02:19 <ifreund> hmm, I think we should just make individual issues for anything that needs doing and add them to an appropriate milestone

12:08 <leon-p> ifreund: The input PR should be done now, I think. I have updated the bash completion script and perhaps the people who contributed the completions for other shells are willing to update those as well.

14:13 <elshize> hi everyone, I tried out river yesterday, and I gotta say I'm really impressed; I love the layout protocol, have been playing with kile layout manager; I realize it's work in progress but so far everything works perfectly. to all contributors: great job so far! very exciting project, and your work is very appreciated!

14:15 <leon-p> elshize: if you want to play around with layouts a bit more, in the contrib directory in rivers repository you'll find a simple layout written in C that is very hackable and has comments explaining all the important bits

14:17 <elshize> leon-p: yep, I saw that, I'll definitely play around with that at some point

17:10 <ifreund> elshize: thanks for the kind words!

17:10 <ifreund> aaay 666 stars

17:10 <ifreund> hail satan

17:24 <leon-p> yeah, nice to see river approved by the metal god :P

17:27 <ifreund> leon-p: one last bit of bikeshedding on the input PR, otherwise everything looks good :)

17:30 <leon-p> ifreund: all done!

17:31 <ifreund> sweet, I'll just lest the CI run and then merge it

17:31 <leon-p> is our CI un-broken again?

17:32 <leon-p> I have gotten used to just ignoring it :P

17:32 <ifreund> indeed

17:32 <leon-p> nice

17:32 <ifreund> I should probably make it either download a zig tarball or build zig from source

17:33 <ifreund> we should be fine until zig 0.9.0 releases though

17:33 <ifreund> well, wlroots might break it again for a bit

17:33 <leon-p> I don't think there is a wlroots release planned right now, so we are fine for a while

17:36 <leon-p> I really enjoyed working on this. Using @"" and @tagName to have a single source of truth instead of some sort of enumToString function is absolutely amazing

17:36 <ifreund> yeah, it's that kind of thing that I now miss in every other language

18:07 <elshize> what problems do you have with CI? is it because github keeps updating images? I started using docker images within Actions in some of my projects. our C++ builds basically kept breaking with each PR, no kidding

18:22 <ifreund> elshize: CI problems are purely due to my decision to build river using zig and wlroots with are both pre-1.0 and have breaking changes with every release

18:22 <ifreund> so the distro I use for the CI lags behind the upstream releases we track

18:23 <ifreund> building both zig and wlroots from source would make things a bit more reliable

18:27 <elshize> ifreund: got you. out of curiosity, what made you choose zig? is there something behind this decision, or do you just like the language?

18:27 <elshize> it's pretty niche, isn't it?

18:31 <ifreund> elshize: yeah it's pretty niche, but it's the only language I found that made me want to try using it instead of C

18:31 <ifreund> river is my first zig project for reference

18:31 <ifreund> since starting river I've gotten involved in working on the zig compiler though

18:32 <ifreund> basically Zig offers significant upgrades in safety and power over C while remaining simple

22:41 <elshize> ifreund: I see. have you tried rust? not that i'm suggesting switching, just curious :) I haven't really tried zig myself, I'm curious to see what it offers, might give it a spin one day.

22:43 <leon-p> elshize: rust and Wayland aren't exactly the greatest match. Wayland is dangling pointers everywhere. You can design abstractions to get around that, but it gets awkward.

22:56 <elshize> leon-p: whoa, "dangling pointers everywhere" sounds like a nightmare :D but I hear you. I can imagine it being tricky, though I haven't done anything wayland, so no experience with it.

22:59 <leon-p> realistically "dangling pointers everywhere" is just what you get when you do systems programming

23:00 <leon-p> even with rust, if you go deep enough everything is unsafe

23:00 <ifreund> I've got some ideas of how to make wayland code less pointer-soupy

23:00 <ifreund> these ideas would not be very wlroots/libwayland compatible though

23:01 <ifreund> elshize: RE rust: https://way-cooler.org/blog/2019/04/29/rewriting-way-cooler-in-c.html

23:01 <ifreund> and if you want to see me ramble on this topic: https://www.youtube.com/watch?v=mwrA5IRGpfU

23:03 <ifreund> rewriting *everything* in rust is theoritically possible, but that means reimplementing all of wlroots for (IMO) very little gain

23:03 <ifreund> also I just don't like rust very much any more, I find it too complex for the benefit it provides, especially in single threaded programs

23:21 <leon-p> ifreund: speaking of pointer soup, maybe we should make debug builds use zigs gpa to make it easier to catch leaks

23:36 <elshize> ifreund: I'm familiar with way cooler story, though never looked into details. I very much like Rust, but I do agree that just simply rewriting everything in Rust is not necessarily the way to go. because Rust approach is so much different, it doesn't necessarily work well with some existing code or APIs, and rewriting everything from ground up is usually a way to nowhere

23:46 <leon-p> huh, river does not like zigs gpa

23:49 <waleee-cl> wouldn't wlroots need c's allocator?

23:51 <leon-p> yeah, although I think wlroots stuff cleans itself up