| + mjw (~mjw@gnu.wildebeest.org) | 00:29 | |
| + svp (~svp@host-79-7-240-189.business.telecomitalia.it) | 00:51 | |
| + casparvitch (~casparvit@130.102.140.49) | 01:00 | |
| minute | josch: could you ping me tomorrow somehow about this again? :D | 01:04 |
|---|---|---|
| minute | the admin interface is kind of nasty on a phone | 01:04 |
| minute | (the one for assigning runners) | 01:05 |
| - mjw (QUIT: Ping timeout: 252 seconds) (~mjw@gnu.wildebeest.org) | 01:05 | |
| - svp (QUIT: Remote host closed the connection) (~svp@host-79-7-240-189.business.telecomitalia.it) | 01:06 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 02:57 | |
| - nsc (QUIT: Ping timeout: 245 seconds) (~nicolas@129-97-142-46.pool.kielnet.net) | 03:29 | |
| + nsc (~nicolas@62-98-142-46.pool.kielnet.net) | 03:31 | |
| - paperManu_ (QUIT: Ping timeout: 265 seconds) (~paperManu@72.10.128.164) | 03:37 | |
| + photomattmills (~photomatt@192-184-254-28.fiber.dynamic.sonic.net) | 07:32 | |
| photomattmills | is there a 3d model for the friction hinges in the pocket around anywhere? I see the top and bottom case in the repo, but no hinges | 07:33 |
| - photomattmills (QUIT: Ping timeout: 272 seconds) (~photomatt@192-184-254-28.fiber.dynamic.sonic.net) | 08:04 | |
| + photomattmills (~photomatt@192-184-254-28.fiber.dynamic.sonic.net) | 08:25 | |
| josch | minute: the problem lost its urgency because I just created a branch with vimja's changes in my own namespace and tested on rk3588 pocket and classic. And vagrant tested it on non-dsi classic. It seems to work fine for everybody so i just pressed the merge button here: https://source.mnt.re/reform/reform-debian-packages/-/merge_requests/127 | 08:32 |
| josch | and now the pipeline is green again: https://source.mnt.re/reform/reform-debian-packages/-/pipelines | 08:32 |
| josch | please shout if dropping that patch was a mistake :) | 08:32 |
| - ZetaR (QUIT: Ping timeout: 272 seconds) (~user@c-98-208-141-90.hsd1.fl.comcast.net) | 08:47 | |
| - photomattmills (QUIT: Ping timeout: 272 seconds) (~photomatt@192-184-254-28.fiber.dynamic.sonic.net) | 09:03 | |
| - casparvitch (QUIT: Ping timeout: 265 seconds) (~casparvit@130.102.140.49) | 10:10 | |
| + casparvitch (~casparvit@130.102.140.49) | 10:15 | |
| - casparvitch (QUIT: Ping timeout: 260 seconds) (~casparvit@130.102.140.49) | 10:20 | |
| + mjw (~mjw@gnu.wildebeest.org) | 10:43 | |
| + casparvitch (~casparvit@36-255-114-132.ip4.superloop.au) | 11:43 | |
| - casparvitch (QUIT: Remote host closed the connection) (~casparvit@36-255-114-132.ip4.superloop.au) | 11:47 | |
| + casparvitch (~casparvit@36-255-114-132.ip4.superloop.au) | 11:48 | |
| - mjw (QUIT: Ping timeout: 260 seconds) (~mjw@gnu.wildebeest.org) | 12:27 | |
| * Guest9868 -> mjw | 12:40 | |
| - mjw (QUIT: Quit: Leaving) (~mjw@2001:1c06:2486:a800:7602:5eff:dc71:a72c) | 12:58 | |
| + gustav28 (~gustav@c-78-82-52-252.bbcust.telenor.se) | 13:02 | |
| minute | josch: ok, thanks for the initiative! i think i put that in originally because i still thought we could have a unified hdmi/dsi dts, but nah | 13:05 |
| minute | (i.e. where eDP could be disconnected and the system knows that) | 13:05 |
| + paperManu (~paperManu@72.10.128.164) | 13:12 | |
| minute | FYI we just had an instance here of the internal UART cable between the keyboard and the motherboard (LPC connection) lost connectivity on two pins (still analyzing which part of the cable broke exactly) | 13:14 |
| minute | and it was intermittent first | 13:14 |
| josch | classic reform? the new motherboard 3.0? | 13:18 |
| + mjw (~mjw@2001:1c06:2486:a800:7602:5eff:dc71:a72c) | 13:20 | |
| - casparvitch (QUIT: Ping timeout: 252 seconds) (~casparvit@36-255-114-132.ip4.superloop.au) | 13:45 | |
| minute | josch: classic reform, yeah but the motherboard doesn't matter. the problem can be in crimped cables | 14:18 |
| minute | just another variable to consider for some failure modes | 14:18 |
| dok | now called crippled cables | 14:20 |
| minute | dok: https://wid.org/ableist-language-phrases-that-you-may-unknowingly-use/ | 14:22 |
| - paperManu (QUIT: Ping timeout: 260 seconds) (~paperManu@72.10.128.164) | 14:45 | |
| dok | that was a bad joke... sorry | 14:46 |
| bremner | any thoughts from the hive-mind on the risks of operating a pocket reform without the back attached? Looking at "sensors" it seems to be steady at about 60C, which is higher than with the heatsink attached, but maybe OK? | 14:58 |
| josch | ACTION is doing that all the time because of all the cables hanging out from the back XD | 15:03 |
| bremner | exactly. | 15:03 |
| josch | [insert "i also like to live dangerously" meme] | 15:05 |
| bremner | some kind of actual connector for serial console would be very nice | 15:06 |
| josch | i've had that for the classic -- turns out, after most bugs were gone that wasn't so useful anymore | 15:08 |
| josch | so my hope is to reach this state with the pocket as well :) | 15:08 |
| bremner | hrm. This looks important https://lore.kernel.org/linux-pci/1744940759-23823-1-git-send-email-shawn.lin@rock-chips.com/ | 15:08 |
| bremner | (rk3588 PM) | 15:08 |
| josch | ah i see you got a reply from Niklas Cassel | 15:10 |
| bremner | yes | 15:10 |
| [tj] | for next it could just be on an alt mode of usb-c | 15:10 |
| [tj] | if I could figure out what the hell the usb standard actually says about usb-c we could expose the soc controllers debug interface | 15:11 |
| ch | :) | 15:16 |
| ch | not so easy to find all the relevant bits | 15:16 |
| [tj] | I'm not sure its been written down anywhere | 15:17 |
| [tj] | this is a future futile task, last months was "is a uefi pcie allocation reserved in the run time services block?" | 15:17 |
| - bkeys (QUIT: Ping timeout: 248 seconds) (~Thunderbi@h193.131.19.98.dynamic.ip.windstream.net) | 15:20 | |
| josch | ch: could you give me some pointers on how to read the firmware version of the sysctl and keyboard from a shell script? :) | 15:20 |
| josch | (without looking it up via a /sys path maybe?) | 15:21 |
| ch | josch: lsusb | 15:22 |
| ch | josch: but i was gonna write some code in reform-mcu-tool | 15:22 |
| josch | ch: please do -- the lsbusb output is not very friendly to be read by machines | 15:24 |
| josch | ch: and then we can display that information in reform-check | 15:24 |
| + bkeys (~Thunderbi@140.235.133.176) | 16:01 | |
| - bkeys (QUIT: Ping timeout: 248 seconds) (~Thunderbi@140.235.133.176) | 16:17 | |
| + bkeys (~Thunderbi@140.235.133.176) | 16:28 | |
| - bkeys (QUIT: Ping timeout: 245 seconds) (~Thunderbi@140.235.133.176) | 16:53 | |
| + bkeys (~Thunderbi@140.235.133.176) | 17:00 | |
| - bkeys (QUIT: Quit: With every step we take, danger will follow closely) (~Thunderbi@140.235.133.176) | 17:12 | |
| + bkeys (~Thunderbi@140.235.133.176) | 17:12 | |
| - bkeys (QUIT: Ping timeout: 248 seconds) (~Thunderbi@140.235.133.176) | 17:19 | |
| + wielaard (~mjw@gnu.wildebeest.org) | 17:23 | |
| + _justin_kelly716 (~justinkel@user/justin-kelly/x-6011154) | 17:23 | |
| - _justin_kelly71 (QUIT: Ping timeout: 260 seconds) (~justinkel@user/justin-kelly/x-6011154) | 17:25 | |
| * _justin_kelly716 -> _justin_kelly71 | 17:25 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 17:28 | |
| + bkeys (~Thunderbi@140.235.133.176) | 17:47 | |
| - wielaard (QUIT: Ping timeout: 265 seconds) (~mjw@gnu.wildebeest.org) | 17:56 | |
| - bkeys (QUIT: Ping timeout: 245 seconds) (~Thunderbi@140.235.133.176) | 18:01 | |
| - RandyK (QUIT: Remote host closed the connection) (~RandyK@user/randyk) | 18:29 | |
| + RandyK (~RandyK@user/randyk) | 18:29 | |
| + photomattmills (~photomatt@2a09:bac3:635e:1232::1d0:cd) | 18:31 | |
| - photomattmills (QUIT: Ping timeout: 272 seconds) (~photomatt@2a09:bac3:635e:1232::1d0:cd) | 18:36 | |
| vagrantc | for those of you wanting to use guix on mnt/reform ... i proposed updating the kernel to 6.15.x https://codeberg.org/guix/guix/pulls/1223 although the 6.12.x kernel i have there works, the patchset is a bit old and 6.15.x just works with the current patchset | 18:38 |
| vagrantc | would be nice to bisect the issues with 6.12.x sometime, though it is not a trivial git bisect, as the patches were not maintained in git (at least, not in a way that makes git bisect feasible, as far as i can tell) | 18:40 |
| vagrantc | oh, i see new kernel packages from reform.debian.net that i can at least try | 19:02 |
| josch | vagrantc: before you dump time into bisecting, you could also try and see if you find a good workflow which would make bisecting easier in the future | 19:05 |
| josch | vagrantc: maybe using dgit or whichever other git-based mechanism | 19:05 |
| vagrantc | problem is with rebasing workflows ... it is hard to bisect | 19:06 |
| vagrantc | bisect kind of only really works well when you have a linear set of patches that goes from A to B to C ... | 19:06 |
| + bkeys (~Thunderbi@140.235.133.176) | 19:07 | |
| vagrantc | though i would be happy to be proved wrong :) | 19:07 |
| vagrantc | josch: but yeah, figuring out a good workflow | 19:07 |
| josch | vagrantc: that rebasing makes bisecting hard is also my experience -- do you know of a workflow which would be easier to bisect? | 19:08 |
| vagrantc | rather than rebasing patches, applying newer and newer patchsets with changes? | 19:09 |
| vagrantc | this is just a fundamentally hard problem with maintaining patchsets outside of mainline | 19:10 |
| vagrantc | which is both necessary to get stuff working ... but hard :) | 19:10 |
| vagrantc | i can confirm that 6.15.6+1-mnt-reform-arm64 appears to work ok for me :) | 19:11 |
| vagrantc | ACTION tries 6.12.3x | 19:11 |
| josch | nice! | 19:11 |
| - bkeys (QUIT: Ping timeout: 248 seconds) (~Thunderbi@140.235.133.176) | 19:11 | |
| vagrantc | try 3, need to remember to interrupt the boot process to boot an older kernel | 19:12 |
| josch | also when you say "guix doesn't work on reform with 6.12.x" then you mean "guix doesn't work on RK3588 reform with 6.12.x" -- non rk3588 reforms might work already! :) | 19:12 |
| vagrantc | i have no experience or reports of testing my kernels on other variants on guix :) | 19:12 |
| josch | vagrantc: build me a system image and i flash sd-cards with it :) | 19:12 |
| josch | vagrantc: can i build a guix system image from Debian? | 19:13 |
| vagrantc | josch: should be able to ... although some serious CVEs ... https://bugs.debian.org/1108318 ... with no obvious fix in sight | 19:15 |
| josch | oh dear o0 | 19:16 |
| vagrantc | i mean, it is fixed upstream, but there's no backported patches to the last release and changes since .... anyways, getting a bit off-topic | 19:17 |
| vagrantc | 6.12.35 still having same troubles for me with the Debian kernels ... but the backports kernels work | 19:18 |
| vagrantc | would be curious if it is just me, or are there other rk3588 non-dsi mnt/reform classic users out there to test it? :) | 19:18 |
| vagrantc | at least it isn't crashing the LPC lately | 19:19 |
| josch | vagrantc: maybe gsora or bkeys | 19:20 |
| vagrantc | josch: strangely enough the mechanisms for building system images assume cross-compilation and do not work natively for guix, so i have balked at making an image! :) | 19:20 |
| vagrantc | but in theory it is straightforward to make a working image | 19:21 |
| josch | vagrantc: they assume cross-compilation because you compile everything on-the-fly which you want in the image? | 19:21 |
| josch | i'm on arm64 so it wouldn't be cross-compilation even though i'm on debian... | 19:21 |
| + bkeys (~Thunderbi@98.19.131.193) | 19:22 | |
| vagrantc | josch: i think too much special-casing, forcing cross-compilation for aarch64/arm64 | 19:22 |
| josch | vagrantc: oh it assumes that you are on amd64? | 19:22 |
| vagrantc | josch: basically ... although building amd64 images on amd64 works fine, so ... there is just some stupid code somewhere i haven't tracked down | 19:23 |
| + jogu (~jogu@user/jogu) | 19:23 | |
| vagrantc | josch: but yeah, i should figure out how to build an image for it | 19:24 |
| + AnimaInvicta (~AnimaInvi@88-120-179-216.subs.proxad.net) | 19:24 | |
| jogu | So I've been poking at a ZMK port for the reform keyboard and have the basics done! https://github.com/joguSD/zmk-mntre-module | 19:32 |
| jogu | Tbh, things at the moment are even more rough than my previous attempt at a QMK port but I'm convinced that ZMK is probably a better fit in the long term | 19:33 |
| jogu | Unfortunately, the USB HID instability has been present in all 3 firmwares I tried (QMK, KMK, and now ZMK) | 19:34 |
| vagrantc | josch: now you've got me looking at https://codeberg.org/vagrantc/mnt-reform-guix-config/src/branch/main/config-mnt-reform.scm which could relatively easily be adapted to make an image ... which also contains a few things i have not yet properly pushed into guix | 19:39 |
| vagrantc | josch: i am guessing the different variants use different kernel commandlines ... and probably different sets of kernel modules ... so might take some fiddling to get those working across multiple devices | 19:40 |
| vagrantc | and bring-your-own u-boot, etc. ... and you would want to probably adjust the filesystem and user and hostname and whatnot ... but that is basically "it" :) | 19:40 |
| josch | vagrantc: the different kernel cmdlines are listed in reform-tools machines.conf files | 19:43 |
| + wielaard (~mjw@gnu.wildebeest.org) | 19:43 | |
| josch | and yes, those come from u-boot anyways | 19:43 |
| josch | vagrantc: the modules are the same everywhere -- remember that we now have the "any" image which boots everywhere (if the platform already has u-boot on eMMC) | 19:44 |
| josch | jogu: nice!! | 19:44 |
| josch | jogu: is it ready for trying it out or still too much alpha-stage? | 19:44 |
| vagrantc | josch: yes, but guix has very primitive and crude module handling in initramfs, so you have to list all appropriate modules for all the platforms to get to their rootfs | 19:45 |
| vagrantc | josch: which, i might add, changes depending on the kernel version (yes, i've filed bugs!) | 19:46 |
| vagrantc | i basically start by booting the platform with one of the debian kernels and seeing what modules it has loaded and excluding the "obvious" ones | 19:47 |
| vagrantc | did i say it was primitive and crude? | 19:47 |
| vagrantc | modprobe is for the weak | 19:48 |
| josch | oh interesting | 19:48 |
| vagrantc | josch: i have much more loaded terms i would use. :) | 19:48 |
| jogu | josch: Thanks! The basics are functional (keys / rgb works) but it will require flipping the flash switch to get back into the flashing mode. | 19:48 |
| jogu | As for the USB HID stabilty -- I'm not sure if it's a firmware issue or perhaps a hardware issue with my pcb | 19:49 |
| vagrantc | nice work on alternate keyboard firmware :) | 19:49 |
| vagrantc | josch: if you could get me the lsmod output for 6.12.x and 6.15.x from other device variants .... that would go a ways towards making a multi-device image. still not sure how to implement the different kernel commandline stuff (unless u-boot always has $bootargs set or something?) | 19:52 |
| vagrantc | there was interest from someone with an imx8mq variant to boot guix, but basically have to hand them the image. | 19:53 |
| vagrantc | currently, guix just does extlinux.conf so not sure how to handle different platforms | 19:56 |
| jogu | The QMK firmware is in a better state. I have a prebuilt uf2 here: https://github.com/joguSD/qmk_firmware/releases/tag/alpha1. Hyper + R will reboot the keyboard into flash mode so it shouldn't requiring flipping the internal flash switch to go back and forth. | 20:02 |
| jogu | If there's a brave soul who's willing to test and tell me if they have USB HID stability issues :) | 20:03 |
| jogu | If someone is more interested in trying ZMK this build is best: https://github.com/joguSD/zmk-mntre-module/actions/runs/16260686996 It was before I made a migration to a newer version, but it's slightly more function in that it can reboot itself into flash mode via Hyper + X | 20:13 |
| * mjw -> Guest2879 | 20:14 | |
| - Guest2879 (QUIT: Killed (silver.libera.chat (Nickname regained by services))) (~mjw@2001:1c06:2486:a800:7602:5eff:dc71:a72c) | 20:14 | |
| * wielaard -> mjw | 20:14 | |
| + Guest2879 (~mjw@2001:1c06:2486:a800:7602:5eff:dc71:a72c) | 20:14 | |
| - robin (QUIT: Remote host closed the connection) (~robin@user/terpri) | 20:26 | |
| + robin (~robin@user/terpri) | 20:26 | |
| jfred | I've been meaning to try actually migrating over to Guix on my rk3588 Reform at some point. I booted an image built from https://codeberg.org/lykso/mnt-reform-nonguix on my SD card and it seemed to work, though didn't have much of a chance to poke at it at the time | 20:55 |
| vagrantc | jfred: i'm running it on NVMe with the guix system config referenced above ... it pulls in a couple things not appropriate for mainline guix, but those are all in the system config | 21:03 |
| vagrantc | have not looked at lykso's config in a while | 21:04 |
| vagrantc | but the kernel i am using is coming from guix official channel | 21:04 |
| vagrantc | breaks periodically with kernel updates and having to refresh the patches, but works well overall | 21:05 |
| vagrantc | need to test HDMI output to see if i feel comfortable actually dong a presentation with it :) | 21:06 |
| vagrantc | don't have a convenient HDMI monitor to plug into set up | 21:07 |
| jfred | I kinda wish guix had a standardized way of separating hardware-support config from the rest of your system config like nixos seems to(?) - I've tried to emulate it a bit in my system config but would be nice if I could just hook a reform hardware config into my system config | 21:09 |
| jfred | I've been taking advantage of the fact that several of the fields in an `operating-system` record have defaults and setting up a chain of inherited `operating-system` records: https://github.com/jfrederickson/dotfiles/blob/master/guix/guix/system/machines/terracard/hardware-configuration.scm | 21:11 |
| jfred | but it doesn't feel as clean as I would like | 21:12 |
| vagrantc | probably more savvy than my clumsy cut-and-pasting :) | 21:14 |
| jfred | oh right - no, it wasn't that a bunch of the fields had defaults, it's that I manually set some nonsensical defaults to override later: https://github.com/jfrederickson/dotfiles/blob/master/guix/guix/system/machines/desktop-base.scm#L16-L20 | 21:18 |
| jfred | it feels like what I want is maybe some new records to keep track of hardware-specific things and hardware independent things, and a more intelligent way of combining the two to output an operating-system record | 21:20 |
| - gustav28 (QUIT: Quit: Quit) (~gustav@c-78-82-52-252.bbcust.telenor.se) | 22:15 | |
| - bleb (QUIT: Ping timeout: 260 seconds) (~cm@user/bleb) | 22:26 | |
| + bleb (~cm@user/bleb) | 22:27 | |
| - bleb (QUIT: Ping timeout: 260 seconds) (~cm@user/bleb) | 22:40 | |
| + bleb (~cm@user/bleb) | 22:40 | |
| - bleb (QUIT: Ping timeout: 248 seconds) (~cm@user/bleb) | 22:52 | |
| + bleb (~cm@user/bleb) | 22:58 | |
| + photomattmills (~matt@192-184-254-28.fiber.dynamic.sonic.net) | 23:07 | |
| photomattmills | finally realized since the pocket doesn't sleep it's a perfect irssi/screen host | 23:07 |
| + casparvitch (~casparvit@36-255-114-132.ip4.superloop.au) | 23:08 | |
| - vagrantc (QUIT: Ping timeout: 276 seconds) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 23:19 | |
| - photomattmills (QUIT: Changing host) (~matt@192-184-254-28.fiber.dynamic.sonic.net) | 23:49 | |
| + photomattmills (~matt@user/photomattmills) | 23:49 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 23:51 | |
| - photomattmills (QUIT: Quit: leaving) (~matt@user/photomattmills) | 23:53 | |
| + photomattmills (~matt@user/photomattmills) | 23:54 | |
| + svp (~svp@host-79-7-240-189.business.telecomitalia.it) | 23:56 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!