Logs on 2022-01-07 (liberachat/#kmonad)
| 02:31:26 | → | neurocyte0917 joins (~neurocyte@IP-046243081042.dynamic.medianet-world.de) |
| 02:31:26 | × | neurocyte0917 quits (~neurocyte@IP-046243081042.dynamic.medianet-world.de) (Changing host) |
| 02:31:26 | → | neurocyte0917 joins (~neurocyte@user/neurocyte) |
| 02:33:29 | × | neurocyte091 quits (~neurocyte@user/neurocyte) (Ping timeout: 256 seconds) |
| 02:33:29 | neurocyte0917 | is now known as neurocyte091 |
| 09:48:47 | × | neurocyte091 quits (~neurocyte@user/neurocyte) (Quit: The Lounge - https://thelounge.chat) |
| 09:49:59 | → | neurocyte0917 joins (~neurocyte@IP-046243081042.dynamic.medianet-world.de) |
| 09:49:59 | × | neurocyte0917 quits (~neurocyte@IP-046243081042.dynamic.medianet-world.de) (Changing host) |
| 09:49:59 | → | neurocyte0917 joins (~neurocyte@user/neurocyte) |
| 15:46:36 | → | kraM joins (~kraM@user/kram) |
| 15:50:16 | <kraM> | Hey guys. I have a working setup on node a which I want to use on node b as well. https://paste.xinu.at/HfONk/ xcape like: Have CAPS used as ESC on short keypress, as META when holding it down. |
| 15:50:17 | → | GavinFreeborn[m] joins (~gavinokma@2001:470:69fc:105::cb9) |
| 15:50:17 | → | contessa joins (~contessa@2001:470:69fc:105::a9f) |
| 15:50:17 | → | Tisoxin joins (~ikosit@user/ikosit) |
| 15:50:17 | → | meain joins (~meain@2001:470:69fc:105::2db) |
| 15:50:57 | <kraM> | On node b activating this configuration results in the Enter key doing nothing anymore. I have no idea what I'm doing and what this configuration exactly tells me. |
| 15:51:15 | <kraM> | (Did this once with some guy who understands kmonad.) |
| 15:51:39 | <kraM> | Anything reasonable in what I said and any idea where I might want to learn more? |
| 15:53:26 | <kraM> | Ah. Hm. I fear the input device I'm targeting on node b isn't exactly the keyboard only. It seems to disable the touchpad completely as wall. |
| 16:07:25 | <kraM> | Okay. I guess I understand what happens, but I have no idea how to address this. From kernel point of view I only have /dev/input/by-path/platform-i8042-serio-0-event-kbd only. |
| 16:11:10 | <kraM> | Hm. Nope. cat $device doesn't do anything when moving the mouse via touchpad. |
| 16:16:12 | <kraM> | So, device-file seems to be correct, kmonad configuration looks legit (as far as I understand), but touchpad and enter key on internal keyboard stop working. Any hint greatly appreciated! |
| 16:29:36 | <neurocyte0917> | kraM, maybe evtest can tell you more about what is going on? |
| 16:33:28 | <kraM> | sudo evtest -> choose input 4 -> press Enter -> all fine. Start kmonad, press Enter -> new blank line in terminal, EV_MSC or EV_KEY output. |
| 16:33:55 | <kraM> | Ah. Sure. Um. kmonad creates a virtual keyboard, right? |
| 16:34:42 | <kraM> | Right. Now I'm completely confused. |
| 16:37:29 | <kraM> | Reason for my confusion: sudo systemctl start kmonad -> broken. Running as unprivileged user -> everything works as expected. |
| 16:37:57 | <kraM> | A wayland thing, maybe? |
| 16:38:47 | <kraM> | Like, root/sudo doesn't know about the wayland session, does something, and sway doesn't recognize any new input devices (or something like that)? |
| 16:39:49 | <kraM> | (Node a is a desktop with xorg and everything feels simpler. Node b is a notebook with wayland/sway.) |
| 16:44:47 | <neurocyte0917> | I use kmonad with wayland all the time. It should work fine. |
| 16:46:04 | <neurocyte0917> | and evtest with the kmonad virtual device doesn't produce any events at all? |
| 16:50:19 | <kraM> | kmonad device gives me, e.g.: type 1 (EV_KEY), code 28 (KEY_ENTER), value 1 |
| 16:51:28 | <kraM> | (Need to do that via ssh right now. Currently I don't have any Enter key functionality again, whatever user I start the service with.) |
| 16:51:57 | <kraM> | ESC/LMET an CAPS is working fine though. |
| 16:52:10 | kraM | is confused. |
| 16:52:48 | <Solid> | that sounds weird indeed |
| 16:54:55 | <kraM> | So to sum up: I guess there's nothing horrible wrong with my kmonad keyboard definition, right? |
| 16:55:03 | <Solid> | not that I can see |
| 16:56:30 | <kraM> | That's good. Checkmark. |
| 16:58:35 | <kraM> | kmonad seems to do everything correctly. It grabs on the physical keyboard, creates a virtual one, and evtest recognises keypress events on virtual keyboard correctly. |
| 16:58:38 | <kraM> | Checkmark. |
| 16:59:52 | <kraM> | In sway I cannot launch terminals (LMET+Enter) and in running terminals (alacritty) pressing Enter does nothing after starting kmonad. I'll test some gui application. |
| 17:02:36 | <kraM> | No Enter functionality in wayland native gui application. No Enter functionality in Xorg (xlsclients) application. |
| 17:04:25 | <kraM> | For idiots: Should I disable `input "*" { xkb_options "caps:escape" }` in sway? |
| 17:05:04 | Solid | has no idea about Wayland as he will cling to X11 for at least another 10 years |
| 17:07:36 | <kraM> | As far as I understand these xkb_options shouldn't bother kmonad. |
| 17:07:51 | <Solid> | they should not |
| 17:08:07 | <Solid> | kmonad reads directly from uinput, long before any environment has a chance to change the keys |
| 17:08:30 | <kraM> | Thought so. |
| 17:11:21 | <kraM> | Okay. Thank you guys for rubberducking. I'll give up for today. I still appreciate any further hint. |
| 17:20:18 | <neurocyte0917> | kraM, now that I think about it, I remember having similar problems with kmonad and wayland to you. My "fix" was to start kmonad before wayland (via systemd on startup) and since then I've had no problems. |
| 17:20:40 | <kraM> | I'll give that a try. |
| 17:22:18 | <kraM> | Works. |
| 17:22:59 | <kraM> | When it's not the simple things to fix an issue, it's the more simple things. |
| 17:23:39 | <kraM> | such happy much wow very thanks |
| 17:23:48 | Solid | adds that to his bucket list of reasons why wayland is not ready yet :> |
| 17:27:36 | <kraM> | A lot of things feel quite good nowadays with wayland, assumed not to use nvidia proprietary drivers. |
| 17:27:53 | <kraM> | Especially if you come from i3 and go to sway. |
| 17:29:41 | <neurocyte0917> | The nvidia drivers are catching up quickly in the last few months. I have a nvidia desktop running wayland and most things work just fine. Not everything though. |
| 17:30:03 | <Solid> | kraM: I do not |
| 17:30:22 | <Solid> | I suspect that to feel at home on Wayland we would need to port my current wm to it |
| 17:30:36 | <Solid> | (and we will, at one point) |
| 17:30:59 | <kraM> | I'll stick to Xorg on my desktop with nvidia. Desktop/wm is one thing, games, ... Before I did that, I'd buy an AMD GPU. |
| 17:31:00 | <Solid> | but I still object to some fundamental architectural decisions so... it'll probably never feel quite right :/ |
| 17:31:18 | <kraM> | What wm are you using, Solid? |
| 17:31:31 | <Solid> | kraM: xmonad (of course!) |
| 17:31:56 | <neurocyte0917> | sway is a bigger problem than wayland for me. i3 just works. sway has some catching up to do. |
| 17:32:31 | <neurocyte0917> | I used xmonad for many years. I switched to i3 at some point, but I don't really remember why. |
| 17:32:36 | <kraM> | Hm. Ja. I'm too stupid for xmonad, herbstluftwm is missing a feature I'm used to too much (and hlwm is xlib only), so i3/sway it is for me. |
| 17:33:37 | <Solid> | no one's too stupid for xmonad---the website even says "that was easy!" :P |
| 17:35:05 | <kraM> | neurocyte0917: Did you say you run wayland with i3? |
| 17:35:37 | <Solid> | I think they meant they switched xmonad -> i3 on X11 and then to sway for wayland |
| 17:36:18 | <kraM> | xcb is X only, right? |
| 17:36:24 | <Solid> | yes |
| 17:38:09 | <kraM> | That would've grinded my gears, if switching to sway wasn't necessary. |
| 17:44:17 | <drdo> | I run wayland and kmonad works well. But I have had very strange issues when working with evdev directly myself |
| 17:44:53 | <drdo> | Even though the generated events were exactly the same I sometimes got strange results in terminals depending on how fast I emitted the events |
| 17:45:02 | <drdo> | I never figured it out |
All times are in UTC on 2022-01-07.