00:15 <ifreund> leon-p: it should work in theory, I dont know if hiding windows works at all yet

01:17 <leon-p> ok, (apply foo a b '(c d e)) actually does map to (foo a b c d e), this makes a lot of things a lot easier

02:50 <leon-p> "solved" the rectangle issue by iteratively splitting an area (starting with the output area) into multiple areas, checking how many window center coordinates are inside each sub-area, and repeating with the least occupied one until either a depth N is reached or all sub-areas are empty

02:51 <leon-p> An improvement would probably be to include the window corner points in the occupancy count as well, but that's for tomorrow to think about

04:52 <leon-p> unfortunately the window dimensions event does not get send for windows that are not shown, so that trick doesn't work yet :(

04:57 <ifreund> I can take a look tomorrow

04:59 <Nickli> Maña

04:59 <Nickli> +na

05:00 <ifreund> leon-p: its kinda funny that our sleep schedules seem to be pretty much in sync despite my current time zone being 7 hours behind yours :D

05:00 <ifreund> anyhow, good night :)

05:00 <leon-p> don't call me out like that

05:00 <leon-p> night o7

05:01 <Nickli> do tell more :P

06:44 <wbr> /

07:23 <TheAnachron> I am thinking about writing a mode for river that will let me spawn new terminals with ssh sessions to my dominant devices, so I dont need to ssh "manually" so often. :D

09:01 <larstiq> leon-p: I did say silly. More seriously, with the layout management protocol I had to do hacky things to alter geometry for identified windows, I might have another go at toggling windows to a monocole layout/hiding them completely without otherwise changing the base layout

09:01 <larstiq> I very often have ~4-5 windows one one monitor and I just for a couple minutes want to focus on one, or not see one

09:06 <Momentum> good morning

09:06 <Momentum> anyone had problems with colors in chromium?

09:07 <TheAnachron> Momentum: can you specify what problems exactly?

09:07 <Momentum> TheAnachron: some colors are basically inverted like red is blue and blue is red

11:28 <leon-p> that's odd

11:29 <leon-p> what's your graphics hardware?

12:15 <Momentum> leon-p: intel hd and it's not being used

14:51 <bwbuhse> I have a tag I use as a scratchpad. I have a keybind to toggle that tag on/off and another to move windows to it. Is there an easy way to move windows out (besides just sending them to whatever tag)? i.e. something like set the tags on a window to all the active tags (except the scratchpad tag)? I'm thinking no until rwm protocol but just wondering

17:35 <ifreund> leon-p: your hide/propose_dimensions(0,0) trick should work now

17:38 <ifreund> bwbuhse: I don't think there is any way to have an acually ergonomic scratchpad with river currently

17:38 <ifreund> there are lots of half solutions floating around, but iirc the focus behavior of river makes them pretty annoying to use

17:39 <ifreund> scratchpads are one of those things I don't personally use or care about

17:40 <ifreund> the rwm protocol is my way of making such things possible for others to implement without needing to review/maintain the code myself

18:34 <bwbuhse> ifreund: yeah fair enough, thanks! I don't even use it super often but occasionally I use it just to put a floating window in and toggle it on/off when I need

18:34 <bwbuhse> that's even worked totally fine for me for a while now, I was just thinking about it when trying to move a window out earlier and figured I'd ask

18:49 <leon-p> ifreund: nice, thanks!

18:49 <leon-p> now I can (almost) stop trying to overengineer initial positions of floating windows for a reference window manager

18:50 <leon-p> the only thing bothering me right now, apart from the missing stuff of course, is that the output dimensions event sends the wrong dimensions at first (0,0) so I always have to resize the river window for stuff to work

18:54 <ifreund> leon-p: hmm, I'm pretty sure that's a quirk of the wayland backend

18:55 <leon-p> oh

18:55 <ifreund> perhaps we can work around it somehow though

18:56 <ifreund> wait, are you sure about that?

18:56 <ifreund> output dimensions seem to be sent properly for me

18:57 <ifreund> https://0x0.st/8slB.txt

18:57 <leon-p> ifreund: do you have the river window floating or tiled? Bec. I have all wlroots windows auto-floating so they don't get a second configure

18:57 <leon-p> after resizing the window once, it works fine

18:58 <ifreund> tiled, I'll try floating

18:59 <ifreund> yeah, can confirm the 0,0 when it's floating

18:59 <ifreund> kinda seems like a wlroots bug to me

18:59 <leon-p> makes sense

23:16 <leon-p> ooohhh, I really like that multiple windows can be moved at once. For a floating WM, a user could f.e. click-drag to draw an area and move multiple selectedd windows at once

23:17 <leon-p> Now I only need to implement that selection ui, I guess

23:19 <leon-p> although I am wondering whether resize/move could be the same operation globally, so that you don't have resize_add_window but pointer_operation_add_resize. I.e. allowing both resizing and moving simultaneously

23:19 <leon-p> either way, I'll have fun with it

23:27 <ifreund> leon-p: resize_add_window requires edges, move doesn't

23:28 <ifreund> we could model move as a resize with all edges set I guess, it feels a bit hacky though

23:29 <ifreund> also, click-drag to select an area isn't possible with the current protocol

23:30 <ifreund> I have thought about sending absolute pointer coordinates when a pointer binding is triggered though

23:31 <leon-p> you could have a surface above all windows where you do the actual selection, like slurp does

23:32 <ifreund> oh, yes that should work I guess

23:32 <ifreund> let the hacks done with river_shell_surface_v1 begin

23:32 <ifreund> do you have a use-case for moving and resizing at the same time?

23:33 <leon-p> Not one I would actually do myself

23:33 <leon-p> but I wondered about "attaching" one window to the side of a another window. Resizing the "host" should move the attached window, maybe

23:33 <ifreund> I think I figured out what you mean, we could have op_start_pointer, op_start_serial, op_add_resize_window, op_add_move_window, op_end

23:34 <leon-p> again, not something I am actually do / use

23:34 <leon-p> s/am/will/

23:34 <ifreund> I did dislike how much protocol text is duplicated between the new requests, so I may change to that simply because it's cleaner design

23:35 <leon-p> yay for hygiene