<leon-p>
ifreund: other than the issue of users not being able to know the input identifiers in a good way, are there any other blockers for the input PR?
<ifreund>
no, not that I'm aware of
<ifreund>
what do you think of `riverctl list-inputs` or similar?
<leon-p>
I see the need for it, but I don't like it.
<leon-p>
Because it means river formatting the output and sending it over Wayland for riverctl to display
<ifreund>
what if riverctl linked libinput?
<leon-p>
Does libinput expose the vendor id and product id? If yes, that would be ideal. I'll look into it
<ifreund>
Don't know for sure if libinput does, but I was assuming that's where wlroots is getting it
<ifreund>
hmm, I think riverctl might need to be run as root to print this list though if we took this approach
<leon-p>
libinput list-devices does not need root, does it?
<ifreund>
oh, maybe not then
<ifreund>
for some reason I thought it id
<leon-p>
Maybe it does and it uses polkit to not require a password or something. I'll look into it.
sjtio has quit [Remote host closed the connection]
priner[m] has quit [Remote host closed the connection]
sjtio has joined #river
priner[m] has joined #river
novakane has joined #river
<novakane>
leon-p: libinput list-devices need root for me
<novakane>
it read /dev/input/*
yyp has joined #river
<leon-p>
novakane: /dev/input/* does not require root on my system (Parabola).
<novakane>
hmm I needed it on arch and need it on void
<leon-p>
novakane: is your user in the plugdev group?
<novakane>
nope I don't think I even have this group on my system
<leon-p>
novakane: all devices in /dev/input should be of the plugdev group. If you add your user to that group you can access them without root.
<leon-p>
(unless you did something weird with udev)
<novakane>
hmm yeah I don't have that group
<leon-p>
weird. That means we can't go that route since it is apparently not standard to be able to access /dev/input without root.
<novakane>
leon-p: yeah I search about plugdev and it apparently not used on all distro
<leon-p>
that leaves us with two options: 1) let river format the output of the riverctl command, or 2) change the way identifiers are build using only information available using the libinput tool.
<leon-p>
2) is probably not possible, since the identifiers would not be unique enough
<novakane>
yep so only one option
<novakane>
I understand why you don't like it, but I don't see any other way :/
<novakane>
and if there is not an easy way to get the devices indentifiers, I bet a lot of issues will be open about that
<leon-p>
ifreund: am I missing something or is `out` never freed for commands that return successfully?
<leon-p>
oh, yes I am missing something :P
<leon-p>
added `riverctl list-inputs`
<leon-p>
ifreund: ^
<leon-p>
novakane: ^
<leon-p>
it just displays all input device identifiers, their type and whether they have been configured
<novakane>
all right I relaunch river to test it
novakane has quit [Quit: WeeChat 3.1]
novakane has joined #river
<novakane>
leon-p: works good for me
<novakane>
looks good to me, maybe not the perfect solution, but it's a good one
<leon-p>
good enough for 0.1.0, but before 1.0 I'd like to do it differently
<novakane>
sure, I don't see how though
<leon-p>
by having an input config protocol
<novakane>
oh right of course
<novakane>
you'll need to have it validated by wayland first, then wlroots right? A long way
novakane has quit [Read error: Connection reset by peer]