ifreund changed the topic of #river to: river - a dynamic tiling wayland compositor - https://github.com/ifreund/river - channel logs: https://freenode.irclog.whitequark.org/river/
<daurnimator> leon-p: session selection is cog at bottom right after you select a user.
maringuu has quit [Ping timeout: 250 seconds]
maringuu has joined #river
<daurnimator> ifreund: for some reason I can't start river any more
<daurnimator> from inside of X: http://sprunge.us/JeLMdH
<daurnimator> from a tty: http://sprunge.us/9Klt02
waleee-cl has quit [Quit: Connection closed for inactivity]
leon-p has joined #river
<leon-p> daurnimator: yeah I know where the changing button /should/ be, but when I installed gdm from the repos and run it it wasn't there. I am not going to try gdm again though. Might try the kde display manager, but not today.
leon-p has quit [Quit: leaving]
<novakane[m]> in frame: wl.Listener(*wlr.Output) = wl.Listener(*wlr.Output).init(handleFrame),
<novakane[m]> i'd like to return error from handleFrame function
<novakane[m]> I add handleFrame() !void {}
<novakane[m]> but how can I make it works in the init(handleFrame)
<novakane[m]> I'm not sure I'm very clear
<novakane[m]> probably better https://zigbin.io/7d8b09
leon-p has joined #river
<leon-p> novakane[m]: not sure returning an error from that function makes much sense, considering it is called by wlroots, not by your own code.
<leon-p> if you encounter an error, just handle it in the function and abort
<novakane[m]> leon-p: yeah ok, it was mainly to see if everything works as intended step by step to debug, and probably remove it later
<leon-p> so you want an error trace if something fails?
<leon-p> in that case, if the function encounters an error, let it print the error trace and then call std.os.exit(1) or something like that
<novakane[m]> yep that's what I want
<novakane[m]> thanks I'll try that
<leon-p> if you want to copy the error trace printer, you'll want to look at fn _start in root.zig in std. Not sure where exactly, but that is where the logic to print the error trace should live IIRC
leon-p has quit [Quit: leaving]
<novakane[m]> leon-p: hmm where root.zig should be? I can only find _start here https://github.com/ziglang/zig/blob/0.7.x/lib/std/start.zig
<ifreund> start.zig is correct
<ifreund> daurnimator: I assume this is after updating wlroots to 0.13.0?
<ifreund> that's the only recent change that could have anything to do with that afaik
<novakane[m]> ifreund: thanks
yyp has joined #river
<daurnimator> ifreund: possible/likely
<novakane[m]> could be related to this https://github.com/swaywm/wlroots/pull/2834
<yyp> https://github.com/ifreund/river/pull/112#discussion_r615375314 - do I not understand Wayland errors? I thought they are just a different response to a request indicating that something went wrong, do they do anything more?
<ifreund> yes, they disconnect the client
<yyp> Ok, makes sense. But do we really need to keep the layout client connected when it's namespace is taken?
<yyp> s/it's/its
<ifreund> we need to allow clients to try again with a different namespace or report an error as they see fit
<ifreund> wayland errors should only be used for protocol violations
<ifreund> and protocol violations should never happen in a bug free client/compositor
<yyp> yeah
<novakane[m]> launching river in river and having a passthrough mode is really the best thing ever :p
<novakane[m]> really make the life easier to debug
<ifreund> for sure, it was one of the first things I implemented :D
<ifreund> pretty sure that came before riverctl existed
<novakane[m]> haha really? Well that was a great idea!
<ifreund> I mean, I had a similar passthrough mode in sway which is how I tested river before I started dogfooding so I can't really take credit for the idea :D
<novakane[m]> oh there is one in sway? or you added it?
<yyp> novakane[m]: yes, and it was there for a very long time in the default config
<novakane[m]> yyp: oh ok I didn't know, didn't really used sway a lot
c7s has joined #river
leon-p has joined #river
<ifreund> Hmm, I kinda want to make the rivertile layout namespaces `rivertile_*` instead of `tile-*`
<leon-p> sure
<leon-p> related: What about directional options? Still planned or not?
<ifreund> We don't have a use-case yet do we?
<leon-p> other than rivertile only having a single layout, no
<ifreund> that would be nicer tbh
<leon-p> we could also just do it with a string though
<ifreund> yeah, I think I'll do it with a string for now
<ifreund> `main_location`?
<leon-p> fine with me
<ifreund> I think I'm down to the last bug here
<leon-p> nice!
<ifreund> I love the fact that std.meta.stringToEnum() exists in a systems programming language
<leon-p> I am overall quite impressed by the std lib. Not quite a kitchen sink, but enough parts for me to build my own without much effort
<ifreund> parts of it are excellent, and other parts need a lot of work
<ifreund> it's pretty good considering that it has mostly grown organically according to immediate needs and has yet to see focused attention
<ifreund> definitely will benefit from some cleanup before 1.0
<ifreund> I was more commenting on how sweet comptime is tbh :D
<leon-p> yeah, std.meta is magic :D
<leon-p> but the good kind of magic, not the "let's use glib to make c niver" kind of magic
<yyp> ^^
<yyp> Kinda reminds me of Go's reflect package :)
leon-p has quit [Quit: leaving]
exception has quit [Quit: ZNC 1.8.2 - https://znc.in]
exception has joined #river
leon-p has joined #river
<ifreund> river-options-v2 is complete, taking a short break then I'll rebase it and push to your branch leon-p
<novakane[m]> nice!
<novakane[m]> so we are near the merging of river layout branch to master then ?
<ifreund> yup :)
<novakane[m]> cool :D
exception has quit [Quit: ZNC 1.8.2 - https://znc.in]
exception has joined #river
waleee-cl has joined #river
leon-p has quit [Quit: leaving]
yyp has quit [Quit: now it's safe to turn off your computer]
<Bonicgamer[m]> Im unable to launch river or sway anymore. What fun.
<Bonicgamer[m]> Nvm fixed it
c7s has quit [Ping timeout: 260 seconds]