| chartreuse | Might give bookworm a try then upgrade to testing to see how that behaves. Though the only breakage I'm seeing right now is FreeCAD on unstable | 00:00 |
|---|---|---|
| josch | the reform repo only has unstable, but you can mostly mix that with testing | 00:00 |
| - xktr (QUIT: Ping timeout: 264 seconds) (~xktr@user/xktr) | 00:02 | |
| + xktr (~xktr@user/xktr) | 00:02 | |
| josch | chartreuse: one problem with mixing the reform repo with testing is every time that gcc didn't transition to testing yet | 00:03 |
| chartreuse | Ah okay, so some stuff in that repo is depending on the latest gcc | 00:04 |
| chartreuse | Does it need the latest gcc though, or would setting it to something like gcc > 11 be enough | 00:04 |
| josch | chartreuse: the dependency relationship is very strict | 00:07 |
| josch | chartreuse: the problem is linux-libc-dex | 00:07 |
| chartreuse | Ah that makes sense | 00:08 |
| josch | and you need that because you want to build the reform-lpc dkms module | 00:08 |
| josch | *linux-libc-dev | 00:08 |
| chartreuse | I didn't think you needed the latest libc to build a dkms module, just the same one that the kernel is build on. But I guess the kernel is in unstable so it's being built on the latest | 00:09 |
| chartreuse | I might give the bookworm image a try later. I imagine it (besides those issues) should work fine on an 8mq since I have a ath9k wifi module so won't lose that | 00:10 |
| josch | chartreuse: why do you want a kernel later than bookworm if you have the imx8mq? | 00:12 |
| - Gooberpatrol66 (QUIT: Ping timeout: 264 seconds) (~Gooberpat@user/gooberpatrol66) | 00:14 | |
| - qbit (QUIT: Remote host closed the connection) (~qbit@mail.suah.dev) | 00:26 | |
| + qbit (~qbit@mail.suah.dev) | 00:28 | |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 00:36 | |
| - qbit (QUIT: Remote host closed the connection) (~qbit@mail.suah.dev) | 00:41 | |
| + qbit (~qbit@mail.suah.dev) | 00:43 | |
| + chomwitt (~chomwitt@2a02:587:7a1b:1e00:1ac0:4dff:fedb:a3f1) | 00:44 | |
| - qbit (QUIT: Remote host closed the connection) (~qbit@mail.suah.dev) | 00:48 | |
| + qbit (~qbit@mail.suah.dev) | 00:50 | |
| chartreuse | I don't really, just was thinking of putting something more "stable" on here. But might just stick with sid | 00:57 |
| - xktr (QUIT: Quit: leaving) (~xktr@user/xktr) | 01:23 | |
| - mjw (QUIT: Ping timeout: 264 seconds) (~mjw@gnu.wildebeest.org) | 01:39 | |
| + xktr (~xktr@user/xktr) | 01:57 | |
| - q66 (QUIT: Quit: WeeChat 4.0.2) (~q66@q66.moe) | 02:39 | |
| + q66 (~q66@q66.moe) | 02:46 | |
| - q66 (QUIT: Client Quit) (~q66@q66.moe) | 02:46 | |
| + q66 (~q66@q66.moe) | 02:47 | |
| - nsc (QUIT: Ping timeout: 240 seconds) (~nicolas@i5C74DD2D.versanet.de) | 03:11 | |
| + nsc (~nicolas@29-98-142-46.pool.kielnet.net) | 03:13 | |
| violet | the pocket reform is here :) | 03:28 |
| - Manis (QUIT: Quit: Gateway shutdown) (01a66df340@77-56-188-94.dclient.hispeed.ch) | 04:03 | |
| - chomwitt (QUIT: Ping timeout: 246 seconds) (~chomwitt@2a02:587:7a1b:1e00:1ac0:4dff:fedb:a3f1) | 04:18 | |
| - bkeys (QUIT: Remote host closed the connection) (~Thunderbi@45.134.140.153) | 04:23 | |
| - akira (QUIT: Ping timeout: 268 seconds) (~akira@37.4.230.225) | 05:57 | |
| + akira (~akira@37.4.230.225) | 06:00 | |
| - akira (QUIT: Read error: Connection reset by peer) (~akira@37.4.230.225) | 06:12 | |
| + akira (~akira@37.4.230.225) | 06:14 | |
| reform12872 | @josch: it seems i can't manage to build the image | 07:19 |
| reform12872 | it sill dies with -j2 | 07:20 |
| reform12872 | https://mister-muffin.de/p/MfSh.txt | 07:20 |
| reform12872 | https://mister-muffin.de/p/MfSh.txt | 07:20 |
| reform12872 | any other ideas i could try?; | 07:20 |
| josch | fatal error: Killed signal terminated program cc1 | 07:41 |
| josch | interesting! | 07:41 |
| josch | building the kernel image works for me even with six parallel processes | 07:41 |
| josch | maybe that is because i am using swap? | 07:42 |
| josch | reform12872: do you have any swap configured? | 07:42 |
| - akira (QUIT: Ping timeout: 268 seconds) (~akira@37.4.230.225) | 07:50 | |
| + akira (~akira@2a01:599:a25:97c3:38b2:1449:228:176) | 07:51 | |
| + Manis (01a66df340@77-56-188-94.dclient.hispeed.ch) | 08:01 | |
| chartreuse | If you look in dmesg you can probably check if the OOM killer is what killed it | 08:24 |
| chartreuse | It'll have a line like "Out of memory: Kill process ..." | 08:24 |
| chartreuse | Having some swap would be a good idea, though make sure you're not running something memory hungry like a browser while doing the compiling. I believe you should just be able to also restart the compilation with the same command and it should pick up assuming no make clean has been done | 08:25 |
| josch | on my a311d reform i have 8 GB of swap configured and am running a browser and all my other normal stuff while compiling src:linux | 08:28 |
| chartreuse | I don't remember having any swap setup when I was originally messing with compiling some earlier 6.x kernels a year ago or such. Though maybe I did. I know I was trying to setup zram or zswap back then doing that | 08:33 |
| josch | chartreuse: but was that on the imx8mq? | 08:34 |
| chartreuse | Yes | 08:34 |
| josch | then i had the same experience | 08:35 |
| chartreuse | I didn't think that'd make much difference with compiling, but dunno | 08:35 |
| josch | i think the problem is that the a311d has less ram but more cores | 08:35 |
| josch | so by default, sbuild will use all available cores which is six on the a311d | 08:35 |
| josch | while having less memory than imx8mq with four cores | 08:35 |
| josch | so sbuild will have spawned less processes in parallel on a machine with more memory | 08:35 |
| chartreuse | Both have 4GB though? I would have though then doing a -j4 or -j2 would have solved that then | 08:36 |
| josch | since a311d has 50% more cores, i learned to be much more careful with my ram than i had been with imx8mq | 08:36 |
| josch | chartreuse: no, a311d has 3.6 GB | 08:36 |
| chartreuse | Ah the video buffer takes from ram on there then | 08:36 |
| josch | yes, i have zero use for 400 MB video memory but what can you do :) | 08:37 |
| chartreuse | Another option to swap space on a nvme or such is setting up zswap to use some of your ram as a compressed swap space. That's my normal setup on here with 2GB given to that | 08:37 |
| chartreuse | er zram not zswap (similar but different) | 08:37 |
| josch | with my 8 GB of swap i usually do not run into memory problems -- no idea if it's faster to use zram (and pay the cpu cost) or use more swap (and pay the ssd slowness cost) | 08:38 |
| chartreuse | With the higher core count on the a311d I'd assume it'd be more worth while on there, I have mine set to zstd compression and I've not really noticed it too bad, can even set lz4 for even faster speed at cost of compression ratio | 08:39 |
| chartreuse | Can actually use both too with zram having a higher swap priority. zram with CONFIG_ZRAM_WRITEBACK can also then fall back nicely to a swap file behind that. | 08:41 |
| chartreuse | Though depending on the workload maybe having the full 3.5G of main memory makes more sense, rather than giving up half of that to compressed ram | 08:41 |
| chartreuse | I don't think anyone really knows what's the best or benchmarked all the swap approaches | 08:42 |
| chartreuse | Main reason I was going for it was just general aversion to swap hammering ssds | 08:43 |
| josch | makes sense | 08:47 |
| josch | and clearly benchmarks would be nice | 08:47 |
| josch | but those are obviously highly system dependent :) | 08:47 |
| chartreuse | Was just loading a bunch of applications to see how it was going to respond to memory pressure, and noticed that Prusaslicer no longer is working right just showing a while area for the build volume and a warning about an unsupport OpenGL version. Guess they increased their minimum since last year | 09:01 |
| chartreuse | It wants 3.2 now, while the etnaviv driver on here is 2.1 | 09:01 |
| - jjbliss (QUIT: Ping timeout: 252 seconds) (~jjbliss@1464766-static.elnsmiaa.metronetinc.net) | 09:07 | |
| minute | chartreuse: tried overriding to 3.2? | 09:10 |
| chartreuse | I don't remember the commands for it, though I figure it was just trying anyway since it launched fine and didn't error out | 09:12 |
| chartreuse | MESA_GL_VERSION_OVERRIDE=3.2 then gives a list of missing shaders and then basically the same inapp behaviour | 09:13 |
| chartreuse | With the override set the errors are basically being unable to compile the shaders since it's requiring 1.4 and overriding that won't change it | 09:17 |
| chartreuse | https://github.com/prusa3d/PrusaSlicer/issues/12334 | 09:20 |
| + jjbliss (~jjbliss@1464766-static.elnsmiaa.metronetinc.net) | 09:20 | |
| chartreuse | Sounds like it was changed early this year, and quite a few people are complaining and finding workarounds for it, since lots of people used it on the RPi and even the Pi 5 only has 3.1 support | 09:21 |
| - cobra (QUIT: Ping timeout: 268 seconds) (~cobra@user/Cobra) | 09:24 | |
| pandora | I will try to run the script with like 8 GB of swap | 09:39 |
| - reform12872 (QUIT: Remote host closed the connection) (~user@business-90-187-186-49.pool2.vodafone-ip.de) | 09:46 | |
| minute | chartreuse: you also need to set MESA_GLSL_VERSION_OVERRIDE=320 but it might still not work | 09:54 |
| josch | pandora: you can just use a swapfile for testing purposes and create a real partition for it later | 10:00 |
| josch | minute: do we have these kind of workarounds documented somewhere? | 10:00 |
| josch | they pop up every once in a while and I find myself grep-ing the irc history too often :D | 10:01 |
| minute | josch: yes, in the operator handbook | 10:01 |
| josch | aha! | 10:01 |
| minute | josch: https://mntre.com/reform2/handbook/software.html#gpu-hacks | 10:02 |
| minute | (it's new since rev 2 of the handbook) | 10:02 |
| josch | perfect, thank you! | 10:02 |
| josch | shame on me -- i had reviewed that and should've remembered :D | 10:02 |
| josch | is anybody using emulate.sh in the reform-system-image repo to test the system images? | 10:03 |
| josch | i have a bunch of changes/improvements and maybe somebody wants to test them | 10:03 |
| minute | josch: haha no shaming necessary | 10:03 |
| minute | josch: well, i haven't used it in a while (i think last time i tried it was broken). it's useful for sure | 10:03 |
| vkoskiv | I feel quite tempted to grab the v2 book on my next order. I don't think my name has ever appeared in a book before :D | 10:04 |
| chartreuse | minute: Same issue with that override, it's trying to use GLSL 1.4 | 10:04 |
| vkoskiv | (Thank you for including it, made me smile!) | 10:04 |
| minute | chartreuse: weird but ok! | 10:04 |
| minute | vkoskiv: nice :3 | 10:04 |
| chartreuse | And some 1.5 ones too: Unable to compile vertex shader of shader program 'dashed_thick_lines': | 10:04 |
| chartreuse | 0:1(10): error: GLSL 1.50 is not supported. Supported versions are: 1.00 ES | 10:04 |
| josch | vkoskiv: my name never appeared in a book together with Lukas' so I also totally will get myself a physical copy for the bookshelf the next time i order something from MNT :) | 10:05 |
| chartreuse | Oh hey I'm in there too, so maybe I should finally get a physical copy. And also get some upgrades like the protected battery boards XD | 10:06 |
| chartreuse | One of the few people in there under a pseudonym XD | 10:06 |
| josch | chartreuse: you really want the protected boards -- they will let you sleep so much better at night :D | 10:07 |
| josch | chartreuse: would you have preferred the name your government knows you by instead? | 10:07 |
| chartreuse | Heh yeah I know, I already had managed to drain the cells once and had to replace some of them | 10:08 |
| chartreuse | I was one of the ones concerned about the lack of undercharge protection in the first place too XD | 10:08 |
| chartreuse | Looking at the store keeps making me eye the RK3588 boards as expensive as they are, would be a fun upgrade | 10:10 |
| josch | the trend goes to a second reform! ;) | 10:10 |
| pandora | @josch: hehe yeah just use a 8gb swap file. It’s compiling rn. Gonna see the results in a few hours I guess :D | 10:10 |
| josch | https://tooting.ch/@vimja/112668406067498207 | 10:11 |
| chartreuse | Going to kill me through my wallet with upgrades, like a Pocket Reform, or a Reform Next XD | 10:11 |
| josch | i'm currently half jobless but got a new half-job starting in october, so that means more money for MNT then and since the rk3588 is still a few months away, this might align nicely :) | 10:14 |
| minute | :D | 10:15 |
| chartreuse | Yeah I'm applying for jobs right now, so hoping to be able to afford it when it comes out, kinda like when I got the reform in the first place | 10:15 |
| chartreuse | Need someone to come up with some 18650 lifepo4 cells at >2000mAh capacity to make up for the higher power draw | 10:16 |
| chartreuse | Though honestly the life span of the 8mq on these 1800's is plenty fine for me, even with basically no automatic power savings | 10:17 |
| minute | power draw on rk3588 is not that bad, except if you heavily exercise it of course :D | 10:17 |
| - lexik_ (QUIT: Read error: Connection reset by peer) (~lexik@93.185.97.218) | 10:17 | |
| josch | chartreuse: you might be interested in the reform-next then, as that may support multi-chemistry charging, so li-ion cells with higher power density would be a possibility | 10:17 |
| josch | minute: my cpu is idle 99% of the time, i think personally i'm mostly interested in the power draw of "sway is running with 10 terminals open" :) | 10:18 |
| minute | josch: i think the rk itself draws around 4W then | 10:18 |
| josch | that's real nice! | 10:18 |
| minute | huh, this might be a "light" version of rk3588 combined with rcm4 https://www.cnx-software.com/2024/07/03/orange-pi-cm5-a-raspberry-pi-cm4-alternative-with-up-to-16gb-ram-256gb-emmc-flash/ | 10:19 |
| josch | reform.d.n just had the first successful run of its scripts via cron since may 15 :) | 10:20 |
| minute | has a third connector though so might not be immediately 100% useful | 10:20 |
| josch | pandora and ex-parrot gave me some good motivation -- so thank you two! :) | 10:20 |
| chartreuse | Oh so it's actually more variable than the 8mq is, seems to draw 240-280mAh depending on load | 10:20 |
| josch | wow, three clicks to get the cnx-software cookie banner to go away... | 10:21 |
| chartreuse | I do like the use of lifepo4 though, even though it's not as energy dense as li-ion (1800vs3000) for 18650's | 10:21 |
| + lexik (~lexik@93.185.97.218) | 10:22 | |
| josch | chartreuse: for me, i really like the argument regarding conflict material in my cells | 10:22 |
| josch | i'm surprised companies are announcing cm4 modules still. I weren't aware they are still that popular. I also wonder if/when there might be a cm5 or a raspi5 in cm4 form factor (hopefully the latter i guess) | 10:23 |
| vkoskiv | Salt batteries aren't there yet wrt. energy density, but I did see one for the first time on a store shelf a few weeks ago! | 10:24 |
| pandora | @josch: I will continue to generate more motivation since I got my own reform now and want to push it a bit :P | 10:24 |
| chartreuse | The salt ones do seem to have a much more sloped discharge curve and still lower that lifepo4 but always neat to see newer techs | 10:24 |
| josch | pandora: yeeeeees! :D | 10:25 |
| chartreuse | I did miss out on a year of my Reform so getting back into the full swing of it again | 10:25 |
| vkoskiv | Salt is neat because it's salt, kind of hard to make a conflict mineral out of that :D | 10:25 |
| vkoskiv | Though I'm sure some people will try anyway | 10:25 |
| josch | vkoskiv: re salt batteries, you saw this, right? https://tiny.tilde.website/@lykso/112221544025551357 | 10:26 |
| chartreuse | Clearly the solution is to use a material found everywhere on earth, lead. And mix it with some kind of acid as an electrolyte :P | 10:26 |
| josch | i really like lead batteries -- i'd not give our students any other chemistry ;) | 10:27 |
| chartreuse | Though laptops with them went away fairly quick in the early 90s XD | 10:28 |
| chartreuse | For their experiments they really would have benefited from an original plain battery board it seems, since the protection is setup for lifepo4. Plus needing to change the resistor divider on the charger circuit | 10:30 |
| chartreuse | I have no desire to use them, but presumably that voltage divider could be changed too use li-ion cells in the reform as well, by setting the charge current to 4.2 instead of 3.65, then tweaking the LPC firmware to suitable values | 10:32 |
| chartreuse | s/current/voltage/ | 10:33 |
| josch | chartreuse: maybe but you then also need some temperature monitoring for safety, no? | 10:33 |
| chartreuse | Could use protected cells, and set the charge current to be low enough to not be as much as a worry. | 10:34 |
| chartreuse | I was looking at the LPC firmware again since I replaced the hacktheplanet header I accidentally melted last time. And messing up the balacing logic is keeping me from hacking that controller all too much even though it has plenty of gpios free | 10:35 |
| chartreuse | For Li-ion could make a custom protected battery board with li-ion protection circuits as well. It's not an impossible setup | 10:36 |
| josch | impossible it certainly is not :) | 10:37 |
| chartreuse | The reform is currently charging at around 1C I imagine (-1.46A) That's not unreasonable for li-ion, though could be set lower for safety/longevity too | 10:37 |
| josch | but it's probably easier to wait for Lukas to do the work for us in the Reform Next :) | 10:37 |
| chartreuse | But then I'd have to buy another laptop instead of hacking :P | 10:37 |
| josch | hehe :) | 10:38 |
| josch | new emulate.sh for anybody who'd like to give it a try: https://source.mnt.re/reform/reform-system-image/-/merge_requests/101 | 10:39 |
| chartreuse | Just like my mind was looking at the audio circuit and noticing the IC has a 3rd mono output and thinking about adding a "subwoofer" type amp on there. I was originally thinking it wouldn't work since the pin isn't broken out to a pad, but I did just look under my reform and noticed that there's solder visible around the chip so a wire *could* be tacked on | 10:39 |
| chartreuse | I know you and minute have added the pocket reform speakers in parallel, but not sure how good that is for the quality/chip as that would lower the impedence to 4 ohms instead of 8 | 10:40 |
| chartreuse | Was also looking at digikey and seeing there was another speaer brand that has an 13x18 or such speaker the same size as the normal ones and wondering if it'd be any louder 1:1 swapped | 10:41 |
| chartreuse | I'm sure the whole lack of loudness thing on the built in is that there's no room for a large resonance box in such a compact speaker | 10:41 |
| josch | since my personal use is "make episodes of my little pony louder for the little ones", speaker quality was of little importance to me :D | 10:41 |
| chartreuse | Heh yeah. I already improved the headphone output on mine, so that takes care of most of the quality concerns. | 10:42 |
| chartreuse | With the big 330uF electrolytic dc blocking caps to improve bass response | 10:42 |
| chartreuse | Soldered to the tiny 0402 pads where the 47uF ceramics lived | 10:43 |
| josch | uff i'd like to have your fingers XD | 10:43 |
| amospalla | josch: yeah, it was not the first login, but I remember one root login had no bash. | 10:44 |
| chartreuse | https://tiny.tilde.website/@lykso/112221544025551357 | 10:44 |
| chartreuse | Oops | 10:45 |
| josch | amospalla: thank you for confirming -- i have to figure out what's going on there... | 10:45 |
| josch | funnily i cannot reproduce it in qemu | 10:45 |
| chartreuse | https://community.mnt.re/t/speakers-too-quiet-try-this/375/6 | 10:45 |
| amospalla | josch: it happens, but it has bash on /etc/passwd | 10:46 |
| josch | yeah, i was confused by that too | 10:47 |
| josch | how does it give me /bin/dash if it says otherwise in /etc/passwd | 10:47 |
| chartreuse | Is /bin/bash symlinked to /bin/dash? | 10:47 |
| amospalla | never saw that before | 10:47 |
| chartreuse | Might do that if bash isn't installed to help with scripts | 10:47 |
| chartreuse | or /usr/bin/bash alternatively | 10:47 |
| josch | chartreuse: nope, typing "bash" gets me to bash | 10:48 |
| chartreuse | If you type the path it shows in /etc/passwd does that bring you to bash though? | 10:48 |
| chartreuse | Perhaps it's /bin/bash while bash is installed at /usr/bin | 10:48 |
| josch | it is installed in /usr/bin but /bin is a symlink to /usr/bin | 10:48 |
| chartreuse | Hmm | 10:49 |
| josch | i suspect it is something to do with the custom profile.d scripts the system image ships | 10:49 |
| chartreuse | Is bash in /etc/shells? Though I figure if it wasn't a valid login shell it's fail rather than dash | 10:49 |
| chartreuse | Or something in systemd yeah | 10:49 |
| josch | yes, both /bin/bash as well as /usr/bin/bash is in /etc/shells | 10:50 |
| chartreuse | INB4 setting up the only sysvinit MNT reform :P | 10:51 |
| chartreuse | (Though I guess people are running BSD on them... so they do have a plain init type reform) | 10:52 |
| amospalla | pstree shows: "greetd --session-worker 12" -> sh -> bash | 10:52 |
| josch | aha, maybe it's greetd | 10:52 |
| amospalla | that is, root logged in, into sh, and then I typed bash | 10:52 |
| josch | command = "/usr/sbin/agreety --cmd /bin/sh" | 10:52 |
| josch | amospalla: it's greetd! | 10:52 |
| josch | thank you!! :D | 10:53 |
| chartreuse | Oddly /usr/bin/sh is a symlink to /usr/bin/dash, but running it and echo $SHELL gives /bin/bash | 10:53 |
| amospalla | glad to help | 10:53 |
| chartreuse | I guess dash just reports bash | 10:54 |
| chartreuse | Since it's ignoring my .bashrc | 10:54 |
| chartreuse | minute: Did try that LIBGL_ALWAYS_SOFTWARE=1 you mentioned for prusa-slicer and it does work in software at least, but quite slow as expected | 10:56 |
| + mjw (~mjw@gnu.wildebeest.org) | 10:59 | |
| josch | and that's also why the problem does not appear in qemu, because there i have a login via serial and greetd is configured to show up on virtual terminal 7 | 11:01 |
| vkoskiv | What I want is a flag like LIBGL_ALWAYS_SOFTWARE, but where it only emulates a feature in software if it isn't supported in hardware | 11:03 |
| vkoskiv | I think 99% of the things Blender does are supported in hardware, but I have to set everything to software, or nothing at all | 11:04 |
| + cobra (~cobra@user/Cobra) | 11:04 | |
| vkoskiv | I don't mind unsupported features being slow, as long as the basic stuff is still accelerated | 11:04 |
| chartreuse | Presumably can be done at the driver level adding emulations in there to get a higher version number, though would be nice if mesa could do that like you said for just ones that no accelerated routine exists | 11:09 |
| - colinsane (QUIT: Quit: bye) (~colinunin@97-113-64-230.tukw.qwest.net) | 11:10 | |
| + colinsane (~colinunin@97-113-64-230.tukw.qwest.net) | 11:16 | |
| erle | > where it only emulates a feature in software if it isn't supported in hardware | 11:29 |
| erle | aren't the gallium drivers working like that? | 11:29 |
| amospalla | On my Pocket Reform, greetd's shutdown fails with "Command failed: no such file or directory (os error 2)", I added "--power-shutdown 'sudo poweroff'" to /etc/greetd/config.toml and a sudo entry to allow _greetd user to execute without asking for a password /usr/sbin/poweroff and it works. | 11:30 |
| amospalla | When I find time, where should I try to upload PR with these kind of changes? | 11:32 |
| josch | amospalla: i think those would go either into the Debian packaging or into greetd upstream | 11:32 |
| amospalla | I have never submitted patches though, but I would enjoy doing so. | 11:32 |
| amospalla | Yeah, I saw there is nothing related inside reform-tools. | 11:33 |
| josch | and making the shutdown button work is also not reform-specific | 11:33 |
| josch | but possibly debian-specific | 11:33 |
| josch | and if it's not even debian-specific, it should go to upstream | 11:33 |
| amospalla | Would Debian accept these kind of patches into stable? | 11:33 |
| josch | into stable? probably not, no | 11:34 |
| josch | argument would be: you are free to adapt your /etc/greetd/config.toml to your liking as a stable user | 11:34 |
| josch | this is just about sane defaults | 11:35 |
| amospalla | That is what I thought, so the correct thing should be opening a patch against debian or upstream, and maybe adding something into reform-tools that patches stable images? | 11:35 |
| josch | we can add this into the image, yes | 11:36 |
| josch | but we cannot add it to reform-tools because users might have their own custom /etc/greetd/config.toml and reform-tools must not overwrite that | 11:36 |
| amospalla | Or are these kind of changes into the image itself at image creation? | 11:36 |
| josch | it would be even better if greetd upstream would default to reading a config file in /usr/share/greetd/config.toml and if there exists /etc/greetd/config.toml, then it would prefer the latter | 11:37 |
| josch | then packages could update /usr/share/greetd/config.toml without damaging the user's config in /usr/share/greetd/config.tom | 11:37 |
| amospalla | I see. What do you expect on this case, for example, somebody to notify, open bug, submit pr? | 11:37 |
| amospalla | I've never submited participated into open source project, I'm new. | 11:38 |
| josch | amospalla: i do not know greetd upstream and thus i don't know how friendly they are to newcomers | 11:38 |
| josch | if you submit this as a bug to debian, then i can promise you to help you out | 11:38 |
| amospalla | I mean, about Reform itself, about Debian stable on it. | 11:38 |
| josch | if you want to submit a bug or mr to the reform tooling, the right place would be reform-system-image repo | 11:39 |
| amospalla | Ok, thank you. | 11:40 |
| - cobra (QUIT: Quit: ZNC 1.8.2 - https://znc.in) (~cobra@user/Cobra) | 12:04 | |
| + cobra (~cobra@user/Cobra) | 12:09 | |
| minute | chartreuse: how do you mean "in parallel" @ speakers? reform has separate l/r speaker outputs | 12:09 |
| minute | chartreuse: ah, in parallel to the old ones. i haven't done that, i have disconnected the old ones | 12:10 |
| minute | amospalla: do you have an approved account on source.mnt.re? | 12:10 |
| amospalla | minute: no, I don't. | 12:10 |
| minute | amospalla: if you want one, please tell me your username after signing up so i can approve it | 12:11 |
| amospalla | minute: thank you. | 12:11 |
| amospalla | @minute I created the username amospalla on gitlab. | 13:06 |
| - Aard (QUIT: Remote host closed the connection) (~bwachter@217.11.60.132) | 13:19 | |
| + Aard (~bwachter@edna-edison.lart.info) | 13:19 | |
| + chomwitt (~chomwitt@2a02:587:7a1b:1e00:1ac0:4dff:fedb:a3f1) | 13:24 | |
| - akira (QUIT: Read error: Connection reset by peer) (~akira@2a01:599:a25:97c3:38b2:1449:228:176) | 13:29 | |
| + akira (~akira@37.4.230.225) | 13:29 | |
| - mjw (QUIT: Killed (zinc.libera.chat (Nickname regained by services))) (~mjw@gnu.wildebeest.org) | 13:32 | |
| * Guest1668 -> mjw | 13:32 | |
| + Guest4377 (~mjw@gnu.wildebeest.org) | 13:32 | |
| gsora | oooh got a delivery, wifi card for my reform! | 16:12 |
| gsora | got myself a mt7922 | 16:13 |
| bluerise | nice | 17:09 |
| mtm | minute: ticket filed, looks like plom is out of office until next week | 17:09 |
| bluerise | ETA for the RK3588 SoM? Wondering if I should get one | 17:09 |
| bluerise | Got a NanoPi R6C incoming soon | 17:09 |
| bluerise | The Firefly ITX 3588 that I got was complete shit :/ | 17:10 |
| bluerise | I hope the SoM is better | 17:10 |
| gsora | why was it bad? | 17:13 |
| bluerise | mostly because they heavily rely on their linux fork | 17:17 |
| bluerise | which has like 20 dts that include each other for modularity | 17:17 |
| bluerise | and don't just apply to Linux mainline | 17:18 |
| bluerise | They do all this fancy UART/RS323/RS485 stuff and then the actual serial console is on a tiny 3 pin connector no one uses | 17:18 |
| bluerise | I don't think I found proper schematics for board/som | 17:19 |
| bluerise | with radxa/nanopi, I know that eventually linux mainline will have stuff | 17:19 |
| gsora | ah so the issue is not with the rk3588 itself, rather the board maker isn't very upstream friendly | 17:21 |
| bluerise | Oh yeah, I'm happy with the Rockchip stuff. They've been delivering with every chip. | 17:22 |
| bluerise | It's just that everytime I have things from Firefly it's all... semi-nice | 17:22 |
| gsora | we could say they make nice eval boards for us who get a reform with the rk chipset :P | 17:25 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 17:48 | |
| - xktr (QUIT: Quit: Lost terminal) (~xktr@user/xktr) | 18:08 | |
| + xktr (~xktr@user/xktr) | 18:12 | |
| - amk (QUIT: Ping timeout: 246 seconds) (~amk@user/amk) | 18:17 | |
| + amk (~amk@user/amk) | 18:19 | |
| - sterni (QUIT: Read error: Connection reset by peer) (~quassel@user/sterni) | 18:33 | |
| + sterni (~quassel@user/sterni) | 18:35 | |
| - akira (QUIT: Ping timeout: 268 seconds) (~akira@37.4.230.225) | 18:48 | |
| + akira (~akira@2a01:599:a1c:be58:4a6e:e9db:7e29:9dfd) | 18:49 | |
| - mjw (QUIT: Killed (lithium.libera.chat (Nickname regained by services))) (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 20:18 | |
| * Guest4377 -> mjw | 20:18 | |
| + Guest2827 (~mjw@2001:1c06:2488:1400:4fd:39a7:74ac:7bae) | 20:18 | |
| - akira (QUIT: Read error: Connection reset by peer) (~akira@2a01:599:a1c:be58:4a6e:e9db:7e29:9dfd) | 20:39 | |
| + akira (~akira@ip2504e6e1.dynamic.kabel-deutschland.de) | 20:39 | |
| - akira (QUIT: Read error: Connection reset by peer) (~akira@ip2504e6e1.dynamic.kabel-deutschland.de) | 20:49 | |
| - hairu (QUIT: Remote host closed the connection) (m-uotkmd@user/hairu) | 20:49 | |
| + akira (~akira@ip2504e6e1.dynamic.kabel-deutschland.de) | 20:49 | |
| + hairu (m-uotkmd@user/hairu) | 20:50 | |
| - akira (QUIT: Read error: Connection reset by peer) (~akira@ip2504e6e1.dynamic.kabel-deutschland.de) | 20:58 | |
| + akira (~akira@ip2504e6e1.dynamic.kabel-deutschland.de) | 20:58 | |
| violet | i wonder, is it technically possible to do USB device-mode over usb-c with the pocket with any of the computer modules? | 21:00 |
| - hairu (QUIT: Remote host closed the connection) (m-uotkmd@user/hairu) | 21:15 | |
| + hairu (m-uotkmd@user/hairu) | 21:16 | |
| - akira (QUIT: Ping timeout: 256 seconds) (~akira@ip2504e6e1.dynamic.kabel-deutschland.de) | 21:17 | |
| + akira (~akira@2a01:599:a1c:be58:4a6e:e9db:7e29:9dfd) | 21:23 | |
| - Twodisbetter (QUIT: Quit: Gateway shutdown) (2cc0e4ea1c@irc.cheogram.com) | 21:25 | |
| + Twodisbetter (2cc0e4ea1c@irc.cheogram.com) | 21:27 | |
| + reform7299 (~user@business-90-187-186-49.pool2.vodafone-ip.de) | 21:42 | |
| reform7299 | @josch: i think it was successful. tho there is a weird rsync error at the end | 21:43 |
| reform7299 | https://mister-muffin.de/p/pQGY.txt | 21:43 |
| josch | well... :) | 21:43 |
| josch | that one should probably be made configurable indeed | 21:43 |
| josch | reform7299: the script is meant to run on reform.debian.net | 21:44 |
| reform7299 | took 11h ish on the pocket :D | 21:44 |
| josch | and there it's supposed to fill /var/www as that is shared by the webserver | 21:44 |
| josch | reform7299: takes 5 hours on reform.debian.net | 21:44 |
| reform7299 | maybe it should get a "local_build" option | 21:44 |
| josch | reform7299: you can now set the WWWDATA environment variable to choose another location than /var/www | 21:46 |
| reform7299 | ah maybe next time :D | 21:47 |
| reform7299 | from the logs it is a bit unclear where all the builds are saved to | 21:47 |
| josch | reform7299: the last executed line in your build log has this: | 21:48 |
| josch | rsync -Pha --delete reform-debian-packages/repo/ /var/www/debian | 21:48 |
| josch | the directory reform-debian-packages/repo/ contains the debian repository | 21:48 |
| pandora | Ah yeah | 21:49 |
| pandora | I see it | 21:49 |
| pandora | I should learn to read :D | 21:49 |
| minute | our final stock of LS1028A modules (6!) now available at crowd supply: https://www.crowdsupply.com/mnt/reform#product-3748 | 21:54 |
| minute | also, stock of MNT Reform USB camera! https://www.crowdsupply.com/mnt/reform#product-3749 | 21:55 |
| minute | bluerise: yeah the som is robust so far | 22:00 |
| [tj] | gsora: the usb controller has otg | 22:04 |
| [tj] | oh sorry, I meant violet | 22:04 |
| mhoye | Oh no, did I miss my chance at the bananapi CM4 upgrade? I don't see it on the mnt site... | 22:23 |
| josch | mhoye: you will probably have an easy time get an a311d second hand from those who want to upgrade to the rk3588 :) | 22:25 |
| mhoye | haha | 22:29 |
| mhoye | Is that a drop-in option, or do I need to worry about reinstalling anything? | 22:30 |
| josch | mhoye: what do you want to drop-in for what? | 22:32 |
| bluerise | minute: etwa for the RK3588 SoM? 3 months? | 22:35 |
| mhoye | sorry - I guess my two questions are "is the a311d a drop-in replacement for the imx8 one I have" | 22:36 |
| vagrantc | mhoye: i think it is a mixed bag as i understand it | 22:36 |
| vagrantc | oh, wait, i was assuming the ls1028a ... | 22:37 |
| josch | ls1028a is a mixed bag indeed :) | 22:37 |
| josch | mhoye: when i switched from imx8mq to a311d i wrote this: https://community.mnt.re/t/migrating-encrypted-nvme-to-different-som-imx8mq-to-a311d/1783 | 22:37 |
| mhoye | my second question is, if I really want to go rogue, are "compute modules" actually a standard thing? | 22:43 |
| mhoye | by which I mean, if I am looking at http://www.orangepi.org/html/hardWare/computerAndMicrocontrollers/details/Orange-Pi-CM5.html | 22:44 |
| bluerise | This one *again* doesn't have the PHY on the module | 22:44 |
| bluerise | minute: iirc the Ethernet PHY needs to be on the CM4 module, right? | 22:45 |
| minute | bluerise: hmm yeah i think so. this is also the case on rcore, where i squeezed a phy in the corner | 22:55 |
| minute | mhoye: banana pi stock is also at crowd supply! ships immediately | 22:56 |
| mhoye | minute: what what? | 22:57 |
| mhoye | I'll take a look directly! thank you. | 22:57 |
| minute | mhoye: check the mnt reform page on CS | 22:57 |
| mhoye | Got it. So, what are the odds that an orangepi will just drop into that slot? | 22:57 |
| mhoye | I guess zero | 22:57 |
| minute | bluerise: eta for rcore: it depends. 2-3 mo realistically :/ i'm doing a final spin to iron out some pcie SI issues | 22:58 |
| minute | bluerise: in general it is in very good shape | 22:58 |
| minute | i've been daily driving it for a longer time now | 22:58 |
| minute | it is kind of like how reform should have always been, performance wise ^^ | 22:59 |
| minute | and really curious what's coming up with panvk... | 22:59 |
| bluerise | cool cool! | 23:04 |
| vagrantc | mhoye: what's the orangepi got that the banana pi module doesn't? | 23:09 |
| mhoye | The A55 chip is (reportedly) a fair bit less power hungry | 23:48 |
| mhoye | I haven't seen an apples-to-apples comparison though... | 23:48 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!