a macro( .. ) are triggered once per press and do not get repeated as normal keys when the key sequence is held down.

To reproduce:

leftshift = macro(xx)

Pressing and holding leftshift will print "xx" only once and will never repeat.

Expected behavior:

Repeating is expected by default.

In case of macros (IMHO) repeating the whole macro would be logical, also by default.

Would be nice to be able to control in config:

• toggle repeating for a key,
• for macros set key sequence (or sequences) to be repeated.

Thanks for a great piece of software!

P.S. I'm not suggesting a specific syntax here as I both do not have the feel for it yet and I did not dive into the implementation (yet)...

When using keyd to intercept keys on a laptop with a touchpad, the "disable while typing" feature is neutered: with keyd enabled and while typing, it is very common for the touchpad to emit a "BTN_LEFT" event aimed at wherever the mouse pointer happens to be hovering (I usually can't type two complete sentences without it happening).

In the Readme, Recommended Config where you have capslock = overload(esc, C) I actually had to switch it to capslock = overload(C, esc) to get it to work! Just wanted to let you know

Sometimes keyd gets "stuck" on an oneshot layer. This causes every key press to have the stuck oneshot layer applied to
it. This problem is reproducible with the following config:

# Makes the shift key sticky for one keystroke.
leftshift = oneshot(S)
rightshift = oneshot(S)

To trigger the stuck state, the following can be done:

1. Hold down left shift and right shift
2. Press any letter on the keyboard once
3. Release left shift and right shift

Now every key press will have the shift modifier applied which looks like caps lock is enabled. To get out of the stuck
state either left shift or right shift has to be pressed again.

This stuck effect is not specific to left and right shift. I have managed to trigger this effect on oneshot mappings for
AltGr aswell sometimes. However, I only found left shift and right shift to consistently trigger this effect for me.

So I would guess, that there is some issue when two oneshot layers are triggered at the same time or with a small time difference to each other.

Tested with keyd version: 1.1.2 (8131320)

i want to remap F1 to esc because my escape key is broken but F1 = esc doesnt work

Hot off the heels of #66... :)

I'm curious if you've considered using libinput rather than interfacing directly with evdev? I know this would imply a large rewrite of some key parts of the code, but for the sake of imagining possibilities, perhaps we could consider it just for a moment?

(This is also related to linuxtouchpad discussion -- the last question in particular, "should key remappers be implemented at the same level as touchpad remappers (like Touchégg and Gebaar)?")

1. I learned recently about how gesture remappers like gebaar and Touchegg are similar to key remappers. I've diagrammed the "overview" from the perspective of a touchpad here (the idea is exactly the same for keyboards). In essence, instead of opening evdev file descriptors directly, gesture remappers have to rely on libinput to detect the gestures before they can be remapped. Then, the next layer in the stack opens their "virtual device" and uses that input instead of the "real" device.
2. libinput handles keyboard events, mouse events, touchpad events, touch screen events, and tablet events. These events are all at a fairly "low level" in the sense that there is some normalization going on, but not a lot of interpretation. And yet, it seems like input devices generally are often related--sometimes multiple kernel devices are the same physical device, and sometimes two or more devices interact. In fact, #42 and #66 are sort of proof of how they are related in sometimes surprising ways.
3. libinput is now used almost universally in X.Org (Xserver) and Wayland compositors (Weston, mutter, KWin). The one place it may not be universal is in text terminals.

So there are clearly some benefits to using libinput as a source of events, at least at the boundary around something like keyd:

1. You get some normalization for free (for example if combining a keyboard with a mouse--libinput has a hardware database that provides accel curves and other benefits to make mice behave)
2. You get beneficial device interactions for free (e.g. touchpad "disable while typing")
3. You get support for both major graphical desktop systems (granted, this is true of evdev too)
4. You can start to piece together advanced remapping scenarios where gestures and keys combine to make other possibilities (imagine "Ctrl + 3-finger-swipe right" for instance).

So I guess I'm just spitballing a bit here, but am I missing anything important? Let's say keyd 3.0 uses libinput and we're all done. Is it a net improvement?

Hello, I am running version 1.1.2 and my config is:

capslock = overload(arrows, esc)

[arrows:main]
h = left
j = down
k = up
l = right

But this config seems to only be applied sometimes.
The config is stored in '/etc/keyd/AT Translated Set 2 keyboard.cfg' and the device name is AT Translated Set 2 keyboard.

When the config is applied properly the logs are:

-- Boot 9322d9a7267e4abfacc26c60cd9beb1f --
ago 29 16:36:21 amos systemd[1]: Started key remapping daemon.
ago 29 16:36:21 amos keyd[366]: Starting keyd v1.1.2 (8df75b749b9131faeb551c84efebca9762047649).
ago 29 16:36:22 amos keyd[366]: Managing AT Translated Set 2 keyboard

Otherwise they are:

-- Boot a83ee4fbd1ad4e1d8866ad2c8c2dedd7 --
ago 29 16:38:45 amos systemd[1]: Started key remapping daemon.
ago 29 16:38:45 amos keyd[365]: Starting keyd v1.1.2 (8df75b749b9131faeb551c84efebca9762047649).

Restarting the daemon fixes part of the problem.
While the capslock is pressed the arrows layer is available, but when it is tapped, it does not work like escape.

The logs are:

-- Boot a83ee4fbd1ad4e1d8866ad2c8c2dedd7 --
ago 29 16:38:45 amos systemd[1]: Started key remapping daemon.
ago 29 16:38:45 amos keyd[365]: Starting keyd v1.1.2 (8df75b749b9131faeb551c84efebca9762047649).
ago 29 16:45:08 amos keyd[365]: SIGTERM received, cleaning up and terminating...
ago 29 16:45:08 amos systemd[1]: Stopping key remapping daemon...
ago 29 16:45:09 amos systemd[1]: keyd.service: Deactivated successfully.
ago 29 16:45:09 amos systemd[1]: Stopped key remapping daemon.
ago 29 16:45:09 amos systemd[1]: Started key remapping daemon.
ago 29 16:45:09 amos keyd[3378]: Starting keyd v1.1.2 (8df75b749b9131faeb551c84efebca9762047649).
ago 29 16:45:09 amos keyd[3378]: Managing AT Translated Set 2 keyboard

Do you know how to fix this issue?

Thank you for your amazing work.

~/repos/keyd (master|✔) [2] ▶ make
mkdir -p bin
cc -DVERSION=\"1.1.2\" -DGIT_COMMIT_HASH=\"231e3ebd1cb58607e21952a0ab93030970d22010\" -DCONFIG_DIR=\""/etc/keyd"\" -DLOG_FILE=\""/var/log/keyd.log"\" -DLOCK_FILE=\""/var/lock/keyd.lock"\" -I/usr/local/include -L/usr/local/lib -O3 src/*.c -o bin/keyd -ludev
src/main.c: In function ‘monitor_exit’:
src/main.c:827:26: error: parameter name omitted
  827 | static void monitor_exit(int)
      |                          ^~~

It looks like a parameter name is missing for monitor_exit.

I maintain the AUR package for Arch Linux.

PKGBUILD: https://aur.archlinux.org/cgit/aur.git/tree/PKGBUILD?h=keyd-git

If you want you can add it to the README. No need for instructions. Arch users should know how to use it.

Add ability to comment a line, maybe ';' or '//' ?

• It would be nice during testing to be able to comment out a line
• Actual comments above a line or section would be helpful for future readers

I'm curious if there's interest in supporting getting the current active window, and automatically switching to different keyboard layers.

A simple use case: VSCode uses Ctrl+, as the "preferences" shortcut key, but Firefox uses "Alt+e, n" (and can't be configured without a special plugin). I'd like to unify this behavior so that Ctrl+, always gets me to preferences.

Two other key remapping tools accomplish this in X11: xkeysnail and autokey. Unfortunately, they both do not support Wayland.

It turns out that in Wayland, getting the "current active window" is not supported. Instead, they punt on it and ask that a layer above handles that task, e.g. at gnome-shell:

gdbus call --session \
  --dest org.gnome.Shell \
  --object-path /org/gnome/Shell \
  --method org.gnome.Shell.Eval "global.display.focus_window.get_wm_class()"

It may also be possible to write a dbus listener that gets notified of window focus changes in gnome. That seems like it could be a separate project--one that perhaps ties in to keyd. Perhaps such a daemon could tap into each of the major shells/window managers and report the necessary information?

I think the "key remapping community" is in need of a solution here, because in general, I don't think it has been solved.

Is it possible to have dead and compose keys (e.g. AltGr+[ for ¨ to create characters like ü etc.) with keyd? The compose option does not seem to do this.

(I could probably create a new layer, but then I think that such characters are not available?!)

Hey man. Been using keyd for about a month now and have zero (no) complains, works like a charm.
I see you've added mouse button support? How can I use this?

I have had this ugly dirthack in my i3 config for years, that triggers xdotool key f5 when i press forward button (btn 5?) and xdotool key backspace on backbutton. Would be awesome if that could be configured with keyd instead.

Could you give me an example on how this can be done, if it can be done?

Also, this is kind of related. Back when i was using windows i had this Autohotkey script that made the left and right buttons act kind of like the overload function. So holding f.i. right mousebutton and scrolling would send control+PageUp/Down. It was amazing, is that possible with keyd, maybe, please ;)

Here is my old mouse remapping script:
https://gist.github.com/budRich/6044613

I also had stuff like tripple click and double click right/middle button for copy,paste and enter. good times.

Keep up the awesome work!

I know you just said that hopefully the output format wouldn't change, but ... since it's so fresh, perhaps I can suggest this?

What if we use tab ("\t") to separate the three fields? I think it would be easier to parse.

So this:

AT Translated Set 2 keyboard(0001:0001): a down
AT Translated Set 2 keyboard(0001:0001): a up

Would become:

AT Translated Set 2 keyboard     0001:0001       a down
AT Translated Set 2 keyboard     0001:0001       a up

This solves several potential problems:

1. It's possible to have parens in keyboard names
2. The : (colon) separator is no longer ambiguous
3. "\t" should be a "safe" separator character
4. It's easier to use a separator than create a regex to pull info out of a line

Hello, I use dvp and russian layout.

keyd didn't work with setxkbmap -option ctrl:nocaps -option grp:win_space_toggle us,ru -variant dvp,typewriter

I disable dvp and create dvp in default.cfg:

layout(dvorak)

[mylayer]
; = $
a = 0
s = 1
d = 2
f = 3
q = $
z = grave
w = 7
e = 8
r = 9
x = 4
c = 5
v = 6

[dvorak]
meta = meta
space = space
rightalt = layer(mylayer)
leftalt = leftcontrol
compose = compose
esc = esc
capslock = leftalt
leftcontrol = leftalt
rightcontrol = rightcontrol
leftshift = leftshift
rightshift = rightshift
enter = enter
tab = tab
backspace = backspace
up = up
down = down
left = left
right = right
mode = mode
q = semicolon
w = comma
e = dot
r = p
t = y
y = f
u = g
i = c
#....etc

after I chose Russian setxkbmap -option grp:win_space_toggle us,ru -variant ,typewriterwith , keymap use old (US). How implement Russian in cfg and change layout without setxkbmap? Only keyd?

Hi,

I have the following settings in my xkbmap:

clear mod1
keycode 108 = Mode_switch
keycode 64 = Super_L
keysym e = e E EuroSign
keysym c = c C cent
keysym w = w W adiaeresis Adiaeresis
keysym o = o O odiaeresis Odiaeresis
keysym u = u U udiaeresis Udiaeresis
keysym r = r R ssharp
keysym q = q Q acircumflex Acircumflex
keysym a = a A abreve Abreve
keysym s = s S U0219 U0218
keysym t = t T U021B U021A
keysym i = i I U00EE U00CE

and I've tried using keyd to generate the same:

rightalt = layer(dia)

[dia]
r = ß
u = ü
o = ö
e = €
...etc...

As I understand I cannot use r = G-r and so on because keyd resets all the xkbmap mappings. is it possible to generate these keysims with keyd? perhaps via a macro calling xdotool or something like that?

Thanks.

It seems like it's absolutely worth spending all my efforts and I'd have loved to find this nice tool a long time ago right now, but at least for today, I've run into some problems that I cannot resolve easily and could not find solutions in the documentation. Probably some FAQ would help.

Additionally to the previous issues reported, I have the following two issues that I could not figure yet.

First: keyd overwrites the keys I have defined in my compositor (and it does not - as I had imagined at first - work in a way that it would only change the keys that I assign in /etc/keyd/* and leave the rest be). I've yet to exactly figure out how to apply setxkbmap or other options on top, or how to re-configure them.

Second (could be more abstract, but I'd like to show the use case here also): I have meta+shift+2 assigned in my compositor, and I've also assigned S-2 = " in keyd. keyd seems to take priority, i.e. I receive " instead of moving my current window to another tag when pressing meta+shift+2. Do I have to define another symbol for this combination, or even do something like M-S-" = 2?

(Second point five: when I have G-2 and S-2 assigned, which one takes priority?)

Can I make something like this:
press ctrl+super+m -> press mouse button 3
release ctrl+super+m -> release mouse button 3

Hello, first of all, thanks for the project, it steps up the usability and automation scene in Linux a lot!

I was wondering if it would be possible to add support to configure the timings that keyd uses in order to consider when a key is held or not.

There's a similar tool for MacOS called Karabiner, and it let's you configure the following parameters:

:alone   80 ; hold for .. ms to register single tap
:held    50 ; If it has been held for this time and no other simultaneous key has been pressed, it considers this as a tap 
:delay   0  ; time after which the key press is delayed
:sim     30 ; keys need to be pressed within this threshold to be considered simultaneous

I'm trying to migrate my MacOS workflow to my Linux machine, but currently I'm having trouble due to this, as I use regular keys to enable layers.

For example: holding space enables the symbols layer, so I can press space + t and it will act as the = key, but with the current timings of keyd it makes very difficult to type, as I have to pause in order to type words, as space is constantly triggering the hold if I type too fast.

I'd really love to get this into keyd and of course, I'm volunteering to try implementing it myself, I just would need some guidelines on where to start.

Thanks again!

Hello!, or as we say in Sweden hallå!

I have been using a very hackish method for years to make a custom layout with setxkb. It has worked fine, but you know only in x. The gist of the layout is that i can use the "extra" key on nordic QWERTY keyboards to act as a modifier for a special layer, and without it i am completely lost.

It seems like i can get all functionality i had before with this keyd except remapping keys to swedish characters. I want this to work:

# Turns capslock into an escape key when pressed and a control key when held.
capslock = overload(C, esc)

# Remaps the escape key to capslock
esc = grave

102nd = layer(bud)

[bud]

h = left
j = down
k = up
l = right

[ = macro(å)
' = macro(ä)
; = macro(ö)

(i also tried without using macro [ = å, but that didn't work either.
When i press f.i å (with normal swedish layout active) when keyd -m is running it just prints [.

I haven't looked that closely at the source. And I imagine this gets tricky since supporting every single foreign character might make the keys.h gigantic? But for me it would be acceptable if macro made a Unicode translation of the chars added and printed that.

Anyways, thanks for a very cool program.

Debian 10, DWM.
tty = dvorak-programmer

/etc/keyd/default.cfg

rightalt = layer(mylayer)

[mylayer]
a = 0
o = 1
e = 2
u = 3

/etc/X11/xorg.conf.d/00-keyboard.conf

Section "InputClass"
        Identifier "system-keyboard"
        MatchIsKeyboard "on"
        Option "XkbLayout" "us,ru"
        Option "XkbModel" "pc104"
        Option "XkbVariant" "dvp,typewriter"
        Option "XkbOptions" "ctrl:nocaps, grp:win_space_toggle"
EndSection

Figured out that keyd -m does not track keycodes like xf86audiodown and similar. Would be nice to have, as some keyboards provide these codes =)

I have noticed this strange issue that everytime i opened htop it seemed like the control key got "stuck". And today I noticed the same thing when opening whiskermeny. it is when i press Control (actual control key) and Escape (actual capslock) in a key combination while having:
capslock = overload(C, esc) in my keyd config, but i think i fixed it with this hack:

capslock = overload(C, esc)

leftcontrol = layer(control)

[control:C]
capslock = C-escape

This might be related to other "overload" issues when using the both of the keys in the overload together.

*edit
the keybinding for whiskermenu is: Control+Esc
htop: Control+Shift+Esc

When I remap < and >, keyd only considers the buttons located at , and . and not the button left to my z key. This one is reported as [keyboard]: 102nd.

How do I assign this key?

< = \

I am using keyd to turn my Dvorak into qwerty when ctrl is held, so far I've been doing it by manually making, for example

A=C-a

But there has to be some way to make left Ctrl be left Ctrl AND switch the layers so I can just hold leftctrl and have that switch to qwerty while held.

Thanks for any help!

In the manpage an example is given to overload the escape key by using the overload action like this: overload(C, esc). However, under the "Actions" section the manpage states that the overload action takes the keyseq as the first argument and a layer as the second argument: overload(<keyseq>,<layer>)

I noticed this inconsistency while migrating my config where the overload action works in the format overload(<layer>, <keyseq>).
I don't know if there are other inconsistencies in the manpage, since I only noticed this issue explicitly. It probably is a good idea to have a look over it to see if everything is correct.

My goal is to bundle keyd in a deb, and then depend on it as part of a higher level app that makes a Mac-like experience easier on X11 & Wayland. To achieve this, there are two autodetection features that I've been thinking about. This relates to part of the discussion in #55 as well.

1. Get key left of spacebar: To detect whether the user has a Mac keyboard or a Windows keyboard plugged in, it seems a "best practice" is to ask the user to press the key "to the left of the spacebar".

2. Figure out which keyboard is being used: To detect which keyboard the user expects to have remapping applied, it would be user friendly to listen for all keyboards, and then check which one is the source of a key press event.

To accomplish both of the above, would it be possible to start keyd in monitor mode, but end after just one key press? I'm thinking I could tell the user "Please press the key to the left of spacebar" (like kinto does) and start keyd in monitor mode / autodetect mode. The keyd would respond with its usual line:

keyd virtual keyboard: alt down

it would then quit and the stdout could be parsed ("keyd virtual keyboard" and "alt down" would be enough to go on... although a device ID would be better).

Description

Typing something with AltGr modifier (eg. € in the English UK layout in X11) does not work while running keyd with empty config. All other modifiers like Alt and Ctrl work with empty config.

AltGr only works when adding the following to config:

rightalt = layer(G)

I am using X11 to change the keyboard layout. I'm not use if that changes anything.

Steps to reproduce

• Set keyboard layout to English UK (it has AltGr)
• Type AltGr+4 to try to type €.

Expected result

Type €.

Actual result

Type 4. AltGr is ignored.

keyd monitor logs

All of them are with the English UK layout set in X11.

Without running keyd

AT Translated Set 2 keyboard: rightalt down
AT Translated Set 2 keyboard: 4 down
€AT Translated Set 2 keyboard: 4 up
AT Translated Set 2 keyboard: rightalt up

Without rightalt = layer(G) in config and running keyd

keyd virtual keyboard: rightalt down
keyd virtual keyboard: rightalt up
keyd virtual keyboard: 4 down
4keyd virtual keyboard: 4 up

With rightalt = layer(G) in config and running keyd

keyd virtual keyboard: rightalt down
keyd virtual keyboard: 4 down
€keyd virtual keyboard: 4 up
keyd virtual keyboard: rightalt up

What I want:

ckb1: Corsair Gaming K95 RGB PLATINUM Keyboard vKB: leftmeta down
ckb1: Corsair Gaming K95 RGB PLATINUM Keyboard vKB: left down
ckb1: Corsair Gaming K95 RGB PLATINUM Keyboard vKB: left up
ckb1: Corsair Gaming K95 RGB PLATINUM Keyboard vKB: leftmeta up

My attempt:

capslock = mods_on_hold(M, leftmeta)

results in

keyd virtual keyboard: leftmeta down
keyd virtual keyboard: leftmeta up
keyd virtual keyboard: leftmeta down
keyd virtual keyboard: leftmeta up

I held capslock, pressed once left arrow and released capslock. But arrow is missing and (slow) hold&release of capslock lead to leftmeta down (hold started), nothing (here should be left arrow down and up) and leftmeta up, leftmeta down, leftmeta up (all tree events fired when key was released).

Some keys (mostly letters on first and second row) are not dropped, but meta goes "up and down" around them which isn't good (also there are excessive meta presses). This is from capslock down, a key down and up, capslock up (both keys pressed only once):

keyd virtual keyboard: leftmeta down
keyd virtual keyboard: leftmeta up
keyd virtual keyboard: leftmeta down
keyd virtual keyboard: leftmeta up
keyd virtual keyboard: a down
akeyd virtual keyboard: a up
keyd virtual keyboard: leftmeta down
keyd virtual keyboard: leftmeta up
keyd virtual keyboard: leftmeta down
keyd virtual keyboard: leftmeta up

I also tried simpler capslock = leftmeta, but that didn't work too (one less down+up of meta, but still meta key was mapped as held, was released when other key was pressed and then pressed down again on its own).

It this a bug, or do I have something wrong in the configuration?

On startup, the daemon sometimes does not load the config.
This is a log of when that happened:

-- Boot a83ee4fbd1ad4e1d8866ad2c8c2dedd7 --
ago 29 16:38:45 amos systemd[1]: Started key remapping daemon.
ago 29 16:38:45 amos keyd[365]: Starting keyd v1.1.2 (8df75b749b9131faeb551c84efebca9762047649).

I am using the keyboard from a Thinkpad T480.

I do not know whether this is relevant or not but this is the output of libinput list-devices regarding the keyboard in question.

Device:           AT Translated Set 2 keyboard
Kernel:           /dev/input/event3
Group:            6
Seat:             seat0, default
Capabilities:     keyboard
Tap-to-click:     n/a
Tap-and-drag:     n/a
Tap drag lock:    n/a
Left-handed:      n/a
Nat.scrolling:    n/a
Middle emulation: n/a
Calibration:      n/a
Scroll methods:   none
Click methods:    none
Disable-w-typing: n/a
Accel profiles:   n/a
Rotation:         n/a

(This issue originated from #22)

I'm having problems creating a config file.
My keyboard is listed as "Apple Inc. Apple Internal Keyboard / Trackpad".
Since this keyboard name contains a forward slash, I'm unable to create a config file, since it's not following the UNIX philosophy:

touch Apple\ Inc.\ Apple\ Internal\ Keyboard\ \/\ Trackpad.cfg
touch: cannot touch 'Apple Inc. Apple Internal Keyboard / Trackpad.cfg': No such file or directory

Do you have any ideas supporting different ways of specifying keyboard id-s? Like the path of a device?:
Apple Inc. Apple Internal Keyboard / Trackpad

Putting this here to track hot config reloading as a feature currently under discussion.

https://gist.github.com/rvaiya/5f2ba4b92f9c56dca76a99df6ca609a0#file-keyd_app_demo-py-L54

#51 (comment)

This would enable a parent process to send keyd a signal to reload config, without having to reload the entire daemon / re-grab the keyboard devices.

I'm using the AUR package on my Framework laptop, and whenever the keyd service is running, the touchpad doesn't work at all. After stopping keyd, I can move my cursor again. Am I doing it wrong?

Env

• Arch Linux rolling
• SwayWM
• Framework laptop

pacman -Q keyd-git
keyd-git 1.1.2.r23.c7ee833-1

Config

cat /etc/keyd/AT\ Translated\ Set\ 2\ keyboard.cfg
capslock = overload(C, esc)

Output

sudo keyd
Starting keyd v1.1.2 (c7ee83350dc0064e9f03a780995efce18266ea4d).
Managing AT Translated Set 2 keyboard

...at this point, there is no more output and it's impossible to move the cursor or click any buttons with the touchpad.

When starting keyd the last keyevent gets dropped. This usually has the benign manifestation of sending enter until it is pressed again but in the case of certain input devices coupled with libinput has the additional implication of rendering the mouse unusuable.

It seems libinput stops generating mouse events for a certain class of devices while the enter key is held. This can be observed by holding enter and trying to move the cursor around. The problem is easily avoidable by simply starting keyd before wayland as intended or switching VTs to force a device reload. Prepending a timeout is also a possibility (E.G sleep 1s;sudo systemctl start keyd). A simple fix would be to bake a timeout of a few ms into the code to allow the user to release any keys but this is ugly and involves a race condition. A more elegant solution is in the pipeline.

I see some dups in the list of keys available via keyd -l (e.g. up, down, menu, record):

prevgroup
nextgroup
accept
cancel
up
down
up
down
menu
menu
11
12
desc
mode
favorite
record
record
vod
unmute

Hi,

Some distros like Nixos use non-standard paths, so it would be cool if the paths for the config file, log file and lock file could be customized. Currently I've patched the source code & Makefile but obviously another option would be to allow the user to pass these as arguments to the program.

diff --git a/Makefile b/Makefile
index 8f90462..3563b80 100644
--- a/Makefile
+++ b/Makefile
@@ -3,9 +3,13 @@
 DESTDIR=
 PREFIX=/usr
 
+CONFIG_DIR=/etc/keyd
+LOCK_FILE=/var/lock/keyd.lock
+LOG_FILE=/var/log/keyd.log
+
 VERSION=1.1.2
 GIT_HASH=$(shell git describe --no-match --always --abbrev=40 --dirty)
-CFLAGS=-DVERSION=\"$(VERSION)\" -DGIT_COMMIT_HASH=\"$(GIT_HASH)\"
+CFLAGS=-DVERSION=\"$(VERSION)\" -DGIT_COMMIT_HASH=\"$(GIT_HASH)\" -DLOCK_FILE=\"$(LOCK_FILE)\" -DLOG_FILE=\"$(LOG_FILE)\" -DCONFIG_DIR=\"$(CONFIG_DIR)\"
 
 all:
 	mkdir -p bin
diff --git a/src/config.h b/src/config.h
index a6465b8..3a53043 100644
--- a/src/config.h
+++ b/src/config.h
@@ -29,7 +29,11 @@
 #include "keys.h"
 
 #define MAX_LAYERS 32
+
+#ifndef CONFIG_DIR
 #define CONFIG_DIR "/etc/keyd"
+#endif
+
 #define MAX_LAYER_NAME_LEN 256
 #define MAX_MACRO_SIZE 256
 #define MAX_MACROS 256
diff --git a/src/main.c b/src/main.c
index 86dbc7e..a697170 100644
--- a/src/main.c
+++ b/src/main.c
@@ -42,8 +42,14 @@
 
 #define UINPUT_DEVICE_NAME "keyd virtual keyboard"
 #define MAX_KEYBOARDS 256
+
+#ifndef LOCK_FILE
 #define LOCK_FILE "/var/lock/keyd.lock"
+#endif
+
+#ifndef LOG_FILE
 #define LOG_FILE "/var/log/keyd.log" //Only used when running as a daemon.
+#endif
 
 #ifdef DEBUG
 	#define dbg(fmt, ...) warn("%s:%d: "fmt, __FILE__, __LINE__, ## __VA_ARGS__)

Thanks.

Today I was trying to remember how to enter monitor mode, and kept trying options that didn't work but were silently ignored. What further fooled me is that monitor mode and non-daemon "main loop" mode are similar in that they mention the name of the keyboard and wait.

It would be handy to have a --help option for a quick reminder of the options.

And/or warn on invalid options.

I mistakenly thought I'd encountered an error state via #47. Thank you to the @rvaiya for the help there. I'm opening this new ticket to discuss what I actually want to achieve, so I can be clearer and potentially get better advice :)

I'm trying to make Linux shortcuts behave like Mac OS:

• I have a Win (Meta) key that can be used for application switching, e.g. Meta-Tab and Meta-Shift-Tab
• I have an Alt key to the left of the spacebar where Cmd normally is, so I'm thinking of Alt as the "do everything" modifier, like Cmd is in Mac OS.
• When I press Alt-C I want to simulate the Linux system-wide "Copy" shortcut key: Ctrl-Insert
• When I press Alt-V I want to simulate the Linux system-wide "Paste" shortcut key: Shift-Insert
• When I press and hold Alt, followed by any sequence of Tab or Grave (tilde) I want to simulate Meta-Tab (for each Tab press/release) or Meta-Shift-Tab (for each Grave press/release), until Alt is released. (In this "held" state, after an initial Tab or Grave has been pressed, it's ok--maybe even desirable?--to do nothing when C or V are pressed, even though technically Alt is also being pressed and that could be considered an Alt-C or Alt-V sequence).

How can I achieve this "dual nature" Alt key in keyd?

(Note: the important part here is that once we enter the "app switching" state, the Alt key cannot be released, or else the app switcher will be shut off and lose its state)