01:27 <leon-p> just PR'd a commit introducing a threshold for move/resize requests of tiled XDG views

01:28 <leon-p> this will solve the issue of those views unexpectedly starting to float; explanation in PR description

08:13 <leon-p> hell, just noticed river depends on glib D:

08:13 <leon-p> how is that library everywhere?

08:15 <leon-p> No, neither systemd nor dbus are the ultimate menace, it's that bloody library.

08:51 <ifreund> huh, I wonder what's pulling it in transitively

09:11 <ifreund> leon-p: it's libwacom's fault :/

09:11 <ifreund> lddtree is pretty neat: https://0x0.st/-XGH.txt

09:58 <ifreund> leon-p: here's an alternative proposal for the issue your PR addresses: ignore request_move/request_resize on non-floating views

09:58 <ifreund> would be a lot simpler for sure

10:04 <yyp> ifreund: why is command to move current window to the top is zoom?

10:05 <yyp> riverctl zoom I mean

10:07 <ifreund> yyp: well the top of the stack is commonly the most prominant/largest view

10:07 <yyp> Ok, how the reverse should be called then?

10:10 <yyp> minimize?

10:12 <ifreund> hrm, I don't see any reason for command to do the opposite tbh

10:12 <ifreund> the idea with the stack is that you zoom the views you care about which has the effect of automatically sorting the stack roughly by priority

10:13 <ifreund> Swap up/down exist, but I personally never use them

10:13 <yyp> I sometimes get into situations where I want to push window to the bottom of the stack (something like a man page open). And since there's no way todo that with a script, I thought it should be here

10:14 <yyp> *to do

10:14 <yyp> nevermind, I forgot that push next on the top window will do the job

10:15 <yyp> s/push/swap

12:50 <leon-p> ifreund: I am not sure about not allowing tiled views to request move/resize. It's a bit arbitrary, considering keybind+move would still work, and I suspect people who use applications with header bars do tend to expect being able to drag them.

12:50 <leon-p> also the threshold does solve this nicely.

12:50 <leon-p> with a bit of work the commit could probably even be made a bit simpler

12:57 <leon-p> if I can somehow deduplicate the moveThreshold and resizeThreshold, which are basically the same code, I can drop at least a third of the code

13:10 <ifreund> I wonder what sway does here

13:11 <leon-p> sway ignores it

13:11 <ifreund> entirely? that's what my personal preference would be tbh

13:11 <ifreund> I wonder how many people complain

13:12 <leon-p> not entirely, just for tiled views. floating ones can request move/resize

13:12 <ifreund> ah, so what I just suggested :D

13:13 <leon-p> if you prefer it, I can change the PR to do just that.

13:28 <leon-p> yeah, I think I'll just do that.

13:31 <leon-p> ifreund: new PR is out

13:32 <leon-p> I did not overwrite the old one so that we have the code if people do complain and have a compelling reason to allow move/resize requests for tiled views

13:41 <ifreund> leon-p: cool, sounds good to me

13:41 <ifreund> your code for the threshold isn't terribly complex, but not having makes me even happier :D

13:45 <leon-p> yeah :D

13:45 <leon-p> now onto the next project: seeing whether libwacom can be liberated from glib...

13:47 <ifreund> heh, good luck :)

13:54 <leon-p> well, the library is only about 20k, so definitely doable. but they use the glib stuff all over and judging by their github issues/PRs they probably like it that way :/

13:55 <leon-p> the actual library appears to be only 2.5k even, but still...

14:03 <ifreund> leon-p: probably be good to ping whot in #wayland and talk about it before putting too much time in

14:05 <leon-p> no worries, I already abandoned the idea. they use glib in basically every other function and make heavy use of it in the testing suite. It's doable, but I highly doubt such a change would be accepted.

14:08 <ifreund> sigh :/

14:09 <ifreund> you can probably build libinput without wacom support if you want

14:20 <leon-p> probably, but it would just have been for peace of mind anyway. In reality at least 50% of things on any modern Linux installation depends on glib in some way, and I don't think that will change anytime soon.

20:49 <ifreund> ;5;44~

21:32 <dnkl> Shift+,

21:32 <dnkl> ?

21:32 <dnkl> :)

21:34 <ifreund> indeed, someone certianly knows their terminal escapes :P

21:44 <dnkl> Not well enough... 5 is ctrl, not shift 😅