elshize has quit [Read error: Connection reset by peer]
elshize has joined #river
<elshize>
when I use river, my computer sometimes freezes and requires a hard reset (can't even access a tty) I'm not sure how to debug this. where can I look for some logs (or turn logs on if needed)? any suggestions would be appreciated
<tmpm697>
elshize: if I were you, I'll try to switch to open source gpu driver and try linux-lts
<elshize>
tmpm697: I'm already on open source drivers (intel) and on lts...
waleee has quit [Ping timeout: 240 seconds]
elshize has quit [Read error: Connection reset by peer]
elshize has joined #river
snakedye has quit [Ping timeout: 240 seconds]
novakane has joined #river
snakedye has joined #river
yyp has joined #river
priner[m] has quit [Quit: Bridge terminating on SIGTERM]
sjtio has quit [Quit: Bridge terminating on SIGTERM]
sjtio has joined #river
leon-p has joined #river
priner[m] has joined #river
yyp has quit [Read error: Connection reset by peer]
yyp has joined #river
<leon-p>
Submitted a PR to fox the bug I found last night. Just a check I forgot to include when I implemented the request_{move,resize} events.
<leon-p>
s/fox/fix
<dnkl>
leon-p: "fox the bug" sounds better :D
<leon-p>
it does :D
<dnkl>
"today I'll out-fox that bug, for sure"
<leon-p>
Well, /this/ bug totally out-foxed me. It just completely slipped my mind that views can be fullscreen when I wrote those 4 lines.
<leon-p>
only found it because I was playing around with drawing CSD motif-style
<tmpm697>
leon-p: with sway, can I have like workspace 1 have tabbed window layout
<tmpm697>
with config: workspace_layout tabbed
<tmpm697>
and workspace 2 like normal one?
tmpm697 has quit [Remote host closed the connection]
tmpm697 has joined #river
<tmpm697>
hi :)
<leon-p>
tmpm697: you should ask sway questions on #sway
<tmpm697>
ok.
<leon-p>
the people there are more likely to have a good answer than some random river contributor
<tmpm697>
:)
novakane has quit [Quit: WeeChat 3.2]
novakane has joined #river
andrea has joined #river
andrea has quit [Client Quit]
andrea has joined #river
andrea has quit [Client Quit]
novakane has quit [Quit: WeeChat 3.2]
novakane has joined #river
novakane has quit [Client Quit]
novakane has joined #river
tmpm697 has quit [Quit: leaving]
tmpm697 has joined #river
waleee has joined #river
leon-p has quit [Quit: leaving]
<novakane>
so the mako issue is the last thing todo before 0.1.0 I guess, don't know if the release would be block by this
dbuckley has joined #river
yyp has quit [Remote host closed the connection]
_whitelogger has joined #river
novakane has quit [Quit: WeeChat 3.2]
<ifreund>
elshize: you can redirect all of river's output to a log file that persists across reboot with e.g. `river > ~/river.log 2>&1`
<ifreund>
you may also be able to avoid a hard reset in that siutation with the sysreq key
novakane has joined #river
<ifreund>
specifically sysreq + r will reclaim the keyboard from the running X server or wayland compositor or similar and allow you to switch ttys
ifreund_gamja has joined #river
ifreund_gamja has quit [Remote host closed the connection]
<novakane>
ifreund: so is smart borders something you interested in to add in river?
<ifreund>
novakane: they should be pretty straightforward to implement in a layout client
<ifreund>
or just using river-status + river-control
<ifreund>
might have issues with frame perfection, but river-control-v2 should address that
<ifreund>
man I've got a lot of open issues to deal with
<novakane>
ifreund: I thought about using river layout client event, like in rivertile with view_count and make a fn returning a bool and just use it with renderBorders(), but I think river only handle river layout server request
<novakane>
and a lot of PR to review too :P
<ifreund>
novakane: river only implements the server side of the protocol yes, it is a compositor not a client
<novakane>
and half of yesterday issues have been already close, you're lucky :P
<ifreund>
you'd need to modify rivertile or write a new layout client
<ifreund>
yeah, I saw the email spam :/
<novakane>
I know that what I did to have smart gaps on my layout generator, but can you handle border rendering from it?
<ifreund>
novakane: yes, using the river-control protocol
<ifreund>
send "border-with" and "0" as the arguments
<novakane>
right I'll do this then, it's better thanks
<ifreund>
no problem, like I said this might cause imperfect frames if you get unlucky currently
<ifreund>
interested to hear how it works in practice
<novakane>
well I'll try and I'll tell you
tmpm697 has quit [Quit: Lost terminal]
tmpm697 has joined #river
<elshize>
ifreund: thanks, I'll try that
elshize has quit [Quit: WeeChat 3.2]
elshize has joined #river
<snakedye>
Setting the border to 0 only really work well with one monitor
<ifreund>
true, river-control-v2 should fix that too
<ifreund>
I
<ifreund>
I'm very confused by one thing in #323, the default background color is dark blue not green
<tmpm697>
ifreund: yes, I can't distingush dark blue and green :)
<ifreund>
tmpm697: ah, that explains that then :D
leon-p has joined #river
<leon-p>
just implemented direct scanout for fullscreen views, just to find out literally no program I run fullscreen uses it -_-
<ifreund>
lol, not even mpv?
<leon-p>
haven't tried it yet. that laptop had no videos on it :D
<leon-p>
anyway, I'll submit a draft PR
<ifreund>
cool
<leon-p>
I'll convert it into a non-draft PR once I know whether it works
<ifreund>
I saw your ping about rendering based on the focus stack. There are a few issues I see with that:
<ifreund>
1. we'd have to double-buffer the focus stack, which would complicate the focus handling code in Seat
<ifreund>
2. I don't see a good way to handle multiple seats
<leon-p>
I see, makes sense
<ifreund>
I also don't see a problem with the way river currently renders things
<leon-p>
and what about my ping regarding the background colour? Changing it to black hides the "problem", so we don't even need to "solve" it :P
<ifreund>
I'm not happy with that, as then it's really hard to tell if river has actually started or not with a default config
<ifreund>
especially if the user doesn't have any spawn terminal/launcher command working
<ifreund>
sway "sovles" this by tightly coupling swaybg and sway, adding special handling for swaybg i believe
<leon-p>
makes sense
<leon-p>
although I don't think tightly coupling swaybg fits to river, so we just have to live with it I guess
<ifreund>
no, I definitely don't want to do that
<ifreund>
there are other possible solutions though, for example add a flag to river to delay rendering the first frame until some command is run
<ifreund>
that wouldn't solve this for new outputs though, for that we would need that layer shell output_ack thing implemented
<novakane>
what if renderer.clear() was black and the background set later?
Misthios has quit [Quit: Misthios]
Misthios has joined #river
<ifreund>
novakane: renderer.clear() is *the* way to implement a background color, I don't know how else you would do it
<novakane>
hmm right ok, otherwise it would be a surface?
<ifreund>
novakane: no, surfaces are what we get from clients. In the compositor we can just render directly using wlroot's renderer abstraction
<novakane>
ah, rendering is too complicated :P
<leon-p>
ifreund: btw, not sure if you have seen already, but I have updated the github wiki a bit. Hopefully the new FAQ covers the most basic things people would ask about.
<ifreund>
leon-p: I had not seen, it looks great :) Thank you!
waleee has quit [Quit: WeeChat 3.1]
waleee has joined #river
novakane has quit [Quit: WeeChat 3.2]
snakedye has quit [Quit: Quit]
snakedye has joined #river
<snakedye>
tmpm697: I cannot replicate your issue. It compiles on my end on Arch.