| minute | josch: you can see the screen refresh? that's not normal i'd say. normally wayland prevents all tearing | 00:19 |
|---|---|---|
| josch | yes, i see the screen refresh from top to bottom -- takes about 1 second | 00:21 |
| josch | this is most bothersome when one wants to go through several slides quickly, for example to go quickly back and forward. The new slide will start rendering before the old one is done. | 00:24 |
| josch | anways, bed for me now. Good night! | 00:25 |
| minute | good night! lets investigate it soon | 00:27 |
| - chartreus (QUIT: Quit: Leaving) (~chartreus@S0106f0f249dfd9c3.cg.shawcable.net) | 00:28 | |
| + chartreuse (~chartreus@S0106f0f249dfd9c3.cg.shawcable.net) | 00:28 | |
| - mtm (QUIT: Ping timeout: 248 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 01:03 | |
| vagrantc | yay, MNT/Reform firmwares are still reproducible as of e5fbd4ed60b67be9c93ce6b1142e958599bb772b ... now I need to test them and make sure they don't catch fire or something | 02:17 |
| vagrantc | though, probably not for another day at least | 02:17 |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 02:26 | |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 03:10 | |
| - klardotsh (QUIT: Quit: nyaa~) (~klardotsh@172.56.105.248) | 03:47 | |
| + bgs (~bgs@212-85-160-171.dynamic.telemach.net) | 07:23 | |
| - bgs (QUIT: Remote host closed the connection) (~bgs@212-85-160-171.dynamic.telemach.net) | 08:28 | |
| - blast007 (QUIT: Remote host closed the connection) (~blast@user/blast007) | 08:55 | |
| + blast007 (~blast@user/blast007) | 08:56 | |
| - swivel (QUIT: *.net *.split) (~swivel@shells.gnugeneration.com) | 09:00 | |
| - eschaton (QUIT: *.net *.split) (~eschaton@li541-49.members.linode.com) | 09:00 | |
| - jnerula (QUIT: *.net *.split) (~jnerula@li1009-93.members.linode.com) | 09:00 | |
| - yankcrime (QUIT: *.net *.split) (~nick@agrajag.dischord.org) | 09:00 | |
| - lastebil (QUIT: *.net *.split) (truck@shell.suomiscene.fi) | 09:00 | |
| - sigrid (QUIT: *.net *.split) (~sigrid@ftrv.se) | 09:00 | |
| - skalk (QUIT: *.net *.split) (~skalk@vond.sysret.de) | 09:00 | |
| - doppler (QUIT: *.net *.split) (~doppler@user/doppler) | 09:00 | |
| - eery (QUIT: *.net *.split) (~eery@172.97.103.152) | 09:00 | |
| - plomlompom (QUIT: *.net *.split) (~plom@play.plomlompom.com) | 09:00 | |
| - ysionneau (QUIT: *.net *.split) (~fallen@tools.sionneau.net) | 09:00 | |
| - _E (QUIT: *.net *.split) (~e@bsd.moe) | 09:00 | |
| + eery_ (~eery@172.97.103.152) | 09:00 | |
| + swivel (~swivel@shells.gnugeneration.com) | 09:00 | |
| + sigrid (~sigrid@ftrv.se) | 09:00 | |
| + yankcrim- (~nick@agrajag.dischord.org) | 09:00 | |
| + skalk (~skalk@vond.sysret.de) | 09:00 | |
| + lastebil (truck@shell.suomiscene.fi) | 09:00 | |
| + _E (~e@bsd.moe) | 09:00 | |
| + jnerula (~jnerula@li1009-93.members.linode.com) | 09:01 | |
| + eschaton (eschaton@2600:3c01::f03c:91ff:fefd:5d92) | 09:02 | |
| + plomlompom (~plom@play.plomlompom.com) | 09:03 | |
| + ysionneau (~fallen@tools.sionneau.net) | 09:03 | |
| + doppler (~doppler@user/doppler) | 09:04 | |
| - chomwitt (QUIT: Remote host closed the connection) (~chomwitt@2a02:587:7a0a:c00:1ac0:4dff:fedb:a3f1) | 09:04 | |
| + chomwitt (~chomwitt@2a02:587:7a0a:c00:1ac0:4dff:fedb:a3f1) | 09:11 | |
| + MajorBiscuit (~MajorBisc@c-001-030-015.client.tudelft.eduvpn.nl) | 09:20 | |
| - MajorBiscuit (QUIT: Quit: WeeChat 3.6) (~MajorBisc@c-001-030-015.client.tudelft.eduvpn.nl) | 09:26 | |
| + MajorBiscuit (~MajorBisc@c-001-030-015.client.tudelft.eduvpn.nl) | 09:35 | |
| - mjw (QUIT: Killed (NickServ (GHOST command used by wielaard!~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440))) (~mark@gnu.wildebeest.org) | 10:10 | |
| * wielaard -> mjw | 10:10 | |
| + mark_ (~mark@gnu.wildebeest.org) | 10:11 | |
| * yankcrim- -> yankcrime | 11:30 | |
| vkoskiv | josch: https://pastebin.com/D1Nk39DK | 12:31 |
| josch | vkoskiv: when you switch on your reform, there is no sd-card inserted, correct? | 12:36 |
| vkoskiv | Yes | 12:36 |
| vkoskiv | my /boot should be on emmc | 12:37 |
| josch | yes | 12:37 |
| josch | what is the content of your /boot? | 12:38 |
| vkoskiv | vkoskiv@reform:~$ ls /boot/ | 12:40 |
| vkoskiv | boot.scr boot.scr.bak config-5.19.0-reform2-arm64 dtb dtb-5.19.0-reform2-arm64 dtbs flash.bin initrd.img-5.19.0-reform2-arm64 lost+found System.map-5.19.0-reform2-arm64 vmlinuz-5.19.0-reform2-arm64 | 12:40 |
| vkoskiv | No 6.x here. | 12:40 |
| josch | aha | 12:40 |
| josch | vkoskiv: then maybe at the time you upgraded your /boot wasn't mounted correctly? | 12:41 |
| vkoskiv | It's possible. Do I just reinstall the linux image package, whatever the name of it was? | 12:41 |
| josch | maybe it's enough to upgrade? we recently had the release of 6.0.7 | 12:41 |
| vkoskiv | testing | 12:42 |
| vkoskiv | Ran an update + upgrade. Haven't rebooted, but I'm not seeing 6.x under /boot still | 12:46 |
| vkoskiv | This happened a day or two ago as well. I know /boot is mounted correctly | 12:46 |
| josch | can you show me the upgrade log? | 12:46 |
| vkoskiv | Yeah, sec | 12:46 |
| vkoskiv | https://pastebin.com/xr9pYTKR | 12:47 |
| josch | we are getting somewhere | 12:49 |
| josch | now the output of this please: apt-cache policy linux-image-arm64 | 12:50 |
| vkoskiv | josch: https://pastebin.com/fSKUYZKB | 12:50 |
| vkoskiv | (still haven't rebooted) | 12:51 |
| josch | there is your problem | 12:51 |
| josch | this probably did not start as a sysimage-v3? | 12:51 |
| vkoskiv | It did | 12:51 |
| vkoskiv | In june I started my setup with sysv3, though it was an early version, I guess | 12:51 |
| josch | then at some point you removed the package linux-image-arm64 -- can i have the output of dpkg -l linux-image-arm64 | 12:52 |
| vkoskiv | I'm not sure I understand the output of this cmd. It says this package is not installed, but it is? | 12:52 |
| josch | it should be a single line so you can paste it here | 12:52 |
| vkoskiv | josch: https://pastebin.com/4vRGCKW0 | 12:53 |
| josch | it says: rc linux-image-arm64 | 12:54 |
| josch | the "r" means that it was installed but then got removed | 12:54 |
| josch | so your problem is, that you at some point removed that package | 12:54 |
| vkoskiv | Yeah, I think I did remove this package when I was trying to see if downgrading to 5.18 would fix the SSD issue I'm having | 12:54 |
| josch | but that package is the one that, when updated, points to the newest kernel package version | 12:54 |
| josch | so just re-install it | 12:54 |
| vkoskiv | re-install linux-image-arm64 with apt? | 12:55 |
| josch | sudo apt install linux-image-arm64 | 12:55 |
| josch | that will pull the latest kernel, replacing your old kernel | 12:55 |
| vkoskiv | It seems to be installing 6.x now! | 12:55 |
| josch | good :) | 12:55 |
| vkoskiv | bigger version number good | 12:55 |
| josch | haha | 12:56 |
| josch | wait for 6.1 | 12:56 |
| vkoskiv | As usual, thanks so much for the help. Much appreciated | 12:56 |
| vkoskiv | And I learned new commands again | 12:56 |
| josch | in 6.1 we get hardware accelerated x265 10bit decoding with hantro g2 :) | 12:56 |
| vkoskiv | That's exciting! Is that something that will work in browsers too? | 12:56 |
| vkoskiv | I've been using mpv, it works nicely | 12:57 |
| josch | it can be used by browsers but i'm not using browsers to watch video | 12:57 |
| josch | mpv is not hardware accelerated because ffmpeg isn't (yet) | 12:57 |
| vkoskiv | Oh. Weird, I guess the CPU is just really good at decoding video then | 12:57 |
| vkoskiv | I was observing really low CPU use when watching youtube with mpv | 12:57 |
| josch | until ffmpeg can make use of hantro, clapper can be used to decode 1080p h264 video at 60fps with very little cpu usage (around 11% in my tests) | 12:57 |
| josch | vkoskiv: oh? i have the opposite | 12:58 |
| josch | when i try to watch youtube videos with mpv it's a slideshow | 12:58 |
| vkoskiv | Smooth for me, but I misremembered - 150+% CPU use | 12:59 |
| josch | vkoskiv: are your youtube videos 1080p? | 12:59 |
| vkoskiv | Yeah, 1080p profile | 12:59 |
| josch | interesting | 12:59 |
| josch | i wonder why that is | 12:59 |
| vkoskiv | rebooting to 6.x now | 12:59 |
| josch | vkoskiv: can you try this: | 13:00 |
| josch | vkoskiv: mpv https://www.youtube.com/watch?v=rk0MTdp0law | 13:00 |
| josch | very choppy for me -- it plays back at 1080p | 13:00 |
| vkoskiv | Linux reform 6.0.0-reform2-arm64 #1 SMP Debian 6.0.7-1+reform20221110T055858Z1 (2022-11-10) aarch64 GNU/Linux | 13:00 |
| vkoskiv | Yay! | 13:00 |
| josch | nice :) | 13:00 |
| vkoskiv | This video is choppy for me too | 13:01 |
| vkoskiv | Maybe I was just watching low bitrate videos | 13:01 |
| josch | aha! | 13:01 |
| vkoskiv | I have a 1080p profile that I select in mpv | 13:01 |
| josch | me too | 13:01 |
| josch | what's yours? | 13:01 |
| vkoskiv | Actually, switching to fullscreen makes it smooth | 13:02 |
| josch | ytdl-format=bestvideo[height<=?1080]+bestaudio/best | 13:02 |
| vkoskiv | smooth(er) | 13:02 |
| vkoskiv | Still choppy | 13:02 |
| vkoskiv | Yeah, my profile is similar | 13:02 |
| vkoskiv | low motion moments in the video are smoother, as expected | 13:02 |
| josch | yes fullscreen is smoother -- you can press "I" to see the dropped frame count | 13:02 |
| vkoskiv | Good thing this is my mobile workstation, and not a media consumption device :D | 13:03 |
| vkoskiv | The less than ideal YouTube situation has been a feature, rather than a bug so far. | 13:03 |
| - mtm (QUIT: Ping timeout: 248 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 13:03 | |
| josch | then better don't install clapper because that fixes the problem ;) | 13:03 |
| vkoskiv | I just recently discovered Helix, I got it compiled for Reform, and it's an amazing editor. | 13:03 |
| vkoskiv | I've used vim for over 10 years, and just a few minutes of helix convinced me to switch | 13:04 |
| josch | "No Electron. No VimScript. No JavaScript" -- oh sounds indeed nice | 13:10 |
| vkoskiv | Seems damn fast to me, so far. | 13:13 |
| vkoskiv | My background is from trying to have a very minimal (50 lines) vim config | 13:13 |
| vkoskiv | And being frustrated at how much work setting up more advanced higlighting/LSP/jumps in vim is | 13:14 |
| vkoskiv | Literally the last week or so I've been thinking about this a lot, looking at alternatives | 13:14 |
| vkoskiv | Then I fire up helix, and it's like someone read my thoughts, made an editor exactly how I have been wanting, and then made it even better. | 13:15 |
| vkoskiv | I've been very happy for the last two days, reading the tutorial and getting used to the differences this has to vim :) | 13:15 |
| vkoskiv | soon, maybe today, alias vim='helix' | 13:15 |
| minute | i've also tried helix recently | 13:32 |
| minute | i'm still too deep into emacs unfortunately... sometimes would like a more modern replacement | 13:32 |
| josch | i'm deep into vim but i feel its limitations -- helix looks useful, i'll check it out, thanks! | 13:34 |
| vkoskiv | The tutor didn't make it obvious, but try hitting space, g, z, and C-w | 14:14 |
| vkoskiv | All those open little windows that show you the things you can do | 14:14 |
| vkoskiv | gd => go to definition. Helix just works with clangd right away, I just need to have a compile_commands.json thing and it picks that up | 14:14 |
| vkoskiv | It *just works*. I love it. | 14:14 |
| vkoskiv | Also, I found 'bear' to work great to generate that compile_commands.json file for projects that just use make | 14:15 |
| vkoskiv | Otherwise CMake has a flag to export that file. | 14:15 |
| minute | vkoskiv: i wonder if it would work with linux kernel codebase? afaik there is a script included to create compile_commands.json | 14:21 |
| vkoskiv | I will actually try that after work and report back. Would be very interesting to see. | 14:22 |
| vkoskiv | For me, a perfectly working "jump to where this thing was declared" is what I need. | 14:24 |
| vkoskiv | ctags was always very meh. It just looked for keywords and sent you to the first one it found. It doesn't understand code. | 14:24 |
| minute | yeah, i feel that | 14:24 |
| vkoskiv | I did consider writing a program that would let you emit a smarter tags file using more modern infrastructure, like clangd/tree-sitter | 14:24 |
| vkoskiv | Then you could still use the nice native tags support in vim, but have it be more useful | 14:25 |
| minute | i work with eclipse sometimes in the form of vivado sdk, and i miss those features when i'm back in emacs land. even if it is possible to integrate lsp, somehow i always hat problems to get it to work | 14:25 |
| vkoskiv | As far as I know, that tool does not exist yet. | 14:25 |
| minute | did you check out nvim, it has more lsp support, no? | 14:25 |
| vkoskiv | I always had problems configuring IDE like stuff into vim. I tried many times. Most times I couldn't get it to work at all, and when I did, it was janky | 14:25 |
| vkoskiv | I did try nvim too, but it just didn't work out. | 14:25 |
| vkoskiv | Forget why | 14:25 |
| minute | s/hat/had | 14:26 |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 15:09 | |
| minute | a big change to etnaviv was merged 6 days ago https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9780 | 15:27 |
| minute | this should improve speed but also could break some things, i'll build it and try it on my own system | 15:27 |
| minute | also pulling in !19582 and !19571 | 15:33 |
| minute | as expected, !9780 breaks all xwayland graphics | 16:15 |
| josch | minute: once you have a patch that works, since we can now fall back to qemu-emulated "native" compilation, we can patch mesa again :) | 16:19 |
| josch | i just removed the mesa patch from reform-debian-packages because the new mesa release includes the patch we had applied | 16:20 |
| minute | cool | 16:22 |
| minute | it appears we need an xserver/xwayland with this https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/936 | 16:27 |
| minute | (which is merged in xserver master) | 16:27 |
| minute | i will now build the latest xserver/xwayland. | 16:28 |
| minute | i can say that firefox with MOZ_ENABLE_WAYLAND=1 works fine with latest mesa | 16:28 |
| minute | chromium does not work anymore in ozone mode, though (but then, it was always a bit buggy) | 16:29 |
| minute | josch: latest xserver/xwayland from git fixes the problems with mesa !9780. | 16:37 |
| josch | if there is a specific commit of xwayland that makes it work again, then we can backport that and build both mesa and xwayland patched | 16:39 |
| minute | weird btw, but i think longer standing issue: on HDMI 1920x1080 (dcss), this has around 30fps, but on eDP 1920x1080 (lcdif), this has 40+fps https://webglsamples.org/blob/blob.html | 16:39 |
| minute | josch: we build xwayland patched anyway, or not? | 16:39 |
| josch | yes, we are already patching xwayland | 16:39 |
| minute | yep, so we just need to make sure it's up-to-date with latest xserver master | 16:40 |
| minute | then all the latest mesa changes will be fine | 16:40 |
| josch | we build xwayland from debian with a single patch on top | 16:40 |
| minute | ahhh. sorry | 16:40 |
| josch | we are not building xwayland from git | 16:40 |
| minute | yes, i was confused. so the patch is https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/936 | 16:40 |
| josch | nice, let me backport that to the version packaged in debian | 16:41 |
| minute | the change is rather small, so hopefully not an issue | 16:41 |
| minute | interesting, i found a faulty (with new mesa) wayland app... wdisplays, unless "show screen contents" is turned off | 16:42 |
| + bkeys (~Thunderbi@2600:100e:be1c:48a:9329:17b8:a23f:1325) | 16:45 | |
| josch | minute: the xwayland patch doesn't require patched mesa, right? Because then I'll just be pushing my rebased patch now and rebuild xwayland and then you can try it out. | 16:53 |
| - bkeys (QUIT: Ping timeout: 260 seconds) (~Thunderbi@2600:100e:be1c:48a:9329:17b8:a23f:1325) | 16:54 | |
| minute | josch: yep, it does not | 16:54 |
| josch | minute: pushed. If it compiles, it will be done in about 2 hours. | 16:56 |
| sknebel | finally getting around to trying the v3 system image. just grab the latest sucessful build from the pipeline, or did I miss a "recommended" starting point somewhere? | 16:59 |
| josch | sknebel: nope, that should work | 16:59 |
| josch | sknebel: report bugs to me please :) | 16:59 |
| sknebel | thx | 17:00 |
| sknebel | will do :) | 17:00 |
| minute | funfact: if you use HDMI with 50hz instead of 60hz, you can get higher fps | 17:02 |
| + bkeys (~Thunderbi@2600:100e:be1a:1cfb:533f:4180:1b4d:718b) | 17:08 | |
| minute | with the latest stack, clapper still can show movies with 0 cpu usage. | 17:08 |
| josch | yes!!! it's so sick :D | 17:09 |
| josch | i'm in the process of packaging gtuber which is youtube-dl for gstreamer -- with that you can watch youtube and twitch with 0 cpu usage :D | 17:10 |
| - bkeys (QUIT: Ping timeout: 252 seconds) (~Thunderbi@2600:100e:be1a:1cfb:533f:4180:1b4d:718b) | 17:28 | |
| sknebel | ok, the early install went a bit wonky because it spewed kernel output through the debian prompts (for hostname, locale and such), but other than that went fine | 17:58 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 18:00 | |
| - mjw (QUIT: Killed (NickServ (GHOST command used by mark_!~mark@gnu.wildebeest.org))) (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 18:30 | |
| * mark_ -> mjw | 18:30 | |
| + wielaard (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 18:31 | |
| + bgs (~bgs@212-85-160-171.dynamic.telemach.net) | 18:40 | |
| - MajorBiscuit (QUIT: Ping timeout: 252 seconds) (~MajorBisc@c-001-030-015.client.tudelft.eduvpn.nl) | 18:51 | |
| josch | minute: pipeline passed and repo is updated with patched xwayland | 19:01 |
| minute | josch: cool! | 19:01 |
| minute | thanks | 19:01 |
| minute | josch: btw mesa master has more problems that shows as corrupted textures when chromium draws SVGs, i've posted this in #etnaviv on oftc | 19:02 |
| sknebel | huh, weird issue with packages. apt update runs into hash errors with files that are fetched | 19:10 |
| sknebel | not sure if thats a "I get pointed to a broken debian mirror" error or what is happening | 19:10 |
| sknebel | (quite possibly not a reform-specific issue, but ...) | 19:11 |
| josch | sknebel: I need your terminal output to tell you more. | 19:21 |
| sknebel | it just resolved itself, so I'm blaming a hiccup in the mirrors :D | 19:23 |
| sknebel | from some online reading the error fits what can happen if a mirror is used in the middle of updating - files on disk are already new, index still points to old hashes or vice verse | 19:24 |
| josch | yes, that's correct | 19:29 |
| josch | it's also possible because we just had a mirror push to the reform mirror because i ran the reform-debian-packages pipeline | 19:29 |
| josch | minute: whoops... sounds like we indeed do not want git master but just cherry-pick the commit we need instead | 19:30 |
| josch | (plus of course figuring out which commit broke chromium...) | 19:30 |
| - bgs (QUIT: Remote host closed the connection) (~bgs@212-85-160-171.dynamic.telemach.net) | 19:35 | |
| - chartreuse (QUIT: Remote host closed the connection) (~chartreus@S0106f0f249dfd9c3.cg.shawcable.net) | 20:00 | |
| - sbp (QUIT: Ping timeout: 255 seconds) (~sbp@apache/doge/sbp) | 20:09 | |
| + sbp (~sbp@2001:19f0:6800:1102:5400:ff:fe11:39a1) | 20:14 | |
| - sbp (QUIT: Changing host) (~sbp@2001:19f0:6800:1102:5400:ff:fe11:39a1) | 20:15 | |
| + sbp (~sbp@apache/doge/sbp) | 20:15 | |
| + ajr (~ajr@user/ajr) | 20:29 | |
| + bgs (~bgs@212-85-160-171.dynamic.telemach.net) | 23:19 | |
| - bgs (QUIT: Remote host closed the connection) (~bgs@212-85-160-171.dynamic.telemach.net) | 23:25 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 23:40 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!