| - chomwitt (QUIT: Ping timeout: 264 seconds) (~chomwitt@2a02:85f:9a5f:900:1ac0:4dff:fedb:a3f1) | 00:15 | |
| + Murphy (~Murphy@76-223-253-78.lightspeed.frokca.sbcglobal.net) | 00:25 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 00:27 | |
| - Murphy (QUIT: Ping timeout: 250 seconds) (~Murphy@76-223-253-78.lightspeed.frokca.sbcglobal.net) | 00:39 | |
| - TechnoWizard (QUIT: Remote host closed the connection) (~TechnoWiz@user/TechnoWizard) | 00:40 | |
| + paperManu_ (~paperManu@107.159.15.124) | 00:51 | |
| + TechnoWizard (~TechnoWiz@user/TechnoWizard) | 01:08 | |
| + chrcav (~chrcav@user/chrcav) | 01:50 | |
| - mjw (QUIT: Ping timeout: 250 seconds) (~mjw@gnu.wildebeest.org) | 02:03 | |
| - paperManu_ (QUIT: Ping timeout: 246 seconds) (~paperManu@107.159.15.124) | 02:11 | |
| - chrcav (QUIT: Ping timeout: 260 seconds) (~chrcav@user/chrcav) | 02:28 | |
| + chrcav (~chrcav@user/chrcav) | 02:29 | |
| - paperManu (QUIT: Ping timeout: 260 seconds) (~paperManu@107.159.15.124) | 02:51 | |
| + paperManu (~paperManu@107.159.15.124) | 02:56 | |
| - rwa_ (QUIT: *.net *.split) (0a82deb4eb@2a03:6000:1812:100::41b) | 03:01 | |
| - jn (QUIT: *.net *.split) (~quassel@user/jn/x-3390946) | 03:01 | |
| - qbit (QUIT: *.net *.split) (~qbit@user/qbit) | 03:01 | |
| + rwa_ (0a82deb4eb@2a03:6000:1812:100::41b) | 03:07 | |
| + jn (~quassel@user/jn/x-3390946) | 03:07 | |
| + qbit (~qbit@user/qbit) | 03:07 | |
| - Ar|stote|is (QUIT: Ping timeout: 244 seconds) (~linx@149.210.67.77) | 03:58 | |
| - jacqueline (QUIT: Ping timeout: 244 seconds) (~jacquelin@user/jacqueline) | 03:59 | |
| - TechnoWizard (QUIT: Remote host closed the connection) (~TechnoWiz@user/TechnoWizard) | 04:03 | |
| - paperManu (QUIT: Ping timeout: 244 seconds) (~paperManu@107.159.15.124) | 04:25 | |
| + TechnoWizard (~TechnoWiz@user/TechnoWizard) | 04:51 | |
| - TechnoWizard (QUIT: Remote host closed the connection) (~TechnoWiz@user/TechnoWizard) | 04:52 | |
| + jogu (~jogu@user/jogu) | 05:05 | |
| - jogu (QUIT: Remote host closed the connection) (~jogu@user/jogu) | 05:42 | |
| josch | minute: i just had a nightmare that you installed phoenix and lost your browser history -- please do not install it anywhere where you could loose important data XD | 07:01 |
|---|---|---|
| + chomwitt (~chomwitt@2a02:85f:9a5f:900:1ac0:4dff:fedb:a3f1) | 07:14 | |
| + jacqueline (~jacquelin@user/jacqueline) | 07:38 | |
| + gidzit (~gidzit@gidzit.org) | 09:03 | |
| grimmware | I’m very interested in playing around with Glide some more at some point soon. It’s a vimium/vimperator equivalent browser based on Firefox but configurable and automateable in typescript. I especially like that you can use the same hints-style interface on the native browser menus. | 09:04 |
| rick_ | Oh we have special stuff for firefox? I didn't know. I mainly use librewolf. Ill look into these changes and also set them in librewolf ^^ | 09:30 |
| - iank_ (QUIT: Ping timeout: 244 seconds) (~iank@fsf/staff/iank) | 10:00 | |
| - cow321 (QUIT: Ping timeout: 244 seconds) (~deflated8@user/meow/deflated8837) | 10:57 | |
| + cow321 (~deflated8@user/meow/deflated8837) | 11:09 | |
| minute | josch: wow :D | 11:19 |
| amospalla | The new "MNT Pocket Reform Headset/Switch Board 2.0" standby switch can be toggled by hand? | 12:07 |
| - chomwitt (QUIT: Ping timeout: 244 seconds) (~chomwitt@2a02:85f:9a5f:900:1ac0:4dff:fedb:a3f1) | 12:08 | |
| amospalla | I received the previus RCORE Pocket upgrade, the one without it. If I buy the new one, can I connect it to my board? I would not mind the wifi card slot not working, I just want to be able to toggle power on and off by hand. | 12:08 |
| + mjw (~mjw@gnu.wildebeest.org) | 12:08 | |
| + paperManu (~paperManu@107.159.15.124) | 12:40 | |
| + gustav25 (~gustav@c-78-82-52-240.bbcust.telenor.se) | 13:02 | |
| - erle (QUIT: Quit: K-lined) (~erle@user/erle) | 13:10 | |
| grimmware | https://lobste.rs/s/wjs6o2/what_laptops_are_lobsters_rocking_2025 is interesting to ctrl-f for MNT through. | 13:47 |
| cwebber | yeah! | 13:48 |
| cwebber | it's also interesting that the macbook air machines have more or less normalized running aarch64 for developers | 13:48 |
| cwebber | so that also seems to mean people will be more open to the Reform Next | 13:48 |
| - paperManu (QUIT: Ping timeout: 260 seconds) (~paperManu@107.159.15.124) | 14:13 | |
| + paperManu (~paperManu@modemcable141.205-200-24.mc.videotron.ca) | 14:45 | |
| minute | amospalla: what do you mean by hand? it still requires a little sharp object to toggle. except if you do a mod like the 3d printed cover/switch cap that debaer and andy piper have | 15:01 |
| amospalla | minute: oh right, I did not read the announcement correctly. The standby switch is the same as the previous one. | 15:02 |
| amospalla | Thank you. | 15:02 |
| minute | amospalla: yeah! | 15:02 |
| amospalla | I understood another thing, my brain translates english sometimes freely. | 15:03 |
| minute | ok, back to building reform-system-image :) | 15:04 |
| + pomel0 (~pomel0@user/pomel0) | 15:13 | |
| - pomel0 (QUIT: Ping timeout: 264 seconds) (~pomel0@user/pomel0) | 15:25 | |
| + pomel0 (~pomel0@user/pomel0) | 15:36 | |
| josch | same here | 15:43 |
| josch | i see that the icons of gnome and sway are again in their correct size :) | 15:43 |
| minute | josch: ah you mean in the wizard. yeah i tried to fix that in the code :D | 15:44 |
| josch | yeah and it's fixed, nice :) | 15:45 |
| + jogu (~jogu@user/jogu) | 15:52 | |
| minute | jogu: gnome and sway both LGTM :3 and no more weird confusing terminal output between wizard + gdm | 15:54 |
| josch | minute: you are testing the artifacts of job 15265 I assume? Just making sure we are both running the same things. | 15:56 |
| minute | josch: yesss | 16:16 |
| minute | josch: i wanted to say the image is good but i don't think i heard the voiceover announcement during the setup wizard. will retest that | 16:21 |
| josch | oh i did hear it | 16:22 |
| minute | josch: ahh ok. probably it was just too loud here because there was construction going on at the time | 16:22 |
| minute | i need to test if the orca shortcut actually works though | 16:22 |
| josch | Yes, i tried again. Screenreader is there. | 16:24 |
| josch | minute: what does not work is super+enter -- that should give me a terminal when the setup-wizard is running, no? | 16:24 |
| josch | oh i think you changed the terminal shortcuts, maybe something broke in the process | 16:24 |
| minute | josch: did i change the terminal shortcut in setup wizard sway config? | 16:27 |
| minute | josch: i think it's because we don't ship foot anymore perhaps | 16:28 |
| josch | bindsym $mod+Return exec foot | 16:28 |
| josch | oops | 16:28 |
| josch | yes | 16:28 |
| minute | ah yeah. that can probably be changed now to xdg-terminal-exec, i'll test | 16:29 |
| minute | hmm, the screenreader shortcut doesn't work | 16:42 |
| minute | because there is no shortcut defined in reform-setup-sway-config... wat | 16:43 |
| minute | looks like i didn't commit that ;/ | 16:44 |
| minute | but so weird, i was sure that i had a branch with that working | 16:44 |
| + iank (~iank@fsf/staff/iank) | 16:46 | |
| minute | uff, conflicts when i try to merge main into my branch | 16:48 |
| minute | ah josch you did the commit https://source.mnt.re/reform/reform-system-image/-/blob/main/reform-setup-sway-config?ref_type=heads | 16:48 |
| minute | but it's incomplete | 16:48 |
| - pomel0 (QUIT: Ping timeout: 240 seconds) (~pomel0@user/pomel0) | 16:50 | |
| minute | i'll restore the missing stuff | 16:50 |
| - iank (QUIT: Quit: ZNC 1.8.2+deb2+deb11u1 - https://znc.in) (~iank@fsf/staff/iank) | 16:51 | |
| + iank (~iank@fsf/staff/iank) | 17:00 | |
| minute | haha, now builds are not working anymore because of debian unstable foo | 17:04 |
| minute | so we're gonna patch that file during install | 17:04 |
| bremner | ACTION sobs | 17:05 |
| bremner | ;) | 17:05 |
| - bkeys (QUIT: Remote host closed the connection) (~Thunderbi@h193.131.19.98.dynamic.ip.windstream.net) | 17:08 | |
| minute | for later reference https://source.mnt.re/reform/reform-factory/-/commit/7de5103de34bc1c0f560e6ba976697d6d97aceae#d299e1acb6ad5c94598c6b489a468caaee0e977f_60_62 | 17:14 |
| + pomel0 (~pomel0@user/pomel0) | 17:25 | |
| minute | ok that worked, first one installed | 17:26 |
| josch | minute: since the pipeline worked a few hours ago, another hack to make it work would be to use snapshot.debian.org -- but we can look into that another time | 17:31 |
| minute | josch: yeah, we have a working setup now, rcores are being installed | 17:32 |
| josch | nice, happy flashing :) | 17:32 |
| minute | thx!! | 17:33 |
| josch | minute: i'd like to preserve known working system images semi-regularly, maybe every 1-2 months? Do you think it's a good idea if i click later (somehow gitlab does not show the button now) on the "keep" button for the reform-system-image job? | 17:33 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 17:44 | |
| minute | josch: yeah i clicked it for the one we're deploying | 17:53 |
| josch | wonderful, thank you! | 17:53 |
| josch | then i'll add that job as another entry to the README.md | 17:53 |
| + istar (~istar@193.62.199.150) | 17:56 | |
| + erle (~erle@user/erle) | 18:15 | |
| - plomlompom (QUIT: Quit: ircplom says bye) (~plom@139.59.138.73) | 18:16 | |
| + plomlompom (~plom@139.59.138.73) | 18:17 | |
| - istar (QUIT: Ping timeout: 240 seconds) (~istar@193.62.199.150) | 18:22 | |
| minute | josch: btw, thinking if we should enable "splash" for rk3588........ | 18:27 |
| - mjw (QUIT: Ping timeout: 246 seconds) (~mjw@gnu.wildebeest.org) | 18:30 | |
| + mjw (~mjw@2001:1c06:2486:a800:158:adb1:a1e7:3c3f) | 18:40 | |
| + chomwitt (~chomwitt@2a02:85f:9a5f:900:1ac0:4dff:fedb:a3f1) | 18:40 | |
| minute | now trying to rebase my gpu patches on 6.17 | 18:55 |
| + istar (~istar@mail.cbreid.com) | 18:55 | |
| - gidzit (QUIT: Ping timeout: 246 seconds) (~gidzit@gidzit.org) | 19:12 | |
| gordon2 | oh i see patches6.17 got merged to reform-debian-packages, are they ready to use? | 19:16 |
| gordon2 | >i'm gonna comment this out of the device tree in any case, because if the gpio is not initialized by system controller, i'm not sure about the initial state | 19:16 |
| gordon2 | minute: may i kindly ask to elaborate? | 19:17 |
| gordon2 | do weird things happen when you just connect the battery and LPC inits? | 19:17 |
| minute | ah, egpu patchset just had a little "delete empty line" error | 19:24 |
| minute | gordon2: if you don't have the sysctl firmware managing the "wake" gpio pin, the gpio value can be random | 19:24 |
| minute | gordon2: and the .dts change i did seems to make a "space" input into the kbd event system if the gpio is high | 19:25 |
| gordon2 | what do you mean by "managing"? isn't stock firmware managing it? | 19:29 |
| gordon2 | sorry, i am confused | 19:31 |
| - istar (QUIT: Ping timeout: 244 seconds) (~istar@mail.cbreid.com) | 19:32 | |
| gordon2 | i assume keyboard pokes wake pin when i do circle+space, and wake pin supposed to be input on the rk3588 side | 19:32 |
| gordon2 | is it about pullup? | 19:32 |
| gordon2 | or am i missing something | 19:33 |
| gordon2 | ACTION pulling up the schematics | 19:33 |
| - mjw (QUIT: Ping timeout: 244 seconds) (~mjw@2001:1c06:2486:a800:158:adb1:a1e7:3c3f) | 19:34 | |
| minute | gordon2: sysctl pokes wake pin | 19:41 |
| minute | gordon2: but only if you have sysctl version that sets this up | 19:41 |
| minute | gordon2: which is a very recent one | 19:42 |
| minute | (on pocket reform) | 19:42 |
| minute | gordon2: i think lpc on classic reform already did this | 19:42 |
| minute | gordon2: keyboard sends wake command to sysctl, sysctl pulses that gpio. but only if sysctl has an implementation for that. which it didn't have on pocket for a longer time. | 19:42 |
| + Ar|stote|is (~linx@149.210.67.77) | 19:44 | |
| gordon2 | oh yes right | 19:49 |
| gordon2 | there's more to that | 19:49 |
| gordon2 | but you commented out this block in the .dts for classic reform | 19:49 |
| gordon2 | so i guess nothing bad should happen if i uncomment this? | 19:50 |
| minute | gordon2: i didn't test. check lpc code | 19:53 |
| minute | gordon2: probably just the "autorepeat" can be removed | 19:54 |
| minute | gordon2: from dts | 19:54 |
| gordon2 | cool, will give it a try, as far as i understand s2idle should just work now? | 19:55 |
| gordon2 | with 6.17 (i'm just gonna test both 6.17 and wakeup since i have my hands on it now) | 19:55 |
| - Ar|stote|is (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~linx@149.210.67.77) | 19:59 | |
| + Ar|stote|is (~linx@149.210.67.77) | 19:59 | |
| minute | gordon2: no, i haven't tested on modular kernel, so not sure if it'll work. i only tested on 6.16 monolithic | 20:08 |
| gordon2 | $ lsmod|wc -l | 20:09 |
| gordon2 | 11 | 20:09 |
| gordon2 | is this monolithic enough? :D | 20:09 |
| gordon2 | or do you mean from initrd | 20:09 |
| - digitalrane (QUIT: Ping timeout: 240 seconds) (~digitalra@user/digitalrane) | 20:29 | |
| cwebber | hi! I have questions! but they're easy ones! | 20:35 |
| cwebber | I'm installing the new battery socket things in the reform! | 20:35 |
| cwebber | but it has this yellow tape at the bottom of it. But the original parts do too! so I assume that yellow tape is important | 20:35 |
| cwebber | the problem is | 20:35 |
| cwebber | the yellow tape, unlike in the original | 20:35 |
| cwebber | covers the screw holes! | 20:35 |
| cwebber | what should I do about this? | 20:36 |
| cwebber | is the answer nothing? :) | 20:36 |
| cwebber | but i need to at least screw them in | 20:37 |
| gordon2 | cwebber: please keep the yellow tape, it is there to cover bare solder joints from shorting on the case (it should not short because of case anodization, but it is just extra precaution) | 20:37 |
| cwebber | ok! | 20:37 |
| cwebber | should I just cut some space around the holes then? | 20:38 |
| cwebber | I assume that must be safe to do | 20:38 |
| vkoskiv | Yeah, that should work. | 20:38 |
| cwebber | ok great | 20:38 |
| cwebber | ty <3 | 20:38 |
| gordon2 | yeah that should be fine | 20:38 |
| cwebber | oh one other question | 20:38 |
| vkoskiv | It's kapton tape, by the way. Space tape! :D | 20:38 |
| cwebber | I did install the wifi chip and the acrylic thing it sits on top of | 20:38 |
| cwebber | it's jiggly tho | 20:38 |
| cwebber | is there a way to keep it from being so jiggly? | 20:38 |
| cwebber | or is it safe to jiggle | 20:39 |
| vkoskiv | Mine is too! I didn't bother affixing it, but I'd put some tape there if I were you | 20:39 |
| cwebber | ok, what kind of tape! | 20:39 |
| gordon2 | cwebber: so it has two parts - hex post and the screw that goes in the hex post, can you tell what part is jiggling? | 20:39 |
| gordon2 | which part is loose | 20:39 |
| cwebber | hm I might be missing that | 20:39 |
| cwebber | I'll take a picture | 20:39 |
| gordon2 | cwebber: so previously there was a screw and a hole you put this screw in to fix up the motherboard, with this acrylic thing upgrade you have to put brass hex post instead of a screw, then mount the acrylic plate on top of that | 20:41 |
| gordon2 | with a screw that goes inside this brass hex post | 20:41 |
| gordon2 | wait, let me look at it again | 20:42 |
| gordon2 | oh it might not be brass, but nonetheless | 20:42 |
| gordon2 | (tbh this captan tape thing makes me a bit nervous, i think i would conformal coat the solder joints in some thick CC just in case) | 20:44 |
| minute | gordon2: cwebber probably has a different version of that acrylic | 20:56 |
| minute | for banana pi | 20:56 |
| gordon2 | ah, i see | 20:56 |
| minute | it just jiggles around, that is ok cwebber | 20:56 |
| minute | ok, installed linux from https://source.mnt.re/reform/reform-debian-packages/-/pipelines/5078 on my test reform here (don't try this at home) | 20:57 |
| + mjw (~mjw@2001:1c06:2486:a800:158:adb1:a1e7:3c3f) | 20:57 | |
| minute | lol i apparently have dualscreen rk3588 dsi/edp + amdgpu hdmi | 20:57 |
| minute | gdm just took a while longer to start | 20:58 |
| minute | lol wat... gnome can handle multi gpus? | 20:58 |
| minute | i have the gnome desktop, and i can move the cursor between my internal display and the hdmi from the amd rx580 | 20:59 |
| f_ | eGPU? | 20:59 |
| minute | yeah | 21:00 |
| f_ | :D | 21:00 |
| minute | some pcie things patched in the kernel for it | 21:00 |
| f_ | I have an rx550 that's basically unused at the moment, but always wanted to use as an eGPU | 21:00 |
| - erle (QUIT: Quit: K-lined) (~erle@user/erle) | 21:00 | |
| f_ | How is it all hooked up with the reform? | 21:00 |
| minute | i can drag windows between those gpus... that's really cool | 21:00 |
| minute | f_: via oculink adapter | 21:01 |
| minute | f_: pcie3 x4 m2 port to oculink dock | 21:01 |
| minute | (and the dock/gpu has its own sfx power supply) | 21:01 |
| f_ | and the oculink dock is plugged in to the reform? | 21:02 |
| + wielaard (~mjw@gnu.wildebeest.org) | 21:02 | |
| minute | f_: yeah into m2 slot | 21:03 |
| minute | f_: m2 slot oculink adapter | 21:03 |
| minute | ok it looks like it's actually the mali gpu rendering to the amdgpu's framebuffer?! | 21:03 |
| f_ | neat! | 21:03 |
| minute | but maybe it's just glxinfo being weird... thinking about how to test this | 21:03 |
| minute | i wonder if amdgpu could render to the internal framebuffer as well? :0 | 21:04 |
| f_ | I wanted one of these before https://pcngon.vn/wp-content/uploads/2023/11/Dock-eGPU-Mini-PCIe-to-PCIe-EXP-GDC-3.jpg | 21:05 |
| - mjw (QUIT: Ping timeout: 246 seconds) (~mjw@2001:1c06:2486:a800:158:adb1:a1e7:3c3f) | 21:05 | |
| f_ | Plugs in to the mini-PCIe port that, on my laptop, is currently occupied by the iwlwifi | 21:05 |
| f_ | there's also an expresscard version that in theory could work too | 21:05 |
| minute | f_: ah i've scrolled past this recently, is it originally for gpd or so? | 21:06 |
| - paperManu (QUIT: Ping timeout: 250 seconds) (~paperManu@modemcable141.205-200-24.mc.videotron.ca) | 21:07 | |
| f_ | Didn't really dig much into it TBH | 21:07 |
| f_ | I never bought it in the end | 21:07 |
| minute | ok running vkcube on the external screen. it says selected GPU 0: AMD Radeon RX 580 Series | 21:08 |
| minute | it also says that on the internal screen wat :0 | 21:08 |
| + paperManu (~paperManu@modemcable141.205-200-24.mc.videotron.ca) | 21:09 | |
| minute | can gpu rendering just be mixed and matched like that? :0 | 21:10 |
| minute | ah, next stumbling block might be debian blender not shipping vulkan support | 21:12 |
| minute | yeah. big meeeh | 21:13 |
| minute | aha! zink works! | 21:13 |
| + istar (~istar@mail.cbreid.com) | 21:19 | |
| minute | so blender runs on the internal reform display as well (aaaaaaa!) | 21:20 |
| vkoskiv | minute: I've been meaning to look into how rendering/piping display stuff works on linux. On desktop PCs it's possible to plug into the motherboard HDMI even though a discrete GPU works. | 21:22 |
| minute | vkoskiv: and the discrete gpu can render onto that HDMI? | 21:23 |
| vkoskiv | Not sure if that is what's happening in your case, but maybe the AMD gpu renders into a buffer that then gets shuffled to the internal GPU for display? | 21:23 |
| minute | vkoskiv: yeah i think the buffer is even composited through the mali gpu | 21:23 |
| vkoskiv | My understanding is that yes, it can, but there may be an associated performance/latency penalty | 21:23 |
| minute | vkoskiv: because gnome says it's using mali | 21:23 |
| vkoskiv | I studied a very small portion of the radeon driver for my latest post. Not enough to really get a good grasp of how things connect together, but I did find like 30-40 interesting looking rabbit holes to consider :D | 21:25 |
| minute | i wish there was like a "little" demo version of cp2077 to try :D | 21:37 |
| josch | minute: re "splash with rk3588" -- i would really like to have that! It would look so cool and shiny and professional to have the reform next boot with a polished splash screen. You may have seen that we build a patched version of plymouth in reform-debian-packages and it has now been used in the MNT repos for a while: https://mntre.com/reform-debian-repo/pool/main/p/plymouth/ | 21:39 |
| josch | i'm trying to get these things into Debian but the plymouth maintainer is ignoring my bug: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1112280 | 21:39 |
| josch | Apart from those patches which seem to do their job, there is the problem that rk3588 can do splash but the other platforms cannot. | 21:40 |
| josch | And I wonder whether the splash screen is enough of a reason to give up the nice property that we have the bit-by-bit identical rootfs across all platforms. | 21:40 |
| minute | josch: ah hmmm @ symlinks bug | 21:45 |
| cwebber | I upgraded to the safer battery socket things! | 21:47 |
| cwebber | it works! | 21:47 |
| cwebber | AND! | 21:47 |
| cwebber | it looks like it "fixed" the some of my batteries being low voltage or not showing up | 21:48 |
| cwebber | and the blinky battery is gone | 21:48 |
| cwebber | next up! reflash firmware! and then! | 21:48 |
| cwebber | upgrade to the banana pi! | 21:48 |
| minute | cwebber: noice noice | 21:50 |
| cwebber | it's pretty cool having this thing I felt bad about not using on the shelf for quite a while | 21:52 |
| cwebber | and then pulling it down and booting it up and upgrading it piece by piece | 21:52 |
| cwebber | and starting to use it in earnest | 21:52 |
| minute | MNT Theseus™ | 21:53 |
| vagrantc | hah | 21:53 |
| cwebber | indeed | 21:53 |
| cwebber | I have to say opening the Reform and doing 'lil upgrades to the Laptop of Theseus does make me feel like a badass | 21:53 |
| cwebber | I mean, I didn't do the work to make doing all this possible! | 21:53 |
| minute | cwebber: very happy to hear | 21:53 |
| cwebber | but it feels really cool to do | 21:53 |
| vagrantc | seconded | 21:54 |
| cwebber | I would also like to install Guix proper on this thing | 21:54 |
| cwebber | but I'll focus on getting a working setup with the Debian land stuff first | 21:54 |
| vagrantc | which variant are you working with? | 21:54 |
| cwebber | of the laptop? | 21:54 |
| cwebber | original crowdfunding campaign model, haha | 21:55 |
| vagrantc | well, i've been updating the kernel for guix for the mnt/reform (classic) with rk3588 module | 21:55 |
| cwebber | oooh | 21:55 |
| cwebber | oh yeah | 21:55 |
| vagrantc | in theory also works with other variants | 21:55 |
| cwebber | well I have the banana pi right now | 21:55 |
| cwebber | well, I'm about to | 21:55 |
| cwebber | gonna install that | 21:55 |
| cwebber | it's running the Original Module right now | 21:55 |
| cwebber | but this evening, gonna upgrade it | 21:55 |
| cwebber | and then I have a trip this weekend, and I'm not gonna have anything else with me so uh, hopefully it'll all work out! | 21:56 |
| vagrantc | i don't think i ever tested it with the original imx8 variant and guix system | 21:56 |
| minute | good luck! the banana pi / a311d is supposed to be quite stable | 21:56 |
| josch | cwebber: the only issues i had with a311d so far arose because i tried using up more ram than i had, usually when trying to compile a C++ project with too much parallelism (some build systems default to -j$(nproc) ) -- apart from that, i much prefer a311d over imx8mq which was already extremely reliable | 22:01 |
| josch | minute: about the splash: the "problem" is that it would be nice to have a way to *only* have the splash on rk3588 without having to special-case it in the system image because building one image which works everywhere is a really nice property. A few weeks ago I think I found out how we can do this. This is what runs very early in the initramfs to enable plymouth: | 22:04 |
| josch | https://sources.debian.org/src/plymouth/24.004.60-5/debian/local/plymouth.init-premount#L32 | 22:04 |
| josch | as you can see, once you enable plymouth it's enabled by default and (in contrast to what I had initially assumed) it does *not* need the "splash" boot parameter to be present | 22:04 |
| - istar (QUIT: Ping timeout: 240 seconds) (~istar@mail.cbreid.com) | 22:04 | |
| + digitalrane (~digitalra@user/digitalrane) | 22:04 | |
| josch | so how about doing the opposite: lets patch the *other* u-boot builds to add plymouth.enable=0 to their ${bootargs} | 22:04 |
| josch | minute: thoughts? | 22:05 |
| gordon2 | random observation: pressing circle+space when the rk3588 is on causes reform2_lpc module to return bunch of garbage in its battery reporting | 22:11 |
| gordon2 | another random achievement today: ffmpeg+mpv with v4l2request patches from sigrid works like a charm with 6.17 | 22:14 |
| - gustav25 (QUIT: Quit: Quit) (~gustav@c-78-82-52-240.bbcust.telenor.se) | 22:15 | |
| gordon2 | had to use hwdec=v4l2request-copy tho, too tired to debug why it does not work with hwdec=v4l2request | 22:17 |
| minute | josch: hmm but plymouth is already enabled, it's just not graphical | 22:19 |
| minute | josch: it already took over the console and does that cyan text etc | 22:19 |
| minute | josch: so i thought on rk3588 we could throw "splash" in the uboot env file | 22:20 |
| minute | gordon2: interesting @ garbage, could be leftover debug output in the lpc code | 22:21 |
| sigrid | gordon2: | 22:21 |
| sigrid | hwdec=v4l2request | 22:21 |
| sigrid | ytdl-format="mp4[height<=1080]" | 22:21 |
| sigrid | oops. sorry. what I was trying to say is that somehow it works for me just fine | 22:21 |
| sigrid | maybe it's a specific video that doesn't? | 22:22 |
| gordon2 | sigrid: http://0x0.st/K9-9.txt hwdec=v4l2request produces this | 22:22 |
| gordon2 | when i do the same video with hwdec=v4l2request-copy everything works fine | 22:23 |
| gordon2 | maybe that's because of gpu-next | 22:24 |
| gordon2 | but probably some permission problem | 22:27 |
| sigrid | I have gpu instead of gpu-next and it fails the same way, but it doesn't switch to software decoding | 22:27 |
| sigrid | [hwupload] hwdec 'drmprime' does not report hwframe constraints. Using static metadata. | 22:28 |
| sigrid | [vf] [out] 1280x720 drm_prime[nv12] bt.709/bt.709/bt.1886/limited/display CL=mpeg2/4/h264 crop=1280x720+0+0 | 22:28 |
| gordon2 | well, anyways, i'm happy with v4l2request-copy for now | 22:30 |
| sigrid | ah nvm: | 22:30 |
| sigrid | [vo/gpu] Loading hwdec driver 'v4l2request' | 22:30 |
| sigrid | [vo/gpu/v4l2request] using libplacebo dmabuf interop | 22:30 |
| sigrid | perhaps that's why it works? | 22:30 |
| gordon2 | libplacebo is gpu-next, isn't it? | 22:30 |
| sigrid | doesn't seem to be the case for me | 22:30 |
| gordon2 | that is a very weird line | 22:30 |
| * wielaard -> mjw | 22:33 | |
| gordon2 | sigrid: if you're interested, https://github.com/mpv-player/mpv/discussions/14748 solves the problem | 22:35 |
| gordon2 | i.e. --gpu-context=wayland | 22:35 |
| sigrid | oh ok, I have a whole bunch of environmental variables set. guess one of them helps | 22:35 |
| gordon2 | okay, waiting for someone to implement VP8/VP9 now, i think v4l2-request patch for ffmpeg is abandoned tho | 22:37 |
| josch | minute: didn't we decide to explicitly revert the commit which added plymouth to reform-desktop-full because it messed up the early (text only) boot on platforms which are not rk3588 | 22:43 |
| josch | ? | 22:43 |
| josch | gordon2: do you know more about the v4l2-request story? There have been patches for years and I have no idea why none of them ends up in ffmpeg proper? | 22:45 |
| josch | or maybe sigrid knows | 22:45 |
| gordon2 | josch: the only thing i know is 1) they were (almost) completely ignored by ffmpeg mailing list 2) according to random reddit testimonies, apparently this is a common behaviour by ffmpeg maintainers | 22:47 |
| gordon2 | also nowdays there they migrated to trac so maybe someone can create a ticket for them about the state of v4l2-request patch | 22:48 |
| gordon2 | or was it gitlab? | 22:48 |
| gordon2 | i don't quite remember, remember just reading a news article that they upgraded from mailing list | 22:48 |
| gordon2 | ah, here we go https://code.ffmpeg.org/FFmpeg | 22:49 |
| sigrid | https://code.ffmpeg.org/FFmpeg/FFmpeg/issues/20484 | 22:51 |
| gordon2 | nice, somebody did it | 22:52 |
| gordon2 | awesome | 22:52 |
| josch | thank you! | 22:52 |
| sigrid | also consider that perhaps nobody wants to bother with v4l2 and instead people are waiting for full-on vulkan video decoding/encoding | 22:52 |
| gordon2 | gstreamer did bother, but i see your point | 22:53 |
| sigrid | I would bother too. vulkan involves a lot more than just an interface to the hardware, to my near-zero knowledge | 22:54 |
| gordon2 | tbh just looking how productive panvk development is i would bet on it being finished earlier than v4l2 | 22:54 |
| + istar (~istar@mail.cbreid.com) | 22:55 | |
| josch | speaking about gstreamer -- that did work out-of-the-box on imx8mq with hantro | 22:55 |
| josch | what's its status on rk3588? | 22:55 |
| sigrid | as long as it doesn't require running full blown wayland compositor to decode a video :) | 22:55 |
| gordon2 | josch: i tested it on my rock5b+, it worked perfectly fine so i see no reason for it not to work on reform | 22:56 |
| josch | gordon2: it worked with hardware decoding? | 22:56 |
| gordon2 | yep | 22:56 |
| josch | nice | 22:56 |
| gordon2 | i used gst-play if i remember correctly | 22:57 |
| josch | i packaged clapper for Debian because that was a nice video player using gstreamer where hardware decoding worked | 22:57 |
| gordon2 | gst-play-1.0 | 22:57 |
| josch | try clapper -- in contrast to gst-play it's a real video player :) | 22:57 |
| gordon2 | oh no thanks, i'll stick to mpv :) | 22:57 |
| josch | the other gstreamer players have the problem that they do not use the new video element | 22:58 |
| josch | haha i understand, i'd also like to stick with mpv :) | 22:58 |
| gordon2 | it would be so much easier if mpv just could work with gstreamer as well, but unfortunately https://github.com/mpv-player/mpv/issues/2512#issuecomment-158663652 | 22:59 |
| gordon2 | also i bet abstracting it is an immense PITA | 23:01 |
| sigrid | yo dawg I heard you like video decoding, so I put ffmpeg with v4l2request as a vaapi library, so that firefox' ffmpeg can use vaapi while it uses ffmpeg while it uses v4l2request | 23:02 |
| sigrid | I kinda agree this all is getting quite ridiculous. no proper way to play a video these days. all my hope is on the mesa folks | 23:02 |
| gordon2 | is there a v4l2-request vaapi lib? | 23:05 |
| sigrid | there are a few and they are all broken | 23:05 |
| gordon2 | oh i vaguely remember one, yes | 23:05 |
| gordon2 | i think i abandoned my tries to make it work | 23:05 |
| sigrid | I tried to fix a couple and it did not work out well - just wasted large number of hours | 23:06 |
| gordon2 | i remember times when ffmpeg was a solution (to playing _everything_), not a problem... | 23:09 |
| gordon2 | or was it libav back then? | 23:09 |
| gordon2 | well, hopefully vulkan will fix everything | 23:10 |
| - paperManu (QUIT: Ping timeout: 255 seconds) (~paperManu@modemcable141.205-200-24.mc.videotron.ca) | 23:11 | |
| gordon2 | oh, yey https://code.ffmpeg.org/Kwiboo/FFmpeg | 23:21 |
| gordon2 | work is happening literally right now | 23:21 |
| gordon2 | and vp8! | 23:22 |
| gordon2 | gonna try it right now | 23:22 |
| sigrid | to me that looks like a rebase on master | 23:33 |
| sigrid | but maybe it has something fixed that i missed when making those patches for alpine | 23:33 |
| sigrid | (patches have vp8 too btw) | 23:34 |
| + paperManu (~paperManu@107.159.15.124) | 23:35 | |
| gordon2 | yeah, nevermind, i have problems playing AV1 not VP8/VP9 | 23:36 |
| + paperManu_ (~paperManu@107.159.15.124) | 23:38 | |
| gordon2 | VP8 works fine, VP9 sometimes complains about absence of VP9F device (whatever it is) | 23:40 |
| gordon2 | damn why did i download half of youtube in AV1 >.< | 23:42 |
| sigrid | heh. i had my analog photo scans in jxl (compresses well) and now i can barely view them with anything | 23:43 |
| gordon2 | sxiv shows them well! | 23:47 |
| gordon2 | (or swiv) | 23:47 |
| - plomlompom (QUIT: Remote host closed the connection) (~plom@139.59.138.73) | 23:48 | |
| + plomlompom (~plom@139.59.138.73) | 23:49 | |
| sigrid | I mean it kinda works with feh but it takes ~10s to load a picture | 23:53 |
| gordon2 | ah, i see, VP9 is still TODO, AV1 tho already there, so just need another patch for ffmpeg for it to work https://gitlab.collabora.com/hardware-enablement/rockchip-3588/notes-for-rockchip-3588/-/blob/main/mainline-status.md | 23:53 |
| sigrid | sounds about right | 23:55 |
| gordon2 | i wonder how hard it is going to be to repeat https://code.ffmpeg.org/Kwiboo/FFmpeg/commit/1f0b9afb74ab2841424c9273a2309bd3c20648d7 | 23:55 |
| gordon2 | c-file full of magic constants | 23:55 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
