<leon-p>
still, a proper service manager is probably the better idea
nemoguest90 has quit [Ping timeout: 256 seconds]
leopoldek has joined #river
alexherbo2 has quit [Remote host closed the connection]
aryak has joined #river
<Momentum>
Interesting! thanks nemoguest and leon-p
<Momentum>
leon-p: I'm curious how would a proper service manager would solve this problem?
<leon-p>
basically you move the responsibility of starting programs to it and out of your init
<leon-p>
this means you can rerun the init without having duplicate instances of things
<leon-p>
and depending on which one you use, you may also get logging and restarting for free
Palanix has quit [Ping timeout: 260 seconds]
<leon-p>
f.e. I use systemd --user services for some things, like wlr-autorotate (which keeps track of my laptops gravity sensor and rotates the screen accordingly). Recently that one was a bit flaky and crashed a lot. systemd tried restarting it three times before giving up and I got an extensive log. If I had just `riverctl spawn`'d it, then I likely would have not discovered it as quickly
Palanix has joined #river
nemoguest has joined #river
<Momentum>
That's very interesting
<Momentum>
and yeah. that makes sense
<leon-p>
there are also dedicated service managers for user sessions, should you run a non-systemd setup
<nemoguest>
leon-p I guess it depends. I prefer my nails with a hammer rather than a drill :) Logging with river &> /tmp/river.log seems pretty sufficient to me.
<leon-p>
screws are more robust than nails :)
<nemoguest>
Good point, though I only need them for a wall painting. Is there a reason why you would prefer 'riverctl spawn <app>' to '<app> &'? Not familiar with the process
<leon-p>
`riverctl spawn foo` and `foo &` in your init should act identically. the & puts them into the shells job control. Your init likely exits or exec's something, so those processes will be adopted by init. riverctl spawn results in a double-fork exec, which also ends up at init
<leon-p>
whoops, I used init to describe both the init script and the init program
<leon-p>
should be readable still though, I hope
<leon-p>
note that exec-ing something in your river init is different thoug since IIRC the init script process is a direct child of river
<leon-p>
but realistically these things are not something you need to think about IMO
<nemoguest>
Make sense, thank you. Though`riverctl spawn`always caused me a little trouble and seemed a little less versatile.
<leon-p>
the most common issue people have with it is that you need to quote the entire command, i.e. `riverctl spawn "foo --flag --another-flag"`
<Momentum>
yeah that can get messy
sm222 has joined #river
sm222 has quit [Remote host closed the connection]
sm222 has joined #river
<Momentum>
so stdout/stderr from the spawn'd processes would go with river's stdout/stderr? or that's stupid to assume?
<Momentum>
i just want to know if i should still redirect errors for spawn'd process
nemoguest has quit [Quit: Client closed]
nemoguest has joined #river
<nemoguest>
Momentum yes, they would.
<leon-p>
it's actually quite annoying to decouple a spawned process's stderr/in/out from its parent, even if double-forked. and river doesn't do that
<leon-p>
well, obscure is prob. a better word than annoying once you know how
<nemoguest>
leon-p hmm, I guess you could try something like `try () (pidof -q $1 || $@ > /tmp/$1.log 2>&1)`?
<nemoguest>
and then `try waybar &` for example
<leon-p>
my brain was more in C mode when i wrote that that in shell mode
<nemoguest>
I don't know C so I guess that's my superpower :)
<nemoguest>
but if you prefer it that way, you could:
<nemoguest>
function try
<nemoguest>
{
<nemoguest>
pidof -q $1 || $@ > /tmp/$1.log 2>&1
<nemoguest>
}
sm222 has quit [Remote host closed the connection]
sm222 has joined #river
sm222 has quit [Ping timeout: 252 seconds]
sm222 has joined #river
angry_vincent has quit [Ping timeout: 246 seconds]
sm222 has quit [Ping timeout: 246 seconds]
sm222 has joined #river
sm222 has quit [Read error: Connection reset by peer]
sm222 has joined #river
sm222 has quit [Ping timeout: 252 seconds]
sm222 has joined #river
Guest19 has joined #river
Guest19 has quit [Client Quit]
nemoguest has quit [Ping timeout: 256 seconds]
talismanick has joined #river
rrix has quit [Remote host closed the connection]
rrix has joined #river
rrix has quit [Remote host closed the connection]
rrix has joined #river
sm222 has quit [Ping timeout: 246 seconds]
Palanix has quit [Ping timeout: 252 seconds]
Palanix has joined #river
sm222 has joined #river
sm222 has quit [Ping timeout: 272 seconds]
Guest53 has joined #river
Guest53 is now known as hi
<hi>
hi guys none of my keybinds work
<hi>
it says that xkeyboard keymap compiler (xkbcomp) reports: warning unsupported maximum keycode 708, clipping. x11 cannot support keycodes above 255
<hi>
and then this twice: warning: could not resolve keysym xf86kbdinputassistprevgrou