ifreund changed the topic of #river to: river - a dynamic tiling wayland compositor || https://github.com/riverwm/river || channel logs: https://libera.irclog.whitequark.org/river/
smach has quit [Ping timeout: 268 seconds]
muirrum has joined #river
eShaev9z has joined #river
eShaev9z_ has quit [Ping timeout: 255 seconds]
waleee has quit [Ping timeout: 248 seconds]
jao has quit [Ping timeout: 252 seconds]
elshize has quit [Ping timeout: 272 seconds]
ohno has joined #river
ohno has left #river [#river]
Ordoviz has joined #river
Ordoviz has quit [Ping timeout: 265 seconds]
Ordoviz has joined #river
talismanick has quit [Ping timeout: 268 seconds]
<andrea> NickH: git submodule update --init
<novakane> git submodule doesn't works in tarball since this is not a git repo
<novakane> but I just tried the tarball and it works for me, submodules are here NickH
captain_chaos has joined #river
<captain_chaos> Hey devs, huge thanks for still developing this! And also for publishing 0.2.0 after almost a year" Hiding cursor is a gamechanger for me. I always moved it in the bottom right corner after logging in before (:
hspak60 has quit [Quit: Ping timeout (120 seconds)]
hspak60 has joined #river
<captain_chaos> I have a question though. I use tags basically like workspaces. Id like to make more use of them. But they are a bit unclear. I mean, when I focus tag 1 and lets say the terminal, and I want to have it on tag 3 too.
<captain_chaos> Then how do I make sure that happens? Is there a way to ADD the tag?
<captain_chaos> I know there is toggle, but it means I need to double check on tag 3 since the terminal might already have been there and now is removed from tag 3
<captain_chaos> Same with removing from a tag.
<captain_chaos> I mean, same with being on tag 1 and wanting to remove it from tag 3 without focusing back and forht. I could just use the set command to set it to tag 1 only, but maybe it is actually on tag 6 too and I don't want to remove it there too.
<captain_chaos> I hope Im not annoying, I wrote a bit much..
<leon-p> captain_chaos: not annoying at all, don'worry
<leon-p> technically you can also use the set command to assign a window to multiple tags, but that's of course not a bit hard to do nicely with keybinds
<leon-p> what you want are dedicated add / remove tag commands. unfortunately we do not have them right now
<captain_chaos> Yeah I mean is there a way to script something like that in shell?
Szadek has quit [Quit: WeeChat 3.7.1]
<leon-p> no, that's not possible right now
<captain_chaos> Ok ):
<captain_chaos> Well thanks for confirming that then...
<captain_chaos> River is the bes window manager I tried so far eihter way
<captain_chaos> *best
<leon-p> that's good to hear! :)
captain_chaos has quit [Quit: Client closed]
Szadek has joined #river
Szadek has quit [Ping timeout: 265 seconds]
Szadek has joined #river
elshize has joined #river
Misthios has quit [Quit: Misthios]
Misthios has joined #river
Misthios has quit [Client Quit]
Misthios has joined #river
hendry has quit [Quit: WeeChat 3.7.1]
<scorpion2185[m]> captain_chaos unclutter can hide the cursor look sxmo service, if you plug a mouse it will appear.
<scorpion2185[m]> ExecStart=/usr/bin/unclutter --hide-on-touch --start-hidden
kdk has joined #river
<codedotjs[m]> Thanks scorpion2185. This programm is really helpful.
Lemanr has joined #river
<codedotjs[m]> Does "unclutter --fork" the same as "unclutter &"?
<codedotjs[m]> Well it looks like unclutter is for X11 only. Does it work for you scorpion2185? On river?
<Lemanr> Also would depend if you build river with xwayland or not and then if xwayland & unclutter work well together.
<Lemanr> I don't know this topic detailed enough to know what xwayland's limitations are in terms of X11 functionality
<scorpion2185[m]> I can't try it I only have unclutter-xfixes, I though that unclutter was the wayland version sorry
<scorpion2185[m]> unclutter-xfixes doesn't work
<scorpion2185[m]> I compile river with xwayland
<scorpion2185[m]> s/compile/compiled/
<codedotjs[m]> I also have xwayland enabled. Doesn't work either. On Archlinux the unclutter package is unclutter-xfixes. But nevermind. There is no need to use unclutter on river. There are builtin commands in riverctl which handle cursor hiding.
<scorpion2185[m]> builtin commands?
<ifreund> are you looking for the riverctl hide-cursor command?
<Lemanr> also this may help answer the first question about & vs --fork https://stackoverflow.com/questions/3096561/fork-and-exec-in-bash . One of the comments refers to a section on a reference manual which could provide details you'd want.
<ifreund> & is implemented by the shell and --fork is implemented by the program
<ifreund> & will always do the same thing, what --fork does depends on the program
<scorpion2185[m]> ifreund yes thanks.
<Lemanr> indeed, as a flag of that program it really does depend on what ever that programmer/team means by --team. Perhaps they have a man page that would provide details? I'd google to see if they have a section on some site if there is no man page.
<Lemanr> *by --fork lol
Lemanr has quit [Quit: Client closed]
<codedotjs[m]> The man page said something like "fork process to background". But I already uninstalled it.
<scorpion2185[m]> unclutter-xfixes is a forkof uncl,utter
Lemanr has joined #river
<ifreund> I think I'll do a purely bugfix 0.2.1 with the handful of small fixes I've landed today and then upgrade to zig 0.10 in 0.2.2
<ifreund> So, y'all have until I finish the Zig upgrade to find and fix any more minor bugs to be included in 0.2.1 I guess
jao has joined #river
Guest51 has joined #river
Guest51 has quit [Quit: Client closed]
Lemanr has quit [Quit: Client closed]
dbuckley has quit [Read error: Connection reset by peer]
dbuckley has joined #river
muirrum has quit [Read error: Connection reset by peer]
muirrum has joined #river
<scorpion2185[m]> I can't see it on the man page how do I use it?
<scorpion2185[m]> >are you looking for the riverctl hide-cursor command?
<scorpion2185[m]> are all the option near this  input name click-method none|button-areas|clickfinger for the touchpad right ?
<leon-p> the man pages are not sorted that way
<leon-p> also it works a bit different than "is touchpad, therefore this option applies". Instead we just ask libinput if an option is supported, regardless of input type, and apply if the answer is yes
Ordoviz has quit [Ping timeout: 272 seconds]
<codedotjs[m]> You can search for options with the search command: /
<codedotjs[m]> e.g. /hide-cursor
<codedotjs[m]> It is around line 290.
<codedotjs[m]> There are two hide-cursor options. One for timeout and one for when-typing.
<scorpion2185[m]> man riverctl?
<codedotjs[m]> Yes.
<scorpion2185[m]> how do you see the line n°?
<codedotjs[m]> There is a line at the bottom.
<scorpion2185[m]> leon-pI wonder if some of those options can be applied to the touchscreen too
<codedotjs[m]> The line number is just for reference. Just open "man riverctl" and then type "/hide-cursor". Now press enter. You will be directly brought to the first match.
<scorpion2185[m]> Do you what options are also available for it?
<scorpion2185[m]> "No matches found"
<codedotjs[m]> For the touchscreen you can probably use one of the options in the input section.
<scorpion2185[m]> I can find other things e.g. /xcurs
<scorpion2185[m]> I didn't update the man pages
<codedotjs[m]> I use those options in my init: riverctl hide-cursor timeout 2000 | riverctl set-cursor-warp on-focus-change | TOUCHPAD=pointer-1267-12363-ELN4690:00_04F3:304B_Touchpad | riverctl input $TOUCHPAD natural-scroll enabled | riverctl input $TOUCHPAD tap enabled | riverctl input $TOUCHPAD disable-while-typing disabled | TOUCHSCREEN=touch-1386-20931-Wacom_HID_51C3_Finger | riverctl input $TOUCHSCREEN events enabled
<scorpion2185[m]> thanks!
<scorpion2185[m]> man pages on .local are superseeded by old man pages in /usr/local
Ordoviz has joined #river
<leon-p> wow...
<scorpion2185[m]> pointer-9011-26214-ydotoold_virtual_device
<scorpion2185[m]> configured: false
<scorpion2185[m]> should I configure ydotool pointer too?
<leon-p> I would be surprised if it would have any effect
<codedotjs[m]> I have no experience with ydotoold at all but if you want to use it you might have to configure it too.
<scorpion2185[m]> I have this in river config `pgrep -x polybar > /dev/null || { waybar & }`
<scorpion2185[m]> works fine but only when I quit river, If I reload the config I get 2 waybars
<scorpion2185[m]> lol
<scorpion2185[m]> they are all like that
taupiqueur has quit [Ping timeout: 252 seconds]
<codedotjs[m]> leon-p sent a pastebin with the spawn_once function. This prevents e.g. bars from appearing twice.
<scorpion2185[m]> isn't it the same of that command ? IIRC
<leon-p> a bit surprised ydotool creates a virtual pointer at all. I am pretty sure last time I looked at it, it only did evdev stuff
<codedotjs[m]> https://matrix.to/#/!tSZeOIVRJwMOVgkcCT:libera.chat/$uIDVi6T6ERXdWQuB9ExXPzQM1U7kTH-nee8fxGlpsy4?via=libera.chat&via=matrix.org&via=exozy.me
taupiqueur has joined #river
<scorpion2185[m]> IDK why here on alpine on firefox besides long press ->right click when I scroll I have to use the scrollbar.
<scorpion2185[m]> While on midori I can scroll with the touchscreen
<leon-p> codedotjs[m]: is that some matrix feature that translates poorly to IRC? Or did you just sent a link?
<scorpion2185[m]> (besides that not working)
<leon-p> scorpion2185[m]: that's because midori uses the touch input and firefox not
<leon-p> maybe firefox is running via XWayland?
<leon-p> if you have the `xlsclients` command, look at its output
<codedotjs[m]> I just sent a link. To your message. But wait I can send the pastebin link directly.
Guest4844 has joined #river
Guest4844 has quit [Client Quit]
<scorpion2185[m]> it is the same but -f instead of -x as pgrep arg
<scorpion2185[m]> I don't such package , maybe I need to export some wayalnd var for firefox
<scorpion2185[m]> *dont have
talismanick has joined #river
<leon-p> whether you want -x or -f depends on your use case. -f compares the string against the entire command line of the process (used because by default pgrep only compares against a fixed amount of characters, which breaks for long command names), while -x searches for the exact command line and won't work if there are flags
<leon-p> or in simpler terms: -x searches for commands that are exactly the same as the search term, -f searches for commands which start with the search term
taupiqueur has quit [Ping timeout: 264 seconds]
<codedotjs[m]> scorpion2185[m]: I can use the touchscreen on Firefox (Archlinux, Wayland). I have got the following env var in my start script (not init script - the start script contains some env vars and runs river with the init script afterwards): export MOZ_ENABLE_WAYLAND=1
<codedotjs[m]> Probably this is needed?
taupiqueur has joined #river
<leon-p> i wonder if we should put some form of that spawn_once function into the default example. reloading the init (aka executing it again) is a use comes up often, I feel
<codedotjs[m]> Yeah that would be great. Also adding a reload function by default would help a lot.
<scorpion2185[m]> yes with a realoding shortcut command too: near super+shift+e quit super+shift+r reload
<codedotjs[m]> Well I personally use quite different shortcuts for that. But adding such an option to the example configuration would also be nice.
<scorpion2185[m]> echo $MOZ_ENABLE_WAYLAND doesn't return anything
<scorpion2185[m]> ah yes Window Protocolxwayland
<scorpion2185[m]> I added it to the script that start river but it is ignored
<leon-p> codedotjs[m]: river doesn't need an extra command for that, the init can just map a keybind to spawn itself
<leon-p> i.e. `riverctl map normal Super R spawn $0`
<scorpion2185[m]> so $0 will reload it?
<leon-p> no
<leon-p> in shell, $0 expands to the first token of the command line
<leon-p> try it. create a shell script and put `echo $0` in it. then run it.
<codedotjs[m]> leon-p: I'll try that. In the moment I use riverctl map normal Super R spawn '$XDG_CONFIG_HOME/init'.
<scorpion2185[m]> export MOZ_ENABLE_WAYLAND=1
<scorpion2185[m]> XDG_RUNTIME_DIR=/tmp river
<scorpion2185[m]> do I need to combine those 2 ?
<scorpion2185[m]> I tried it nice so it will "reload" the init
<codedotjs[m]> scorpion2185[m]: Well I can send you the code of my runscript. In general: if you just want to run a command with certain env vars once you put them in front of it and run it (e.g. "MOZ_ENABLE_WHATEVER=1 XDG_SOME_RANDOM_DIR=$HOME/.local/dir river"). If you want to run them everytime it can help a lot writing a script. There you put the envvars underneath each other and add the "export" command before them. In the end put "exec command" into
<scorpion2185[m]> I would like to see it
<codedotjs[m]> the script.
<scorpion2185[m]> and I would totally put riverctl map normal Super R spawn $0 in the example
<codedotjs[m]> Definitely. Here is my runscript (this was the first pastebin I found on the internet. Can you suggest another one?) https://pastebin.com/bASqAPrG
<scorpion2185[m]> centos pastebin service
<codedotjs[m]> Thanks.
<codedotjs[m]> Don't copy the whole file please. Some of the paths are directly modified for my workflow. They will probably break yours. Just take a look at how it works in theory.
<scorpion2185[m]> that's how I modify the file
<scorpion2185[m]> *modified , but mozilla var is ignored
<scorpion2185[m]> I am still using a var to se the keyboard layout and that is shown with echo
<scorpion2185[m]> I can ask alpine to how set that I guess
<codedotjs[m]> You can set the keyboard in river now e.g. `riverctl keyboard-layout -options "grp:ctrl_space_toggle" "de,us"` (see the man page for more info).
<scorpion2185[m]> .profile worked for firefox I can scroll and right click
<codedotjs[m]> With the screen?
Guest4887 has joined #river
<codedotjs[m]> <leon-p> "i.e. `riverctl map normal..." <- I tested it and it works. Even better than the command I was using before.
Guest4887 has quit [Client Quit]
<scorpion2185[m]> yes. I saw that var in the global .profile once
<scorpion2185[m]> IIRC centos doesn't last longer if you need to have the past there for days
Ordoviz has quit [Ping timeout: 272 seconds]
<leon-p> scorpion2185[m]: glad your touchscreen troubles are solved! with firefox now working as expected, I hope what I have been trying to say is at least a tiny bit clearer now
waleee has joined #river
<NickH> Hmm, surely the tar ball should be self contained and not required a git submodule call. And if it really does it should be documented.
<novakane> NickH: it is self contained, I'm not sure if your download was messed up or something but for me submodules are not empty
<NickH> ok, must be something that uscan is doing.
<NickH> Seems I'll need to indivually package all the bundles dependencies.
<NickH> Neverending fun.
<novakane> oh yeah sounds really fun, glad I don't have to do this
<novakane> it is weird that some dirs would be removed from the tar ball though no? Or you're not using the github release download link?
<NickH> uscan reads a regex that is part of a debian source package. The regex is the url of the upstream packages. It then downloads it. It doesn't surprise me that it strips out dependencies since in the debian world they should be their own self contained packages.
<NickH> Anyway still on year end vacation. Will try to find time for this next week.
<NickH> Been almost a year since the last river release
<NickH> guess I wasn't using uscan back then
alebastr has quit [Ping timeout: 272 seconds]
alebastr has joined #river
Moulberry has joined #river
Moulberry has quit [Client Quit]
Moulberry has joined #river
Moulberry has quit [Client Quit]
Szadek has quit [Quit: WeeChat 3.7.1]
Szadek has joined #river