| ^alex | josch, thanks, we got the new system image flashed over the MMC with that link | 00:02 |
|---|---|---|
| + xktr (~xktr@user/xktr) | 00:02 | |
| - sbp (QUIT: Ping timeout: 246 seconds) (~sbp@tea.infomesh.net) | 00:26 | |
| + sbp (~sbp@tea.infomesh.net) | 00:26 | |
| - sbp (QUIT: Changing host) (~sbp@tea.infomesh.net) | 00:58 | |
| + sbp (~sbp@apache/doge/sbp) | 00:58 | |
| - sevan (QUIT: Ping timeout: 272 seconds) (~sevan@2001:470:1f1d:1d6:5a55:caff:fe24:ed4) | 01:46 | |
| - mjw (QUIT: Ping timeout: 248 seconds) (~mjw@gnu.wildebeest.org) | 01:47 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 01:56 | |
| + sevan (~sevan@2001:470:1f1d:1d6:5a55:caff:fe24:ed4) | 02:02 | |
| - mtm (QUIT: Ping timeout: 276 seconds) (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 02:03 | |
| + mtm (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 02:04 | |
| ^alex | and we've got the keyboard layout _mostly_ tweaked for our tastes :) still not happy with the bottom row at the moment | 02:10 |
| - mtm (QUIT: Ping timeout: 260 seconds) (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 03:15 | |
| - gustav28 (QUIT: Quit: Quit) (~gustav@c-5234524e.019-141-67626730.bbcust.telenor.se) | 03:25 | |
| + mtm (~textual@c-71-228-84-213.hsd1.fl.comcast.net) | 03:39 | |
| - bkeys (QUIT: Remote host closed the connection) (~Thunderbi@45.134.140.153) | 04:59 | |
| - jn (QUIT: Ping timeout: 252 seconds) (~quassel@user/jn/x-3390946) | 05:12 | |
| + jn (~quassel@2a0a-a549-b49d-0-20d-b9ff-fe49-15fc.ipv6dyn.netcologne.de) | 05:13 | |
| - jn (QUIT: Changing host) (~quassel@2a0a-a549-b49d-0-20d-b9ff-fe49-15fc.ipv6dyn.netcologne.de) | 05:13 | |
| + jn (~quassel@user/jn/x-3390946) | 05:13 | |
| ^alex | wooo time to unbrick the keyboard | 05:18 |
| josch | minute: i want to move the files in /etc/skel as well as /etc/motd from reform-tools to reform-system-image. Rationale: packages are not allowed to ship files in /etc/skel in Debian and the {DATE} in /etc/motd should not reflect the reform-tools date but the system-image creation date, so this helps the package being uploaded into Debian. Do you see any downsides of such a move? | 07:12 |
| josch | I'm also thinking about breaking out /usr/share/fonts/Iosevka-Term to its own package because we are not building the font from source right now | 07:13 |
| josch | or also ship the font by reform-system-image and put it into /usr/local/share/fonts/ | 07:13 |
| josch | list of things to package before iosevka can be in debian: https://wiki.debian.org/Javascript/Nodejs/Tasks/Iosevka | 07:17 |
| josch | uuuh Gürkan Myczko is working on packaging it... i contacted them | 08:00 |
| + laumann (~quassel@2a0a-e5c0-2-2-0-c8ff-fe68-bef1.loves.ipv6.at.ungleich.ch) | 10:02 | |
| - laumann (QUIT: Changing host) (~quassel@2a0a-e5c0-2-2-0-c8ff-fe68-bef1.loves.ipv6.at.ungleich.ch) | 10:03 | |
| + laumann (~quassel@user/laumann) | 10:03 | |
| josch | ch: i was thinking more about the firmware update situation and i think i like the idea of a dedicated usb-stick or sd-card more for a reason which only just occurred to me: the systemd solution will not work for people running openbsd, plan9 or any other OS without systemd and i think it would be useful that whatever solution is implemented works independent on what OS the user is running | 10:17 |
| josch | so similar to back in the day where bios updates worked by users inserting a floppy drive with ms-dos on it which also worked independent on the OS the user had installed | 10:18 |
| ch | josch: yes its useful (maybe "required") to have that, but also its annoying to use. if the systemd thing can exist additionally, maybe by reusing 95% of what you have for the boot disk version that'd be nice | 10:37 |
| ch | josch: i guess fwupdate.sh could be make-fwupdate-bootdisk.sh which does the mmdebstrap, sets up uboot/whatever, does the flash using picotool using the same systemd-update infra, and then that resulting image can be reused and mounted as nextroot | 10:38 |
| josch | ch: yes, that can certainly be done -- it's just a question of how much time there is in a day to support X things :) | 10:40 |
| ch | sure | 10:40 |
| josch | ch: i can give you git push commits if you like to help ;) | 10:40 |
| josch | s/commits/rights/ | 10:41 |
| ch | but if someone works on the bootdisk variant anyway, as long as it is systemd-based, its easy to stick into the nextroot machinery | 10:41 |
| ch | i was defering any real implementation work until i got hardware to try it out on, but yeah why not | 10:41 |
| josch | ch: you are waiting for a pocket? | 10:42 |
| ch | yep | 10:42 |
| josch | nice :) | 10:42 |
| ch | (and an RK module, but thats even further out i think) | 10:42 |
| josch | not by that much! apparently the new adapter board revision got ordered recently :) | 10:43 |
| ch | :) | 10:43 |
| - qbit (QUIT: Remote host closed the connection) (~qbit@mail.suah.dev) | 11:15 | |
| + qbit (~qbit@mail.suah.dev) | 11:16 | |
| josch | Creating a "very minimal" rescue disk does not really seem to be possible with requirements like "display is working" and others. So maybe we are back at ch's idea and use the systemd methods and users who do not use Debian on their systems can just download and flash the latest system image onto an sd-card or usb-stick and run the flasher from there. | 12:08 |
| josch | This also makes me again think about having the reform-specific firmware (keyboard, trackball, lpc...) available in a debian package because the rescue sd-card will not necessarily have internet, so we'd like to ship the firmware as a deb to integrate it into the image. This would then also allow the firmware to be distributed via the MNT apt repo instead of downloading it from CI. | 12:11 |
| josch | Since the pico-sdk itself is also packaged, maybe it would not even be too difficult to have a reform-firmware package in Debian main. | 12:12 |
| josch | But then i again think about the way that the firmware is currently packaged in one big mono-repo and how debian tooling is not really prepared for dealing with extracting subdirectories from mono-repos or with writing debian/watch files which make uscan download the necessary parts automatically. | 12:14 |
| josch | ch: any ideas regarding my musings? | 12:14 |
| ch | josch: my expierence from other projects tells me that 'minimal environments for X' that are meant to work 'everywhere' tend to be not very minimal, and indeed more or less often are a copy of the real thing - thus it makes sense to me to just provide everything in, say, "debian", and find a way of reusing 99% and a few special things to make flashing work safely | 12:27 |
| ch | cant the mono-repo be a single (source) package in debian? | 12:27 |
| ch | (not sure what exactly you have in mind) | 12:27 |
| jn | one debian source package can produce many binary packages, so maybe that's how to deal with a big monorepo | 12:27 |
| + mjw (~mjw@gnu.wildebeest.org) | 12:59 | |
| - ptrc (QUIT: Remote host closed the connection) (~ptrc@ptrc.gay) | 13:13 | |
| + ptrc (~ptrc@ptrc.gay) | 13:14 | |
| josch | ch, jn: true but the mono-repo contains nearly 12k files with 720 MB of data... | 13:15 |
| ch | d/copyright has some Exclude filter and i think we have tools using that ;-) | 13:16 |
| ch | (to build an orig tarball) | 13:16 |
| josch | ch: yes, but Files-Excluded is very bad at expressing things like "everything in this directory, except..." | 13:18 |
| ch | :-/ | 13:18 |
| josch | so one way or the other this is a maintenance nightmare for downstreams | 13:18 |
| josch | i may bite the bullet at some point... | 13:19 |
| ch | in the mnt repo there could also be an dist-firmware-tarball make target or similar | 13:19 |
| ch | still not the nicest thing in the world, but maybe better than pushing the filtering downstream | 13:20 |
| josch | yes, maybe having this would be nice anyways, because git tag in a mono-repo get applied to everything and it is not very meaningful to create a tag for the keyboard firmware and let that bump the version for the trackball and lpc firmware as well | 13:20 |
| ch | agreed | 13:21 |
| josch | minute: i'll give packaging the mono-repo another try. I created a new tag to try and package the latest version: https://source.mnt.re/reform/reform/-/merge_requests/70 | 14:18 |
| + bkeys (~Thunderbi@45.134.140.153) | 14:58 | |
| josch | ch: did you know that Files-Included is a thing? :) | 15:04 |
| ch | no :) | 15:05 |
| - bkeys (QUIT: Remote host closed the connection) (~Thunderbi@45.134.140.153) | 15:07 | |
| josch | ch: me neither -- in our defense, there are only 20 source packages in the archive which make use of this uscan feature :) | 15:07 |
| josch | and it's not documented in the uscan man page either | 15:07 |
| ch | does it work? :) | 15:11 |
| josch | i'm waiting for minute to create another tag, then i'll give it a try | 15:11 |
| + bkeys (~Thunderbi@45.134.140.153) | 15:21 | |
| - bkeys (QUIT: Remote host closed the connection) (~Thunderbi@45.134.140.153) | 15:22 | |
| + bkeys (~Thunderbi@45.134.140.153) | 15:24 | |
| + bkeys1 (~Thunderbi@45.134.140.153) | 15:35 | |
| - bkeys (QUIT: Ping timeout: 265 seconds) (~Thunderbi@45.134.140.153) | 15:37 | |
| * bkeys1 -> bkeys | 15:37 | |
| - bkeys (QUIT: Remote host closed the connection) (~Thunderbi@45.134.140.153) | 15:41 | |
| josch | ch: it works kinda | 16:21 |
| josch | ch: the problem is, that Files-Excluded really only excludes files and leaves 740 empty directories | 16:21 |
| josch | "find . -type d -empty -delete" gets rid of them but they are still part of the orig tarball | 16:22 |
| josch | minute: reform2-keyboard-fw/lufa-master and reform2-trackpad-fw/lufa-master seem to be identical -- maybe one can be a symlink to the other? | 16:30 |
| josch | the lufa-master directory is 21 MB large... | 16:30 |
| - jn (QUIT: ) (~quassel@user/jn/x-3390946) | 16:31 | |
| - cobra (QUIT: Ping timeout: 244 seconds) (~cobra@user/Cobra) | 16:57 | |
| + cobra (~cobra@user/Cobra) | 16:59 | |
| + jn (~quassel@cgn-89-1-58-183.nc.de) | 17:21 | |
| - jn (QUIT: Changing host) (~quassel@cgn-89-1-58-183.nc.de) | 17:21 | |
| + jn (~quassel@user/jn/x-3390946) | 17:21 | |
| * mjw -> Guest5415 | 17:57 | |
| * Guest9247 -> mjw | 17:57 | |
| + NanoCodeBug (~NanoCodeB@c-73-35-191-67.hsd1.wa.comcast.net) | 19:05 | |
| NanoCodeBug | I'm interested in bringing up the raspi CM4 for the pocket reform. | 19:10 |
| NanoCodeBug | I have a raspi CM4, pocket reform, and cm4 io board. Is the only known issue adding the dts overlay for the jdi lt070me05000? | 19:10 |
| * mjw -> Guest6299 | 19:56 | |
| * Guest5415 -> mjw | 19:56 | |
| ^alex | w/r/t the Industrial Ethernet port, is there an adapter from that to an RJ45 _jack_? we only see plug-to-plug. | 19:57 |
| ^alex | or could we Just(tm) use an rj45 coupler to extend the cable | 19:58 |
| + gustav28 (~gustav@c-5234524e.019-141-67626730.bbcust.telenor.se) | 20:02 | |
| minute | NanoCodeBug: yep, but also tweak the jdi driver a bit to make sure it is tolerant of non existing gpios | 20:23 |
| minute | NanoCodeBug: take a look at my jdi driver patches in reform-system-image repo | 20:24 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 20:24 | |
| minute | ^alex: i would like to have that kind of adapter too. i forget to design one | 20:24 |
| minute | just a smol pcb with ix plug and eth jack | 20:25 |
| jfred | I was looking around for one of those adapters but yeah doesn't seem like anyone makes them | 20:26 |
| jfred | I ordered an ix cable off crowd supply to go with my pocket which has come in handy already | 20:28 |
| chartreuse | Maybe a very short ix cable with a coupler heat shrunk onto it | 20:29 |
| jfred | I should probably test my usb-c docking station with the pocket too, I know the DP alt-mode won't work but it has an ethernet port and stuff too... whether or not it'll charge will be the big question I suppose | 20:29 |
| jfred | USB-C cable plus microHDMI would be alright for docking the pocket | 20:30 |
| minute | jfred: with updated sysctl fw there is a good chance it will charge | 20:30 |
| jfred | I have had mixed success with USB-C chargers still (unless there's a newer fw version than the one I flashed from the forum a while back;) | 20:32 |
| ^alex | minute, the poster frame of the video in the mouser marketing announcement for the ix ethernet port shows exactly that pcb lmao | 20:32 |
| minute | josch: good call @ lufa | 20:32 |
| josch | minute: i'm preparing a lufa MR for you right now -- i'll finish it tonight | 20:33 |
| jfred | ^alex: lmao you're right | 20:33 |
| minute | reform repo also potentially contains some cruft | 20:33 |
| minute | josch: if you're mainly concerned with packaging firmware sources, are you already cherry picking those folders only? | 20:34 |
| ^alex | also yeah our Belkin USB-C charger only charges the Pocket when it's off, and only through the Thunderbolt cable, and it doesn't like any of our apple chargers; our next thought is to do a coldstart by shutting the power off and tapping the uart on the pins externally | 20:34 |
| jfred | ^alex: oh though I guess you'd want a male ix connector on one side, that PCB has a female connector | 20:34 |
| minute | (typing this from a train being on a mega delayed and diverted train ride) | 20:34 |
| minute | ^alex: this is with updated sysctl firmware? | 20:35 |
| ^alex | minute, yeah, we've flashed the updated sysctl | 20:35 |
| minute | hmm | 20:35 |
| ^alex | we'll probably try and capture debug logs after work today or sometime this weekend | 20:36 |
| josch | minute: yes, it turned out that i didn't know something that was possible to do with debian tooling and i'm using that right now to just pick the foo-fw directories out of the mono repo | 20:36 |
| minute | ah you mean sniffing the usb-c pd comms or the rp2040 sysctl uart? that's available internally as ttyACM0 | 20:36 |
| minute | josch: great | 20:36 |
| ^alex | yes, but we can't sniff that when the machine is off or being coldstarted | 20:37 |
| minute | ah, makes sense | 20:37 |
| ^alex | we don't have a PD protocol debugger | 20:37 |
| minute | there isn't a physical uart to sniff though | 20:37 |
| minute | it's virtual via usb | 20:37 |
| jfred | minute: yeah just tried again, even with updated sysctl my anker USB-PD charger won't charge it | 20:37 |
| minute | you could reuse the uart facing the keyboard and divert printf debugging to that | 20:38 |
| ^alex | wait, the syscon doesn't log to the UART that the SOC isn't using? | 20:38 |
| minute | (from inside sysctl fw) | 20:38 |
| minute | ^alex: afaik not | 20:38 |
| ^alex | rip | 20:39 |
| minute | ^alex: ah, i don't remember if i wired up an uart from sysctl to soc | 20:39 |
| minute | did i? if so, then you can target that in the fw | 20:39 |
| ^alex | lemme stare at the schematics | 20:39 |
| minute | ^alex: what about using acm0 over sysctl usb-c? too slow to react? | 20:40 |
| minute | (with alternate jumper setting) | 20:40 |
| minute | (on coldstart you may miss what's happening in the first second or so) | 20:41 |
| ^alex | if there were exposed SWD pads we could attach our picoprobe to it | 20:42 |
| ^alex | also we're going to try and look for a right-angle USB-C connector with the right clearance so we don't have to carefully balance the motherboard on cardboard while we're doing this | 20:43 |
| ^alex | lol | 20:43 |
| ^alex | minute, the syscon approach should work, we can coldstart it while it's not plugged in and wait for the syscon to enumerate | 20:47 |
| minute | ^alex: yeah. i use tio for that job | 20:49 |
| minute | ^alex: also i use a right angle usb-c plug adapter | 20:49 |
| ^alex | we use `tio` on the pocket but our macbook has `cu` so we use that there | 20:51 |
| ^alex | also yeah we just discovered the existence of those | 20:51 |
| ^alex | actually maybe we'll hack the syscon so it logs to a ring buffer | 20:59 |
| ^alex | and dumps the buffer on command | 20:59 |
| ^alex | there's enough ram in there for a chonky buffer iirc | 20:59 |
| - hairu (QUIT: Remote host closed the connection) (m-uotkmd@user/hairu) | 21:20 | |
| + hairu (m-uotkmd@user/hairu) | 21:21 | |
| - gustav28 (QUIT: Quit: Quit) (~gustav@c-5234524e.019-141-67626730.bbcust.telenor.se) | 22:15 | |
| + reform17660 (~oskar@2001:a61:10f9:b501:3c4f:920d:9282:f0eb) | 22:20 | |
| reform17660 | hi | 22:21 |
| ^alex | henlo | 22:21 |
| reform17660 | this is my first start with pocket reform - looks like much to learn | 22:23 |
| reform17660 | so for now - i need to rtfm ;) | 22:24 |
| - reform17660 (QUIT: Quit: reform17660) (~oskar@2001:a61:10f9:b501:3c4f:920d:9282:f0eb) | 22:26 | |
| - colinsane (QUIT: Quit: bye) (~colinunin@97-113-150-69.tukw.qwest.net) | 23:23 | |
| + colinsane (~colinunin@97-113-150-69.tukw.qwest.net) | 23:26 | |
| ^alex | oh hey openbsd-current boots to something that vaguely recognizes some devices | 23:46 |
| ^alex | but freebsd gripes that there's no device tree and then throws a panic | 23:46 |
| - colinsane (QUIT: Quit: bye) (~colinunin@97-113-150-69.tukw.qwest.net) | 23:48 | |
| + colinsane (~colinunin@97-113-150-69.tukw.qwest.net) | 23:51 | |
| ^alex | but really what we want is ARM64 Haiku >.> | 23:52 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!