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/
Dedguy21 has joined #river
Dedguy21 has quit [Quit: Connection closed]
Dedguy21 has joined #river
Dedguy21 has quit [Quit: Connection closed]
waleee-cl has quit [Quit: Connection closed for inactivity]
leon-p has joined #river
Dedguy21 has joined #river
Dedguy21 has quit [Quit: Connection closed]
maringuu has quit [Read error: Connection reset by peer]
maringuu has joined #river
<leon-p> currently working on zig-shellcomplete, I'll probably have a proof of concept ready this weekend
<leon-p> writing a parser in zig is quite pleasant, especially if you already have a tokenizer :D
<novakane[m]> It's sounds nice, really a great idea to use a parser for this
<leon-p> Yeah, but I will probably leave writing the completion tree to completion script generators for non-bash shells to other people
<novakane[m]> Well I'm not sure I want to touch zsh completion system again so I leave it to other people too :p
<novakane[m]> and I think I'm gonna already spend all my time on damage tracking at the pace it's going :p
<leon-p> I hope you managed to introduce both flickering and freezing :P
<novakane[m]> oh I resolve this whith the good old technique of can't compile it :p
<leon-p> progress!
<novakane[m]> yeah!
<novakane[m]> I pushed it to my fork today, really messy with all the test I tried and doesn't compile but hey it's a begining :p
yyp has joined #river
yyp has quit [Quit: now it's safe to turn off your computer]
<leon-p> am a bit miffed that `if (token == .String) |str| {}` does not work though...
<ifreund> leon-p: gotta use a switch...
<novakane[m]> is it good if I only render the background color and nothing else after? :P
<novakane[m]> I made river a screen locker that you can't unlock :D
<leon-p> and my parser works! It just segfaults when i try to deinit the arena
<novakane[m]> segfault is not ideal though :p
<leon-p> it's progress. It means that the rest more or less works. After a few frustrating rounds of playing "is my parsing code wrong or the input json?" a segfault is refreshing :)
<leon-p> and it was just a double free because I copy-pasted without being careful
<novakane[m]> yeah you were efficient
<leon-p> and now comes the actually painful part: generating the completion scripts
<leon-p> I wonder how well my abstraction maps to bash :D
<novakane[m]> you'll probably have some nice surprise with bash :p
waleee-cl has joined #river
<leon-p> writing code that writes code brings back not so fond memories of c macros…
<ifreund> think of it this way: you're basically writing a comipler
<ifreund> *compiler
<ifreund> for a DSL you just invented :D
<leon-p> that does sound better. Have you though about a job in marketing?
<leon-p> it is somewhat horrifying, but it actually works D:
<leon-p> options with modifiers (--option modifier) work, now only tools and conflict groups are missing
<ifreund> nice :)
yyp has joined #river
josip has quit [Remote host closed the connection]
josip has joined #river
<leon-p> I am beginning to think that using a parser was the wrong approach. It works, but making even minor changes is annoying.
<leon-p> it's probably better to just have a struct and functions to add command options
<ifreund> that would certainly be faster to iterate on
yyp has quit [Quit: now it's safe to turn off your computer]
leon-p has quit [Quit: leaving]