| - mtm (QUIT: Ping timeout: 268 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 01:04 | |
| + freakazoid343 (~matt@2603:9000:cf03:86c3:4cb5:1453:777d:3b64) | 01:46 | |
| - freakazoid333 (QUIT: Ping timeout: 240 seconds) (~matt@2603:9000:cf03:86c3:4cb5:1453:777d:3b64) | 01:49 | |
| - nsc (QUIT: Ping timeout: 240 seconds) (~nicolas@i5C7443CE.versanet.de) | 03:09 | |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 03:09 | |
| + nsc (~nicolas@i5C7440A3.versanet.de) | 03:11 | |
| * nsc -> Guest505 | 03:11 | |
| - Christoph_ (QUIT: Remote host closed the connection) (~Christoph@p54bf62fd.dip0.t-ipconnect.de) | 03:23 | |
| + bleb (~cm@user/bleb) | 04:32 | |
| - sts-q (QUIT: Ping timeout: 268 seconds) (~sts-q@212.53.219.218) | 04:51 | |
| + sts-q (~sts-q@212.53.219.148) | 04:57 | |
| - Ar|stote|is (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~linx@149-210-24-194.mobile.nym.cosmote.net) | 05:32 | |
| + Ar|stote|is (~linx@149-210-24-194.mobile.nym.cosmote.net) | 05:32 | |
| - Ar|stote|is (QUIT: Client Quit) (~linx@149-210-24-194.mobile.nym.cosmote.net) | 05:36 | |
| + Ar|stote|is (~linx@149-210-24-194.mobile.nym.cosmote.net) | 05:36 | |
| - Ar|stote|is (QUIT: Client Quit) (~linx@149-210-24-194.mobile.nym.cosmote.net) | 05:36 | |
| + Ar|stote|is (~linx@149-210-24-194.mobile.nym.cosmote.net) | 05:36 | |
| + chomwitt (~chomwitt@2a02:587:dc0b:200:12c3:7bff:fe6d:d374) | 07:29 | |
| - GNUmoon (QUIT: Ping timeout: 276 seconds) (~GNUmoon@gateway/tor-sasl/gnumoon) | 07:32 | |
| + erlehmann (~erle@ip5f5bd566.dynamic.kabel-deutschland.de) | 07:53 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 08:17 | |
| + MajorBiscuit (~MajorBisc@86-88-79-148.fixed.kpn.net) | 08:45 | |
| - MajorBiscuit (QUIT: Client Quit) (~MajorBisc@86-88-79-148.fixed.kpn.net) | 08:45 | |
| + MajorBiscuit (~MajorBisc@86-88-79-148.fixed.kpn.net) | 08:45 | |
| + Major_Biscuit (~MajorBisc@c-001-018-021.client.tudelft.eduvpn.nl) | 09:25 | |
| - MajorBiscuit (QUIT: Ping timeout: 256 seconds) (~MajorBisc@86-88-79-148.fixed.kpn.net) | 09:28 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:21:21:0:100b) | 09:47 | |
| - bluerise (QUIT: Quit: brb) (~bluerise@user/bluerise) | 10:31 | |
| + bluerise (~bluerise@user/bluerise) | 10:38 | |
| + Christoph_ (~Christoph@p4fe73f2f.dip0.t-ipconnect.de) | 10:44 | |
| + mjw (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 10:58 | |
| Boostisbetter | mntmn, or the community here, does the TRRS jack on the Reform support headsets? If so is there a trick to enabling it? | 11:50 |
|---|---|---|
| - xktr (QUIT: Ping timeout: 240 seconds) (~xktr@37.120.147.186) | 12:00 | |
| mntmn | Boostisbetter: yes it does | 12:52 |
| mntmn | Boostisbetter: but didn't work for me with pipewire yet, only alsa/pulse... you need to make sure that these files are there (they were overwritten for me once by upgrade): https://source.mnt.re/reform/reform-system-image/-/tree/main/reform2-imx8mq/template-etc/pulse | 12:52 |
| mntmn | Boostisbetter: here's where they should go https://source.mnt.re/reform/reform-system-image/-/blob/main/reform2-imx8mq/mkuserland.sh#L85 | 12:53 |
| - mtm (QUIT: Ping timeout: 256 seconds) (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 13:02 | |
| + frank_ (~frank@juicy.frank2.net) | 13:42 | |
| + dodo__ (~dodo@user/dodo) | 13:43 | |
| + robin__ (~robin@user/terpri) | 13:43 | |
| + lexik_ (~lexik@171.25.222.230) | 13:44 | |
| + leonardo1 (~leonardo@user/leonardo) | 13:45 | |
| + skalk_ (~skalk@vond.sysret.de) | 13:45 | |
| - _Bnu (QUIT: Ping timeout: 240 seconds) (~beeanyew@89-160-120-72.cust.bredband2.com) | 13:46 | |
| - frank2 (QUIT: Ping timeout: 240 seconds) (~frank@juicy.frank2.net) | 13:46 | |
| - leonardo (QUIT: Ping timeout: 240 seconds) (~leonardo@user/leonardo) | 13:46 | |
| - kfx (QUIT: Ping timeout: 240 seconds) (~kfx@wopr.sciops.net) | 13:46 | |
| - robin (QUIT: Ping timeout: 240 seconds) (~robin@user/terpri) | 13:46 | |
| - dodo (QUIT: Ping timeout: 240 seconds) (~dodo@user/dodo) | 13:46 | |
| - S0rin (QUIT: Ping timeout: 240 seconds) (~S0rin@user/s0rin) | 13:46 | |
| - skalk (QUIT: Ping timeout: 240 seconds) (~skalk@vond.sysret.de) | 13:46 | |
| - lexik (QUIT: Ping timeout: 240 seconds) (~lexik@171.25.222.230) | 13:46 | |
| * dodo__ -> dodo | 13:46 | |
| + _Bnu (~beeanyew@89-160-120-72.cust.bredband2.com) | 13:46 | |
| + S0rin (~S0rin@user/s0rin) | 13:47 | |
| andrey_utkin | hi all, i'm trying to get back to activity. i see a great progress on package generation :) i thought that the best I can do right now is to write some automated tests for video decoder performance. Could somebody please run `make` on https://source.mnt.re/autkin/tests/-/tree/main on the Reform both with a custom-built ffmpeg and with the vanilla debian ffmpeg, and share the "Average FPS" output? | 13:49 |
| - Guest505 (QUIT: Ping timeout: 256 seconds) (~nicolas@i5C7440A3.versanet.de) | 13:49 | |
| + nsc (~nicolas@i5C7440A3.versanet.de) | 13:50 | |
| * leonardo1 -> leonardo | 13:50 | |
| * nsc -> Guest8449 | 13:50 | |
| andrey_utkin | mntmn: i managed to create the repo https://source.mnt.re/reform/tests/ but couldn't push to it, it says ~call Microsoft~ ask maintainers to create a branch | 13:51 |
| mntmn | andrey_utkin: ok let me see, i will make you a maintainer | 14:00 |
| mntmn | andrey_utkin: btw what does tests mean exactly (purpose of the repo, can we specify it more?) | 14:00 |
| andrey_utkin | right, i will make a readme, the purpose is testing whether intended features work | 14:01 |
| andrey_utkin | another purpose is to track features status on vanilla distros, as some features may start work (just so or differently) outside of our control | 14:03 |
| mntmn | ah, so tests is the right name then | 14:06 |
| mntmn | very good | 14:06 |
| mntmn | andrey_utkin: sorry i had to fiddle a bit there, but now you should be maintainer of tests | 14:14 |
| andrey_utkin | thanks | 14:19 |
| * robin__ -> robin | 14:29 | |
| Boostisbetter | mntmn: Once I confirm or move this 2 files there, does that mean it will be automatic? Or do I need to open alsa and select it when plugging a headset in? | 14:34 |
| mntmn | UPS pickup for first large shipment of MNT Reforms and accessories to Mouser/Crowd Supply scheduled for this friday | 14:41 |
| mntmn | Boostisbetter: should be automatic, see if input is selected in pavucontrol | 14:42 |
| + mtm (~mtm@c-73-27-62-116.hsd1.fl.comcast.net) | 15:09 | |
| + xktr (~xktr@37.120.147.5) | 15:37 | |
| Boostisbetter | mntmn: it seems like it is selected but it only seems to be picking up things like me hitting the mic and not my actual voice. I'll look into it more tomorrow, while I am at work. | 15:57 |
| mntmn | Boostisbetter: there are two types of TRRS headsets, is it CTIA standard or OMTP? | 16:31 |
| sigrid | anyone ever experienced sway freeze every N seconds for a second? idk why this happens, it seems like some sort of a timer that gets fired with exact period. input gets blocked for a second and then it goes on as usual | 16:48 |
| sigrid | but it's damn unbearable to type | 16:48 |
| sigrid | it's working fine in tty | 16:48 |
| sigrid | idk how to even debug this | 16:49 |
| Boostisbetter | mntmn: they are the apple headphones that have a mic in the line controls. I got then cheap for work. I don't really use them but I figured they might make webrtc calls possible on the Reform. | 16:50 |
| + vagrantc (~vagrant@2600:3c01:e000:21:21:21:0:100b) | 17:03 | |
| sigrid | strangely enough it works fine after a reboot | 17:06 |
| mntmn | Boostisbetter: i believe apple is CTIA | 17:12 |
| mntmn | sigrid: could be waybar | 17:12 |
| mntmn | sigrid: calling the status script | 17:12 |
| Boostisbetter | mntmn: meaning it won't work? | 17:13 |
| Boostisbetter | mntmn: no big deal either way. | 17:13 |
| mntmn | Boostisbetter: CTIA is what we support | 17:13 |
| mntmn | Boostisbetter: OMTP we do not | 17:13 |
| mntmn | Boostisbetter: so probably some config isse / alsa controls issue (there are a ton, sorry i don't have time at the moment to tell you exactly the right ones) | 17:14 |
| Boostisbetter | mntmn: OK, then I'll do a restart just to make sure the files are taken into account, and no worries. This isn't a burning issue or anything. | 17:14 |
| sigrid | mntmn: i killed it but it still was there | 17:25 |
| sigrid | my suspicion is dbus actually | 17:25 |
| + kfx (~kfx@wopr.sciops.net) | 17:37 | |
| mntmn | sigrid: weird. but yes, stuff waiting for dbus can cause delays | 17:40 |
| - mjw (QUIT: Ping timeout: 240 seconds) (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 18:25 | |
| + chartreuse (~chartreus@S0106f0f249dfd9c3.cg.shawcable.net) | 18:44 | |
| - Major_Biscuit (QUIT: Ping timeout: 268 seconds) (~MajorBisc@c-001-018-021.client.tudelft.eduvpn.nl) | 18:51 | |
| sigrid | it happens again now after some time, but not as oftennnnnnnnnnnnnnnnnnnn | 19:03 |
| sigrid | oh yeah, that "nnnnn" is the result | 19:03 |
| mntmn | sigrid: weird. but it's not the keyboard itself freezing? | 19:10 |
| sigrid | no, mouse pointer also freezes | 19:13 |
| sigrid | and in tty everything seems to be fine | 19:13 |
| sigrid | so I don't think it's related to usb either | 19:14 |
| sigrid | only happens in sway :( | 19:14 |
| swivel | i think input handling on modern linux windowing systems has regressed substantially in terms of robustness/resilience vs. scheduling delays | 19:14 |
| swivel | sigrid: in your shoes i'd ssh in from another machine and start profiling things during the stalls | 19:17 |
| sigrid | I am not exactly sure what tools to use for profiling | 19:18 |
| sigrid | I can try enabling debug logs in sway, but that's really all I know | 19:18 |
| - exit70[m] (PART: !!unknown attribute: msg!!) (~exit70@user/exit70) | 19:18 | |
| swivel | sysprof can be helpful here | 19:20 |
| swivel | or if gtk is too much pita check out the lower level `perf` cli tooling | 19:22 |
| sigrid | 00:03:42.904 [ERROR] [wlr] [libinput] event1 - MNT Reform Trackball: client bug: event processing lagging behind by 22ms, your system is too slow | 19:24 |
| sigrid | same with the keyboard | 19:25 |
| sigrid | idk if that's related though | 19:25 |
| sigrid | quick search shows "sticky/laggy mouse input" related to that message | 19:26 |
| swivel | so a lot of the modern compositors, and hell even modern XOrg now, are using a dedicated input thread for consuming input events. It works fine until that thread can't get scheduled for whatever reason - system busy, poor scheduling decisions, locks getting in the way, etc | 19:27 |
| sigrid | lovely :( | 19:28 |
| swivel | XOrg used to use a signal handler but that changed when they started wanting to support implementing XOrg atop GL or other non-reentrant backends where they couldn't assume the signal handler could do things like update the mouse pointer | 19:29 |
| sigrid | reading through similar issue reports suggests it has to do with usb autosuspend feature | 19:34 |
| sigrid | why would it be fine while in tty though - no idea | 19:34 |
| swivel | sigrid: I don't have a wayland/sway setup here, but you've piqued my curiosity as this is on my todo list; can you by chance check if the sway process has multiple threads? | 19:47 |
| swivel | also #sway-devel likely has folks w/relevant knowledge to aid troubleshooting this | 19:49 |
| sigrid | it's only one | 19:49 |
| swivel | that'd make it extremely susceptible to stalls then, if it's a classical IO multiplexed single process model | 19:50 |
| swivel | like even if sway got hung up in rendering with a pipeline stall it'd delay input processing | 19:50 |
| sigrid | I would expect it to have mutiple threads :/ | 19:51 |
| sigrid | is that even an option? | 19:51 |
| swivel | I'd assumed they were doing that at least, since XOrg had transitioned to that in the migration to glamor->wayland | 19:51 |
| swivel | this post describes the wlroots side of input handling, but the event loop appears to be entirely delegated to the wayland protocol lib https://drewdevault.com/2018/07/17/Input-handling-in-wlroots.html | 19:53 |
| swivel | maybe there's a writeup on the wayland library side to be found, to better understand what's going on | 19:54 |
| swivel | if swaybar is tightly integrated synchronously into the sway process, what mntmn said sounds very likely - any delay servicing swaybar's stuff would directly hold up the compositor's event loop | 19:57 |
| mntmn | i don't think it is | 19:58 |
| swivel | i hope it's not ;) | 19:58 |
| mntmn | i mean, waybar (not swaybar) | 19:58 |
| mntmn | it's a different process, basically just another application | 19:58 |
| swivel | my mistake re: (s)waybar | 19:58 |
| mntmn | swaybar is the default shipped with sway (it's like i3bar), but we don't use it | 19:59 |
| chartreuse | As an idea with the mouse/keyboard freezing perhaps still try reflashing the keyboard and/or mouse? Could maybe be something flooding the USB bus? Maybe? | 20:00 |
| mntmn | also, i have never experienced this problem that sigrid has, and using my reform pretty heavily, so it must be something special | 20:00 |
| sigrid | should be noted i was using rndis with an android phone | 20:00 |
| chartreuse | I've not had issues like that either on mine using sway either | 20:00 |
| mntmn | by special i mean unusual | 20:00 |
| sigrid | i will try without when i get home | 20:00 |
| sigrid | https://github.com/swaywm/sway/issues/5591 | 20:05 |
| sigrid | some reports are EXACTLY like mine | 20:05 |
| mntmn | :3 | 20:07 |
| swivel | you're likely to find similar reports given the architecture, unfortunately there are *many* ways for event loop stalls to happen | 20:14 |
| sigrid | yeah | 20:15 |
| chartreuse | I had a sway issue a while back, but can't even remember what it was now, but an update fixed it | 20:15 |
| chartreuse | I guess other thing, have you updated your system recently? `apt update && apt upgrade` ? | 20:15 |
| chartreuse | If it's for sure related to sway and not something else you can always try building from source the latest version to see if anything changes | 20:16 |
| sigrid | i use void linux | 20:16 |
| sigrid | everything is latest | 20:17 |
| swivel | are there any sway logs you can look at for spam? | 20:18 |
| sigrid | yeah, sway produces loads of those in debug+verbose mode | 20:19 |
| sigrid | apart from the input stuff i mentioned there was nothing interesting though | 20:19 |
| swivel | one of the comments in that issue claimed it was caused by spamming the log, which makes sense if the log output gets blocked on writeback | 20:19 |
| sigrid | it's very quiet if i don't run it in debug mode | 20:22 |
| swivel | ah so there's no spam flood coincident with the stalls? | 20:23 |
| sigrid | nope :( | 20:24 |
| swivel | well, on the plus side you can check that off the list of suspects ;) | 20:25 |
| sigrid | hah. yeah, considering the number of moving parts in linux distros' stack, that is better than nothing | 20:37 |
| swivel | i used to experience something like this on xorg-intel on i915, caused by gpu stalls - but those were reported in dmesg, and the X architecture made it far less disruptive... just a stutter, not unintended key repeats etc | 20:41 |
| swivel | but looking at https://wayland-book.com/seat/keyboard.html it sounds like wayland clients must all implement keyboard repeating themselves, with the protocol just hinting the rate to use | 20:42 |
| swivel | so that sort of artifact could be caused by any number of things, even the client experiencing scheduling delays | 20:43 |
| swivel | it's as if wayland has gone out of its way to turn the entire desktop into a fragile low-latency video game | 20:44 |
| chartreuse | It kinda is, basically redoing the desktop to the way many modern apps are operating anyways | 20:50 |
| swivel | nice when it works, awful when it doesn't | 20:58 |
| chartreuse | Still some things I'm not a fan of, like apps not being able to bind any global key shortcuts. | 21:00 |
| chartreuse | It's going to make it so basically every Wayland WM is going to implement some API to do that anyway | 21:00 |
| chartreuse | Why not standardize an API for requesting a global key binding from an app to the WM, then the WM can choose if it wants to honour it, make a popup for the user, or what | 21:01 |
| bleb | by the way, is anyone familiar with the bugs that prevent X from running on the reform? | 21:42 |
| bleb | are there open bugs that require testing, is anyone working on it, etc. | 21:42 |
| chartreuse | Not sure if there's any bugs, might just be a need for video drivers? | 21:55 |
| - GNUmoon (QUIT: Ping timeout: 276 seconds) (~GNUmoon@gateway/tor-sasl/gnumoon) | 21:55 | |
| chartreuse | I think it was just setup for wayland by default since it'd take good advantage of the lower end video hardware | 21:55 |
| mntmn | bleb: nope, nobody is working on it. | 22:05 |
| ex-parrot | I tried looking at it but I haven't got very far yet | 22:07 |
| bleb | mntmn: do you know what the issue is? missing driver, buggy driver, etc.? | 22:09 |
| mntmn | bleb: i don't know, no. xwayland works fine which is the same codebase | 22:10 |
| chartreuse | I think it's just that no one has bothered trying | 22:10 |
| mntmn | bleb: unclear what goes wrong with modesetting or whatever | 22:10 |
| ex-parrot | I did try a bit, I got as far as adding more logging to try find where it's crashing | 22:10 |
| chartreuse | You could try just install debian's xorg-core package and go from there | 22:10 |
| mntmn | in ancient times i had it working | 22:10 |
| mntmn | then something in xorg broke it | 22:10 |
| ex-parrot | I got it to not crash, but also not draw anything :P which is kind of progress | 22:11 |
| chartreuse | And try starting with the fbdev environemnt | 22:11 |
| mntmn | and i didn't want to fix it again, as the xorg devs clearly don't care | 22:11 |
| bleb | regardless the xorg bug tracker should have a place for any reports | 22:11 |
| mntmn | tbh i never want to use xorg again | 22:11 |
| ex-parrot | I am still keen to get X working | 22:11 |
| bleb | I might actually order a reform, knowing that xwayland works. if I do I would also help out | 22:13 |
| mntmn | bleb: cool | 22:13 |
| ex-parrot | my forays in to sway and wayland haven't really been satisfying, so it would be good to have xorg working | 22:13 |
| mntmn | bleb: fwiw i am using X all the time, but just through xwayland | 22:13 |
| ex-parrot | it feels slightly gross to have to run X through an extra layer :P | 22:14 |
| * frank_ -> franks2 | 22:15 | |
| * franks2 -> frank2 | 22:15 | |
| chartreuse | xwayland works fine, the default system image even uses it for WindowMaker | 22:15 |
| bleb | yeah why wouldn't you want to run it again without the extra layer, if it were possible | 22:15 |
| ex-parrot | xorg is a bit gross | 22:16 |
| ex-parrot | but it's the gross I know :P | 22:16 |
| bleb | ex-parrot: let me know if you open a report in the xorg bugzilla | 22:17 |
| ex-parrot | sure, I need to do a bunch more work first | 22:17 |
| ex-parrot | and I' not sure when I'll have the time :( | 22:17 |
| bleb | ok | 22:18 |
| mntmn | bleb: well, running x on top of wayland gives you tear free for free, for example | 22:18 |
| sigrid | some of the X apps I run do not render correctly with Xwayland. I get black boxes instead of glyphs in rofi, and drawterm is just a black window | 22:21 |
| ex-parrot | linux desktop without tearing wouldn't feel right :P | 22:21 |
| sigrid | I assume that could be caused by not running a patched mesa? | 22:21 |
| chartreuse | You can do tear free with straight Xorg as well if the video driver supports it | 22:24 |
| ex-parrot | I've recently come around to the idea of using a compositor, because xfce terminal stopped doing transparency without it | 22:24 |
| ex-parrot | in buster I think | 22:24 |
| chartreuse | ex-parrot, without a compositor xfce-terminal would only have shown the desktop background through it | 22:25 |
| chartreuse | Not true transparancy | 22:25 |
| chartreuse | xfce has a built in compositor | 22:25 |
| ex-parrot | same same for me | 22:25 |
| ex-parrot | I just run i3 w/ xfce terminal for tabs | 22:25 |
| chartreuse | Fair yeah, it's a good terminal emulator | 22:25 |
| - frank2 (QUIT: Read error: Connection reset by peer) (~frank@juicy.frank2.net) | 22:28 | |
| + frank2 (~frank2@juicy.frank2.net) | 22:29 | |
| bleb | no drawterm is unfortunate | 22:42 |
| sigrid | wdu mean, bleb | 22:45 |
| bleb | it's unfortunate that drawterm doesn't work on the reform | 22:45 |
| sigrid | it does | 22:45 |
| sigrid | X version works fine on debian | 22:46 |
| sigrid | for wayland I got moody's fork and patched it to bring to a usable state: https://git.sr.ht/~ft/drawterm | 22:46 |
| sigrid | that's what I'm using rn | 22:46 |
| sigrid | ok, the sticky keys/mouse is unrelated to rndis | 22:47 |
| bleb | you had x working in debian on the reform? | 22:47 |
| sigrid | Xwayland | 22:47 |
| sigrid | it's working on void too, but for some reason apps do not render properly, including drawterm built with CONF=unix | 22:48 |
| bleb | so when was drawterm a black window | 22:48 |
| sigrid | when I run the CONF=unix build on void | 22:48 |
| bleb | so the debian x version worked, presumably with different build options than on void | 22:51 |
| sigrid | same build options | 22:51 |
| sigrid | mnt's debian has some packages patched, like mesa | 22:51 |
| sigrid | I suspect that makes the difference | 22:51 |
| bleb | ah I see | 22:52 |
| bleb | then is your version based on wayland instead of x? | 22:52 |
| sigrid | yep, that one works fine anywhere | 22:52 |
| sigrid | it also plays audio through pipewire | 22:52 |
| bleb | cool, thanks for the work | 22:53 |
| sigrid | it was mostly moody, I just fixed a bunch of issues :) | 22:53 |
| sigrid | this will be merged into git.9front.org drawterm eventually | 22:54 |
| bleb | but does it work on x | 22:54 |
| sigrid | if you build with CONF=unix it will, just like before | 22:55 |
| sigrid | wayland build is CONF=linux | 22:55 |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 22:56 | |
| bleb | cool | 22:56 |
| - erlehmann (QUIT: Ping timeout: 256 seconds) (~erle@ip5f5bd566.dynamic.kabel-deutschland.de) | 22:56 | |
| sigrid | good news, I found the culprit for sticky keys/mouse | 23:02 |
| sigrid | it's the damn drawterm | 23:02 |
| sigrid | how exactly does it do that to sway - idk yet, but the more rendering happening there, the more "stuck" everything becomes everywhere else | 23:03 |
| chartreuse | Weird, possibly it's using some out of date practice? Or agressively polling input or something weird | 23:04 |
| chartreuse | wouldn't have thought an x app could do that to the window manager though | 23:04 |
| sigrid | it's a wayland app | 23:05 |
| chartreuse | Oh huh, probably me looking at the wrong site showing last modified 2011 on 9p.io XD | 23:05 |
| chartreuse | Seems also to be an active fork at drawterm.9front.org | 23:06 |
| chartreuse | And a 9fans one | 23:06 |
| sigrid | yes, anything on 9p.io is insanely old | 23:06 |
| chartreuse | Seems like a bit of a mess | 23:06 |
| sigrid | 9front.org is where it's at | 23:06 |
| chartreuse | I only messed around a little with plan 9 but couldn't quite get into it | 23:07 |
| chartreuse | (Installed on my old Thinkpad 600 which is a properly supported laptop for it) | 23:07 |
| ex-parrot | 9front has a bit of a reputation / weird community issue at the moment | 23:07 |
| sigrid | I have a suspicion it might have to do with https://git.sr.ht/~ft/drawterm/tree/front/item/gui-wl/wl-util.c | 23:07 |
| sigrid | ex-parrot: like what? | 23:07 |
| ex-parrot | sigrid: the nazi imagery in the documentation, and then everyone went all weird and doubled down on it | 23:07 |
| chartreuse | (Saying like plan 9 isn't already a weird community :P) | 23:08 |
| ex-parrot | yeah :P | 23:08 |
| sigrid | ex-parrot: it was an extremely old dumb joke about ruby on rails, had nothing to do with nazis | 23:08 |
| ex-parrot | https://news.ycombinator.com/item?id=25778940 <-- some references | 23:08 |
| ex-parrot | sigrid: I see you're in the "doubling down" faction | 23:08 |
| ex-parrot | if it was an extremely old irrelevant joke you'd just... remove it | 23:08 |
| ex-parrot | and everyone would be happy | 23:08 |
| sigrid | I am merely explaining what that is, I am not saying it's great | 23:08 |
| kfx | I made the image. It has been removed. Why is there still a problem, again? | 23:09 |
| chartreuse | Only thing I see on the 9front homepage related is a crossed out swastica going to "nazi punks fuck off" | 23:09 |
| ex-parrot | oh hey kfx | 23:09 |
| sigrid | 9front has always been anti-nazi | 23:09 |
| ex-parrot | sorry, let's not re-tread this again | 23:09 |
| ex-parrot | I'm glad it's been removed | 23:09 |
| kfx | if it's because we don't like internet randos telling us how to run our website, that's not going to change. | 23:09 |
| sigrid | idk what the hell else people want | 23:09 |
| ex-parrot | disclosure: steve is kind of a friend of mine | 23:10 |
| ex-parrot | so I got a bit emotionally invested in it also | 23:10 |
| ex-parrot | and I did hang out in the 9front IRC for a bit | 23:10 |
| kfx | would you like my IMAP password and access to my unix shell, so you can validate that my soul is pure? | 23:10 |
| ex-parrot | but yeah, let's no re-hash that again | 23:10 |
| sigrid | tell him hi. he called me a nazi and linked my webpage on his post about plan 9 rio mods at the same time | 23:10 |
| sigrid | I still get hits from that post often | 23:10 |
| kfx | so do I, nothing like being called a fascist after literally going to war against fascists | 23:11 |
| ex-parrot | kfx: we can judge you on your actions probably, that seems farer :P | 23:11 |
| ex-parrot | fairer | 23:11 |
| kfx | ex-parrot: and yet here we are with "weird community issue" accusations. | 23:14 |
| kfx | surely you can see how this can be frustrating. | 23:14 |
| ex-parrot | that was my experience when I hung out in the IRC and tried 9front 3-4 years ago | 23:14 |
| ex-parrot | but I get a similar vibe off the OpenBSD IRC so it't not strictly a 9front thing either | 23:14 |
| ex-parrot | apologies for being inflammatory, I do think you folks are doing good work | 23:14 |
| kfx | we get a lot of people who join the channel and immediately demand to know where we are hiding our Chrome package | 23:15 |
| ex-parrot | :/ | 23:15 |
| kfx | as a result we have grown distrustful of new users who ask a certain kind of question which historically winds up with a confused and angry /part message | 23:15 |
| ex-parrot | yeah | 23:15 |
| ex-parrot | you guys helped me get a weird IDE controller working in an old thinkpad | 23:16 |
| ex-parrot | and I was appreciative of that | 23:16 |
| kfx | we are overjoyed to get those kinds of questions :) | 23:16 |
| ex-parrot | and I think someone said something like "you're gonna be one of these people who turns up and we do one thing then you never come back" | 23:16 |
| ex-parrot | and I guess that did happen :P | 23:16 |
| kfx | "how do I run Chrome on Plan 9" -> sadness. "how do we get this hardware working" -> at least there's a finish line! | 23:17 |
| sigrid | ex-parrot: that sounds like us | 23:17 |
| ex-parrot | I used 9front to do a "hey check out plan9" demo for my local information security interest group, it was well received | 23:17 |
| chartreuse | plan9 on reform when :P | 23:18 |
| chartreuse | Got enough mouse buttons to at least make it work | 23:18 |
| kfx | some people in our channel will put in a LOT of work and it's disheartening when the hardware owner never shows up and we never get to see if it worked | 23:18 |
| ex-parrot | there's that other raspberry pi fork right | 23:18 |
| sigrid | 9legacy with bcm support on top, iirc | 23:19 |
| sigrid | 9front runs on raspis too | 23:19 |
| sigrid | I am using one rn | 23:19 |
| ex-parrot | where is the official 9front IRC now? the oftc one seems to be dead | 23:19 |
| sigrid | oftc/#cat-v | 23:19 |
| ex-parrot | it would be cool to come back and see what's changed | 23:19 |
| ex-parrot | ah of course | 23:19 |
| sigrid | cat-v has always been the official one | 23:19 |
| sigrid | there is also 9gridchan that one could consider a non-toxic | 23:20 |
| sigrid | it has none of the -chan culture and requires 9p to chat | 23:20 |
| ex-parrot | oh cool | 23:20 |
| ex-parrot | apologies again for dredging up old crap :/ | 23:20 |
| technomancy | I always got cat-v mixed up with catb.org | 23:22 |
| ex-parrot | ooh, yeah that's a big difference | 23:22 |
| ex-parrot | I think we can all agree esr sucks right | 23:22 |
| kfx | chartreuse: the very first email I ever sent to lukas was when I saw the first Reform and begged him to include a third mouse button | 23:22 |
| kfx | ex-parrot: no argument here | 23:22 |
| technomancy | ex-parrot: I hope so | 23:22 |
| ex-parrot | phew | 23:22 |
| technomancy | the keys are right next to each other tho, at least on qwerty | 23:24 |
| chartreuse | Well you ended up getting 5 :P | 23:25 |
| ex-parrot | I re-arranged my trackball buttons to match a Thinkpad nub | 23:25 |
| chartreuse | I've been tempted by that since I keep middle clicking to scroll like on a thinkpad | 23:25 |
| ex-parrot | ^ | 23:25 |
| technomancy | this is why I don't use qwerty | 23:25 |
| ex-parrot | my hack isn't quite right yet | 23:25 |
| sigrid | I got used to the defaults | 23:25 |
| ex-parrot | but I also haven't been running X | 23:25 |
| chartreuse | Should be doable, but not entirely sure how the thinkpad does it so smoothly so that pressing and holding doesn't immediately send a middle click | 23:26 |
| chartreuse | Though for plan 9 wouldn't you need to be able to hold the middle mouse button down for some events? | 23:26 |
| sigrid | yes | 23:27 |
| chartreuse | I'm used to it enough now, but at the start I pressed middle mouse way too much when trying to scroll | 23:27 |
| kfx | libinput broke scroll wheel implementation pretty badly. with evdev you can set a timeout, so both middlemouse scroll and click-and-drag work | 23:28 |
| kfx | here is the ticket, https://gitlab.freedesktop.org/libinput/libinput/-/issues/237 in which the libinput developers completely ignore the fact that this works perfectly in evdev | 23:29 |
| + erlehmann (~erle@ip5f5bd566.dynamic.kabel-deutschland.de) | 23:30 | |
| mntmn | sigrid: if you get black boxes, that's a missing ENV var | 23:34 |
| mntmn | sigrid: you need to set ETNA_MESA_DEBUG=nir in your profile or GPU stuff will suck badly for you | 23:34 |
| mntmn | (it enables the NIR shader compiler) | 23:34 |
| mntmn | sigrid: that could also be a cause for other wonkiness :3 | 23:35 |
| chartreuse | kfx: reminds me of another bug with fonts when the gnome-terminal backend changed libraries which broke bitmap font rendering | 23:38 |
| chartreuse | And the library basically refused to even entertain that it's a bug when the library it replaces had it | 23:38 |
| kfx | yeah, they dropped freetype and use pango exclusively | 23:38 |
| chartreuse | Yeah it's pango or harfbuzz one of the two broke that | 23:39 |
| chartreuse | Oh it's both that's why | 23:39 |
| kfx | pango supports ONE very-specific weird-ass latecoming format for bitmap fonts | 23:39 |
| kfx | which 99% of bitmap font users do not use and have never heard of | 23:39 |
| chartreuse | And harfbuzz which is part of this stack doesn't support even that AFAIK | 23:40 |
| chartreuse | It even break ttf bitmap fonts which is just silly | 23:40 |
| kfx | but they went full "am I out of touch? no, it's the children who are wrong" meme | 23:40 |
| sigrid | mntmn: thanks! indeed, that helped | 23:41 |
| mntmn | sigrid: nice | 23:41 |
| sigrid | though I think quakespasm runs much slower, for some reason | 23:41 |
| sigrid | and it did not help with stuck input when running drawterm | 23:41 |
| mntmn | one day i want to finish bluerise's work and get the internal display framebuffer working in u-boot | 23:41 |
| mntmn | sigrid: ohhh? | 23:41 |
| bluerise | mntmn:I also have many one-days | 23:41 |
| kfx | mntmn: how can I help? | 23:41 |
| mntmn | sigrid: what if you unset the variable in the terminal that runs quakespasm, is it faster? | 23:41 |
| mntmn | bluerise: one day! | 23:42 |
| chartreuse | From a dev "If non-bitmap fonts are destroying my sight, then I'd fork some cash and get myself a high-dpi screen not whine here about people not doing work for me." | 23:42 |
| chartreuse | Like fixing a regression is not a reasonable request :/ | 23:42 |
| sigrid | mntmn: yes | 23:42 |
| kfx | "fork some cash" | 23:42 |
| sigrid | but transparent images (text) isn't rendered properly | 23:42 |
| sigrid | so that's the key to the issue I guess, at least one of them | 23:43 |
| mntmn | sigrid: ok that's very interesting. maybe desktop OpenGL 1 is somehow faster on the old compiler | 23:43 |
| bluerise | my SOQuartz (CM4) and Radxa CM3 (CM4) arrived in Europe. One in .NL, one in .UK. I'm sure the .NL one will arrive faster... thanks, Brexit | 23:43 |
| mntmn | kfx: figure out what part of bluerise's mipi-dsi and dcss setup code is different than what linux is doing (causing it not to display anything) | 23:44 |
| - chomwitt (QUIT: Ping timeout: 268 seconds) (~chomwitt@2a02:587:dc0b:200:12c3:7bff:fe6d:d374) | 23:44 | |
| bluerise | yup | 23:44 |
| mntmn | bluerise: i'm too dumb to find your source tree again... what was the link again? | 23:45 |
| kfx | haha I was just typing that | 23:45 |
| kfx | I found bluerise/u-boot on github, but there are several mntre branches | 23:46 |
| mntmn | kfx: hey then you are already on the right track | 23:47 |
| mntmn | i think mntre-disp2 | 23:47 |
| bluerise | https://github.com/bluerise/u-boot/commits/mntre-disp / https://github.com/bluerise/u-boot/commits/mntre-disp2 | 23:49 |
| bluerise | kfx:I just rebased it | 23:49 |
| bluerise | branch mntre is base machine + pci | 23:49 |
| bluerise | mntre-disp is my first attempt, cherry-picking a lot from boundary devices' u-boot | 23:50 |
| bluerise | mntre-disp2 are changes on top of mntre-disp | 23:50 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!