Something changed recently(?) that #Emacs no longer recognizes the compose key, but other apps do. To be fair, how I've bound it between xmodmap and Emacs is overloaded.
xmodmap: plain capslock is mapped to super, capslock+shift is mapped to compose
Emacs: super+shift+(f/b/n/p) moves windows around
So if I'm hitting capslock+shift, it's ambiguous if that's compose or if it's super+shift+…
But it worked until... recently? And it still works in #EXWM windows, just not regular Emacs windows.
@tfb My fallback WM, when for some reason exwm doesn't work for me, is StumpWM: Common LISP based so sympathetic to Emacs.
I finally have the perfect social media setup for my wide monitors with exwm as the window manager!
Petite question pour les utilisateurs d’#emacs et d’#android ici, qu’est-ce que vous utilisez comme gestionnaire de mots de passe avec un client disponible pour les deux et auto-hebergeable?
J’utilisais #passwordstore qui me convenait parfaitement mais le client android n’est plus maintenu :(
Pour l’instant je teste #vaultwarden mais je trouve le client bureau de #bitwarden pas super pratique à utiliser, c’est trop pensé pour une navigation à la souris et ça rend l’utilisation beaucoup moins fluide que ce que j’ai actuellement.
Password-store encore une fois a un super paquet pour emacs qui me permet en un raccourci clavier de lancer le client, rechercher mon mot de passe, le copier et à nouveau avec le premier raccourci de rebasculer sur l’appli ou j’étais avant (j’utilise #exwm comme environnement de bureau)
Je n’ai pas vu de client natif emacs pour bitwarden, la meilleure piste que je vois c’est le paquet #emacs-bitwarden (https://github.com/seanfarley/emacs-bitwarden?tab=readme-ov-file) qui est en fait un wrapper pour le client en ligne de commande bw et qui permet d’intégrer bitwarden comme auth-source en mode lecture seule.
Ceci-dit les vraies fonctionalités de partage de mots de passe et d’organisation de vaultwarden sont plutôt attirantes.
Avant de me lancer à fond sur vaultwarden je voudrais être sûr qu’il n’y a pas d’autre solution évidente que j’aurais zappée…
@amy I have a bunch of #exwm simulation keys too but never thought to do C-x h, good idea!
For M-f/M-b I map them to C-left and C-right. I also have C-d to delete, C-k to S-end and delete, C-y to C-v, M-w to C-c and a few others.
One challenge is to make C-s work similarly to isearch so that the same key can both initiate the search and search again. I try to do that by sending both C-f and C-g but it's not always perfect.
A couple weeks ago, I posted about my guix setup in-progress, and I think it's finally stabilized a little and the last few things I thought were necessary for a desktop environment were implemented, like a good screen-lock/display power/sleep/hibernate setup, switching from docker to podman and podman-compose, setting up a java language server, and getting all of my emacs config and packages guix-ified.
Oh, and I added two new exwm key translations: C-x h being mapped to C-a, and M-f/M-b being mapped to just four arrow keys in either direction. Really what I use M-f/M-b for is just a faster C-f/C-b, and hopping four letters in each direction is good enough for that purpose. My favorite key translation I've thought of so far is C-k, which is mapped to S-<end> C-x, and works in every application I've tried.
Anyways, I feel a little bad, because @deedend@mastodon.social asked for a screenshot, and I promised to send some of my config upstream to guix, and then proceeded to have a lot of life events happen at me. I am still planning on submitting some merge requests a little later, but I did finally remember to make screenshots, so this is how my system looks now, built from this source code. I've also added pictures of my grub screen, login screen, and screenlock. #guix #emacs #exwm
Think of switching to EXWM and I have 3 questions:
* Is EXWM theme and look as customizable as Sway or i3 ?
* How do you config the sytem (network, printers, etc) from EXWM ?
* Is the performance good when displaying a video game window ?
Deseando escuchar los scripts de #ChucK en el entorno multimedia #EMMS pw-jack no encontraba chuck. Como uso #EXWM intenté usar exec-path pero es interno de #Emacs, aprendí que PATH se hereda a subprocesos. Solución: setenv PATH.
Más detalles en: https://github.com/son0p/folletos/blob/main/social.org#escuchar-scripts-de-chuck-con-emms-en-guixpipewire
Alight, I just saw a post from @ericsfraga and now I want to try #exwm.
Ready for today's session at EmacsConf: mpv running, irc (via erc, of course) channels open, browser window with links to discussions for the talks, tracking the fediverse (with mastodon.el), all within exwm. Cannot be more Emacs than this. 😉
Looking forward to it.
I've installed #GuixSystem this morning and I've discovered #emacs #exwm
So I've installed it.
I constantly use Consult to switch buffers. Consult is very helpful in showing you a preview of other buffers as you search for the particular one you want to switch to.
However, previewing EXWM buffers (i.e. X apps) has the side effect of removing the X app from its original window and messing up the way you've organised your buffers across workspaces.
This doesn't affect normal buffers (of which you can have multiple instances), only EXWM buffers.
Solution follows. 🧵
In a previous toot thread we've been talking about ways to launch processes from Emacs in a way that lets them survive Emacs exiting, crashing or restarting via M-x restart-emacs.
The best way I've found so far:
;;; with separate cmd and args:
(call-process command nil 0 args)
;;; or let the shell parse the command:
(call-process-shell-command command nil 0)
The key is the 0 as the third argument, this causes Emacs to discard the output and not wait for the process.
I use it like this for my EXWM command launcher function:
(defun my-exwm-run (command)
(interactive (list (read-shell-command "$ ")))
(call-process-shell-command command nil 0))
Previously I used start-process-shell-command and nohup but that accumulated defunct processes when using M-x restart-emacs. The call-process-shell-command with the 0 arg doesn't seem to have that problem and it doesn't need nohup. This is a bit counter intuitive because call-process is supposed to be for synchronous processes. Not sure if make-process (the async primitive) can do this the same way.
If somebody is familiar with how/why exactly this works I'd love to hear your thoughts.