| josch | minute: gschema overrides do not support relocatable schemas and org.gnome.settings-daemon.plugins.media-keys.custom-keybinding is exactly that. So we cannot add the <Super>Return shortcut via an override file | 00:15 |
|---|---|---|
| josch | alternative would be to add a dconf key file in /etc/dconf/db/local.d/reform which will then be used to populate the initial dconf database for each user | 00:19 |
| + bkeys (~Thunderbi@h211.16.186.173.dynamic.ip.windstream.net) | 00:22 | |
| josch | (doing that now) | 00:26 |
| josch | minute: you called the thing system image V5 in the merge request title, so I also adjusted /etc/motd accordingly, okay? | 00:33 |
| - mjw (QUIT: Ping timeout: 252 seconds) (~mjw@gnu.wildebeest.org) | 00:43 | |
| josch | okay, i have to give up for today | 01:05 |
| josch | only 5 hours of sleep left | 01:05 |
| josch | sorry | 01:05 |
| josch | please have a look at the top commits in reform-system-image gnome branch to see where i left | 01:05 |
| + robin_ (~robin@user/terpri) | 01:34 | |
| minute | josch: thanks for all your testing and fixing work! i had a rare social meeting outside of work :D also now you know why i often do these "brutalist" solutions like the original reform-gnome-config script file (because it's often quite painful to figure out the "proper" way to do things like this which requires a bunch of contortions and different files) | 01:35 |
| minute | josch: (still, i appreciate very much your perseverance in the name of more correct solutions) | 01:35 |
| minute | josch: oh and yeah, i guess V5 is good | 01:37 |
| - robin (QUIT: Ping timeout: 276 seconds) (~robin@user/terpri) | 01:37 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 02:08 | |
| + robin__ (~robin@user/terpri) | 02:08 | |
| - robin_ (QUIT: Ping timeout: 276 seconds) (~robin@user/terpri) | 02:11 | |
| + aloo_shu_ (~aloo_shu@90.166.99.87) | 02:20 | |
| * aloo_shu -> Guest9192 | 02:20 | |
| - Guest9192 (QUIT: Killed (copper.libera.chat (Nickname regained by services))) (~aloo_shu@85.51.17.10) | 02:20 | |
| * aloo_shu_ -> aloo_shu | 02:20 | |
| - paperManu (QUIT: Ping timeout: 244 seconds) (~paperManu@107.159.213.145) | 03:16 | |
| - nsc (QUIT: Ping timeout: 252 seconds) (~nicolas@234-96-142-46.pool.kielnet.net) | 03:32 | |
| + nsc (~nicolas@113-99-142-46.pool.kielnet.net) | 03:33 | |
| - Gooberpatrol66 (QUIT: Ping timeout: 244 seconds) (~Gooberpat@user/gooberpatrol66) | 04:07 | |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 04:08 | |
| josch | minute: haha no worries, i am quite aware how one can easily pour *days* on any subject matter to do it "properly" :D | 05:44 |
| josch | minute: IT WORKS! \o/ | 06:07 |
| josch | new reform-tools version 1.73 published | 06:19 |
| josch | and also uploaded to experimenta | 06:38 |
| - n_to (QUIT: Quit: quitidiquit) (~n_to@2a03:4000:6:3662:24b1:57ff:fec6:76c1) | 06:49 | |
| + chomwitt (~chomwitt@2a02:85f:9a00:8300:1ac0:4dff:fedb:a3f1) | 06:49 | |
| + n_to (~n_to@2a03:4000:6:3662:24b1:57ff:fec6:76c1) | 06:49 | |
| gsora | seeing you folks working on the reform debian image makes me scared of trying a different OS lol | 09:19 |
| + mjw (~mjw@178.230.17.0) | 09:36 | |
| - Guest2164 (QUIT: Ping timeout: 276 seconds) (~mjw@2001:1c06:2486:a800:7602:5eff:dc71:a72c) | 09:41 | |
| josch | gsora: don't be! we go through the pain now so that others don't have to :D | 09:42 |
| gsora | luckily it'll be linux, just a different flavor, porting over stuff hopefully won't be that hard | 09:45 |
| - mjw (QUIT: Read error: Connection reset by peer) (~mjw@178.230.17.0) | 09:54 | |
| + mjw (~mjw@178.230.17.0) | 09:55 | |
| minute | josch: awesome! | 10:02 |
| - mjw (QUIT: Ping timeout: 276 seconds) (~mjw@178.230.17.0) | 11:09 | |
| + mjw (~mjw@2001:1c00:2a07:3a01:e4e2:af5b:bbd7:5c24) | 11:36 | |
| - chomwitt (QUIT: Ping timeout: 272 seconds) (~chomwitt@2a02:85f:9a00:8300:1ac0:4dff:fedb:a3f1) | 11:36 | |
| grimmware | minute: have you ever had an issue with jlcpcb's support messages throwing a 501 error? | 12:08 |
| grimmware | I ordered some boards from them a while ago and they've been stuck at my local delivery office for 3 weeks and I can't get them to respond to the support case in any way I can actually interact with | 12:09 |
| grimmware | it's infuriating | 12:09 |
| - Gooberpatrol66 (QUIT: Quit: Konversation terminated!) (~Gooberpat@user/gooberpatrol66) | 12:11 | |
| minute | grimmware: had lots of glitches on jlc yes | 12:22 |
| minute | josch: now at work, will check out everything | 12:23 |
| + LainIwakura (~LainIwaku@user/LainIwakura) | 12:38 | |
| + paperManu (~paperManu@107.159.213.145) | 12:40 | |
| + chomwitt (~chomwitt@2a02:85f:9a00:8300:1ac0:4dff:fedb:a3f1) | 12:42 | |
| minute | josch: reform-debian-packages !113 merged | 12:49 |
| - GNUmoon (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 12:59 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 13:00 | |
| + gustav28 (~gustav@c-78-82-54-59.bbcust.telenor.se) | 13:02 | |
| grimmware | minute: oh never mind they've just responded to my email about it and I only have to wait a couple more days until I can complain at them again that the carrier's done nothing to solve the problem - hopefully this means they'll just cook my boards again and priority ship them | 13:08 |
| grimmware | I'll stick to emailing them in the future I think, turns out they were pretty on it, their ticketing system is just broken | 13:08 |
| josch | minute: nice, and i just arrived home again from work -- i can now test thing as well :) | 13:27 |
| josch | minute: i don't mind your "brutalist" solutions -- i rather envy you that you are able to prioritize your time like that. I would 100% easily sink days of work into a tiny detail like "<Super>Enter" should work without a start-up script. :D | 13:28 |
| josch | in this case, i made sure to leave a lot of code comments and write along commit message to make sure that you or me in 1+ years can make sense of all of this (and yes, i'm still going to write a longer article about the architecture) | 13:28 |
| josch | minute: did you already test the latest image and can confirm that <Super>Enter *does* open the terminal? | 13:29 |
| josch | minute: since you merged reform-debian-packages !113, do you want to write a post about what this entails or would you like me to take care of that? | 13:33 |
| - GNUmoon (QUIT: Remote host closed the connection) (~GNUmoon@gateway/tor-sasl/gnumoon) | 13:35 | |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 13:38 | |
| - chomwitt (QUIT: Ping timeout: 276 seconds) (~chomwitt@2a02:85f:9a00:8300:1ac0:4dff:fedb:a3f1) | 13:51 | |
| minute | josch: if you don't mind, i'd appreciate you writing a post of course | 14:05 |
| minute | i'm doing some real life test with the newest any-image now | 14:06 |
| josch | nice! | 14:06 |
| minute | btw about SPI, my current status from last night: in the driver, we always write 4 bytes and then expect to read 8 bytes from lpc. what worries me is that we have to wait (currently 50ms!) between the read and the write to get valid data at least from classic lpc. to me this is a code smell | 14:08 |
| minute | ok, SPI itself doesn't have a mechanism to signal "readiness" | 14:12 |
| minute | so to really fix this, the protocol needs to be adjusted indeed, on both device classes (classic and pocket, and in next i can include this from the start) | 14:12 |
| minute | i guess we can do a handshake for the driver to read the lpc protocol version (or just use the fw version) and if it's not an updated one, fall back to legacy behavior with fixed delays | 14:13 |
| minute | and there can be a sysfs file that we can check from userland to suggest the user should update the fw | 14:14 |
| - LainIwakura (QUIT: Quit: Client closed) (~LainIwaku@user/LainIwakura) | 14:20 | |
| minute | josch: there is an accessibility menu in gdm. happily i enabled screen reader, but it talks about "this is the dummy output module, check speech dispatcher logs slash slash many slashes". could we quickly fix this somehow? | 14:21 |
| [tj] | minute: can't you try and not wait for the read and in that case decided the SPI is running on classic fw? | 14:23 |
| josch | minute: maaaaybe but i'd have to read up on this first -- probably not before tonight | 14:26 |
| minute | josch: ok, no problemo | 14:26 |
| minute | josch: "software" app "explore" is empty, hm | 14:27 |
| minute | josch: do you have any reservations about shipping flatpak also? | 14:30 |
| + chomwitt (~chomwitt@2a02:85f:9a00:8300:1ac0:4dff:fedb:a3f1) | 14:34 | |
| josch | (now testing the latest gnome system image myself as well) | 14:37 |
| josch | minute: no, i don't think installing flatpak should break anything as far as i know | 14:38 |
| minute | josch: cool! btw i got screen reader to work | 14:39 |
| minute | josch: i installed "espeakup" and "speech-dispatcher-espeak-ng" but not sure which of them made it work | 14:40 |
| minute | also it would be great to enable "always show accessibility menu" in the gschemas. the accessibility menu is gone by default after logging in (also the screenreader is off after turning it on in gdm) | 14:40 |
| josch | huzzah i just confirmed myself that <Super>Enter open a terminal! so many hours of my life not completely wasted XD | 14:42 |
| josch | minute: feel free to add any dependencies that you think should be part of a "full" reform install to the reform-desktop-full meta package which you can edit in reform-debian-packages:reform-tools/debian/control | 14:43 |
| josch | minute: if th accessibility menu is gone after logging in even though it was turned on in gdm, i think that is a serious bug in gdm/gnome -- i'll see whether i can reproduce it and then report it to the gnome team | 14:44 |
| minute | josch: thanks! yes, the customizations work very well, well done | 14:45 |
| josch | really, the only problem was the custom key binding... | 14:45 |
| minute | josch: so the general experience is already very good, just some rough edges around "software" and accessibility (which was a bit of a random rabbithole that i just went into because i was curious how far we are away from that) | 14:45 |
| josch | the rest is just copypasting from "gsettings list-recursively" | 14:45 |
| minute | josch: ah yeah the dconf object thing :D | 14:46 |
| josch | yes, i now understand why you manually ran dconf in reform-gnome-setup script | 14:46 |
| minute | damn, suspend happened | 14:51 |
| minute | or something like that, display is off and system can't be woken up | 14:52 |
| minute | josch: after power cycle, gdm does _not_ work | 14:53 |
| minute | josch: i get a console login | 14:54 |
| minute | josch: but it's not greetd's fault, as that's really gone. rather, gdm isn't enabled properly in a reboot-tolerant way | 14:55 |
| minute | josch: systemctl status gdm says "Active: inactive (dead)" | 14:56 |
| minute | so, how to revive it properly... | 14:56 |
| minute | josch: maybe interesting debian discussion topic https://dragonscave.space/@fireborn/114585559582853752 | 15:00 |
| minute | dpkg-reconfigure gdm3 also fails ("gdm.service is not active, cannot reload." | 15:04 |
| minute | i guess the issue is that one can "systemctl disable gdm3" but not "systemctl enable gdm3" (that is broken) | 15:04 |
| minute | ah, maybe the dkpg-reconfigure error is a red herring | 15:06 |
| josch | minute: are you running the latest system image from CI or did you upgrade an existing system? | 15:06 |
| josch | oh dear i just got the same issue | 15:07 |
| josch | this didn't happen before -- what changed since yesterday that could've caused this? | 15:07 |
| josch | i'm quite certain that yesterday it worked but maybe it's not 100% deterministic? | 15:07 |
| josch | yes, i get a console login | 15:08 |
| minute | josch: in any case, the fix is dpkg-reconfigure gdm3 and then reboot. the error is meaningless | 15:08 |
| minute | so i'll add the dpkg-reconfigure line to reform-setup.sh | 15:09 |
| josch | minute: if "dpkg-reconfigure gdm3" fixes it, then there is something in the gdm3 postinst maintainer script that fixes this -- i'm trying to figure out which bit in the script (it's rather length) fixes it | 15:10 |
| josch | because then maybe we can already run whatever it is in mkimage.sh in reform-system-image | 15:10 |
| josch | i am puzzled why it works at all then during the first boot... | 15:11 |
| minute | weird, now the Software app is populated with lots of stuff, but i don't know why | 15:11 |
| josch | classic gnome ¯\_(ツ)_/¯ | 15:12 |
| minute | josch: how should i best add these packages: espeak-ng espeak-ng-data speech-dispatcher-espeak-ng flatpak gnome-software-plugin-flatpak ? | 15:14 |
| josch | minute: add them here: https://source.mnt.re/reform/reform-debian-packages/-/blob/main/reform-tools/debian/control?ref_type=heads#L76 | 15:15 |
| josch | minute: i wrote a small readme about this directory here: https://source.mnt.re/reform/reform-debian-packages/-/blob/main/reform-tools/README?ref_type=heads | 15:16 |
| bkeys | WOooooo! guys my rk3588 module shipping notification just hit | 15:20 |
| josch | \o/ | 15:20 |
| josch | bkeys: the rk3588 is *very* impressive, you will have lots of fun with it! | 15:20 |
| bkeys | It looks impressive in the videos | 15:20 |
| bkeys | I'm hoping I can put edk2 on the emmc and run mainline distros | 15:20 |
| bkeys | I have actually been wanting to try CentOS Stream out as a desktop | 15:21 |
| kfx | it's just like running an old fedora :) | 15:21 |
| kfx | has anyone done an openbsd on rk3588 yet | 15:21 |
| josch | indeed, running dpkg-reconfigure gdm3 fixes gdm. I tried running various parts of the postinst before without success... will continue to investigate later | 15:23 |
| gsora | kfx: does openbsd support rk3588? | 15:40 |
| gsora | oh, you meant in general maybe, thought you were referring specifically to reform rk3588 :) | 15:40 |
| minute | josch: thx!! | 15:41 |
| gsora | apparently openbsd supports https://wiki.radxa.com/Rock5/5b | 15:41 |
| bkeys | minute, josch: What is the extent of what you know about rk3588 and edk2 on the Reform? | 15:55 |
| minute | bkeys: i've tested edk2 a bit on sd card and booted a bunch of OSes over EFI | 15:56 |
| bkeys | Is there a reason you wouldn't be able to have it on emmc? | 15:56 |
| minute | bkeys: no idea | 15:56 |
| bkeys | It'd be neat (and nice) to have the SD card slot freed | 15:57 |
| bkeys | I'll look into it when I get mine | 15:57 |
| josch | bkeys: the sd-card slot can also be freed with u-boot | 15:57 |
| josch | bkeys: i have my u-boot on emmc and thus a free sd-card slot | 15:57 |
| bkeys | Yes I know | 15:58 |
| josch | oh then i misunderstood you | 15:58 |
| bkeys | Yeah I'm saying it would be nice regardless of boot firmware to have a free SD card slot | 15:58 |
| josch | yes, and edk2 on the rk3588 would definitely be a great option to have! | 15:58 |
| bkeys | It's that when I get rk3588 I'm hoping to use edk2 because edk2 will make it far easier for me to run RH based distros, since they usually require EFI (which I know uboot can support, but I haven't achieved yet) | 15:58 |
| bkeys | edk2 in my experience with it on other boards "just works" which is what I want, I like hacking on stuff once I boot my distro, but getting a distro to boot is what I can find frustrating | 15:59 |
| bkeys | That and I could boot from USB devices and such, like a normal PC | 15:59 |
| josch | (my u-boot can also boot from a usb thumb drive :) ) | 16:00 |
| josch | (but i know what you mean) | 16:00 |
| josch | bkeys: i'm interested in your results because then we could prepare system images that just boot with grub instead of flash-kernel/u-boot-menu | 16:03 |
| bkeys | Yeah that's gonna be the first thing I try is getting edk2 on emmc | 16:03 |
| josch | it would make booting the reform more like a "normal computer" which is a plus, i think | 16:03 |
| bkeys | If I succeed at that then I'm gonna get CentOS Stream 10 on there | 16:03 |
| bkeys | It'd be a huge plus to me | 16:04 |
| josch | nice! | 16:04 |
| josch | bkeys: these days, reform-tools tries to be distro-agnostic (it doesn't have a ./debian directory anymore) so send all your CentOS issues with reform-tools to me :) | 16:04 |
| bkeys | Will do | 16:04 |
| bkeys | It's not my business but I think if you got edk2 on emmc and it worked that'd be a good option to have in the store | 16:04 |
| minute | i'll look into this again for the next. but it will be some more work on edk2 driver stuff | 16:05 |
| josch | so... this is a bit of a downer because my charging board is borked but... | 16:12 |
| josch | just now, a battery cell arrived which, from its advertised measurements, should fit just barely into the pocket bottom and has more than double advertised capacity | 16:13 |
| josch | https://mister-muffin.de/p/LFZU.jpg | 16:13 |
| josch | :D | 16:13 |
| josch | pocket-reform-sysctl-fw/inc/sysctl.h has BATTERY_CAPACITY_MILLIAMP_HOURS | 16:14 |
| josch | minute: obviously i can only test this once my new charging board arrives, so we can talk about this another time, but is there much more to hack in the firmware than above value to support bigger cells? | 16:14 |
| minute | josch: i don't think you have to change anything | 16:15 |
| josch | oh? | 16:16 |
| josch | well then :D | 16:16 |
| minute | josch: also, can you give me the source for that battery? i'd like to try that too | 16:16 |
| josch | i have two of these cells and they fit exactly -- i'm very excited to figure out how well these cells work or whether they explode after a few weeks of use XD | 16:16 |
| josch | minute: you find it when searching for 7565121 (its measurements) -- i procured it from amazon.de (alternative would be some aliexpress sellers): https://www.amazon.de/-/en/gp/product/B091YCRQXL/ | 16:17 |
| minute | josch: wow! please be super careful :D but they look like they have integrated protection? | 16:17 |
| minute | josch: nice thx | 16:17 |
| josch | yes, there isa protection board attached to it | 16:18 |
| josch | but in contrast to the eremit cells i do not trust them much, yes :D | 16:18 |
| josch | i was just very happy with how snuggly they fit into the case -- i'll have more results once my unit can charge things again ^^ | 16:18 |
| minute | josch: yeah the size is really fortunate | 16:19 |
| josch | i was thinking about 3d printing a custom bottom plate to fit thicker cells but through sheer luck (i really have no idea how to find these cells methodically) i stumbled across that one in my search a few weeks ago (i started searching right when i saw the pocket shipping notice :D) | 16:20 |
| [tj] | you want a thicker pocket? | 16:21 |
| josch | [tj]: i considered it because i failed to find cells which fit exactly for a long time | 16:22 |
| grimmware | minute: on the topic of sending 4 bytes and expecting 8 back, I've not actually tried the i2c proxying in anger yet but the likelihood is that if that restriction stays in place we're probably restricting ourselves to reading and writing a byte at a time because block reads and writes are going to take up more space than that. Personally I think that's an okay limitiation which should be fine for simple peripherals. | 16:24 |
| grimmware | so like, if you're reworking the proto anyway it's worth considering whether extending the length of that now would provide more flexibility for future use cases. It's probably not a problem worth speculatively solving, just thought I'd mention it. | 16:26 |
| grimmware | for reform mobo 3 I don't think this is a factor because you can just do i2c direct from the SoM | 16:26 |
| minute | grimmware: yes, good call. there's no reason not to have flexible reply sizes | 16:29 |
| Zaba | josch: a snug fit for a li-ion cell is not a good idea because its dimensions change slightly during charge/discharge cycles | 16:29 |
| minute | (and also write sizes) | 16:29 |
| grimmware | might be worth modifying the proto to send the command length up front after the magic byte? | 16:34 |
| minute | grimmware: yeah, both cmd and response should get a header with size | 16:34 |
| minute | and probably a crc at the end | 16:35 |
| grimmware | ++ | 16:35 |
| minute | josch: i guess i can also put this patch in reform-debian-packages somehow? https://source.mnt.re/reform/reform-tools/-/merge_requests/new/diffs?merge_request%5Bsource_branch%5D=gnome-polishing | 16:39 |
| minute | josch: well, it's probably not a blocker for shipping those classic reforms now | 16:40 |
| josch | Zaba: yes, i'll be very careful with this experiment. I can wiggle the cell by about 1mm along its length. I'd hope that it does not increase a mm in thickness and if it does, my cell is probably borked anyways XD | 16:48 |
| josch | minute: i think you need to press the button to turn your gnome-polishing branch into a MR but i see your top commit and will put it into reform-debian-packages now | 16:50 |
| - spew (QUIT: Remote host closed the connection) (~spew@135.233.119.40) | 16:59 | |
| minute | josch: thanks!! | 16:59 |
| josch | hrm... | 17:01 |
| abortretryfail | Anyone got any good ideas for troubleshooting a dead standby3v3 rail on a Pocket without voiding its warranty? My friend brought hers over to try and figure it out. | 17:01 |
| josch | i have to investigate why the CI fails to pull in reform-tools 1.73-1 from experimental | 17:01 |
| josch | oooooh | 17:03 |
| josch | we are using snapshot.d.o XD | 17:04 |
| josch | so of course my uploads to Debian have zero effect | 17:04 |
| josch | but i think we were only doing this to avoid linux 6.15 -- all other packages should not come from unstable if i am not missing anything | 17:04 |
| minute | abortretryfail: charger module failure? | 17:09 |
| minute | abortretryfail: try bypass, josch just posted a guide in community | 17:09 |
| abortretryfail | That was the unplug the batteries, jumper pins 2-3, and try looking for LEDs, right? | 17:10 |
| abortretryfail | I'll go look at the forums and see if that's what we did. She had an email from support we were working from. | 17:11 |
| abortretryfail | Oh yeah this looks like what we tried last night. https://community.mnt.re/t/running-the-pocket-without-batteries-bypassing-a-broken-charging-board/3440 | 17:12 |
| josch | minute: the new upload of linux-base makes more fixes necessary -- i'm working on https://source.mnt.re/reform/reform-debian-packages/-/merge_requests/114 right now and will have to check whether everything continues to work on a live system once this is done building before i can merge to main | 17:16 |
| minute | josch: ohh ok... | 17:25 |
| minute | scrolling in firefox is a bit sluggish on this new rk3588 test system. hmm. is it always like that? | 17:28 |
| minute | gnome help scrolls extremely slow | 17:31 |
| minute | the program is called "yelp" | 17:33 |
| minute | ok gtk3 apps are unusably slow | 17:33 |
| - digitalrane (QUIT: Ping timeout: 250 seconds) (~digitalra@user/digitalrane) | 17:34 | |
| minute | except if the help windows is maximized, then it's fast | 17:35 |
| minute | ah no it's just yelp that has this issue | 17:37 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 17:38 | |
| kfx | does the rk3588 still use an hdmi adapter, or did the drivers sprout edp support? | 17:40 |
| minute | kfx: with mb3.0 we can use dsi now | 17:42 |
| + digitalrane (~digitalra@user/digitalrane) | 17:43 | |
| minute | ok, to populate "gnome software" one has to run "apt update" once. it would be nice if some part of the system could say that | 17:45 |
| minute | altgr/umlauts are not working | 17:47 |
| minute | aha, "show extended input sources" needs to be enabled | 17:48 |
| minute | and "alternate characters key" set | 17:49 |
| abortretryfail | minute: I don't think this one is just the charger module failed. If I plug it in to a USB-C PD port, it will blink for a moment, then cut off power. | 18:04 |
| minute | abortretryfail: ok, sounds bad | 18:05 |
| minute | abortretryfail: which country are you in? | 18:05 |
| abortretryfail | Off an older 5V 15W USB-C phone charger, it powers up sysctl and the keyboard with the jumper between VCC and VCHG | 18:05 |
| minute | abortretryfail: huh ok? | 18:05 |
| abortretryfail | USA, Regarding #5739 for a friend. | 18:05 |
| minute | abortretryfail: sysctl fw revision? | 18:05 |
| abortretryfail | Uhhh, 1 sec | 18:05 |
| minute | abortretryfail: sounds like PD negotiation doesn't work, so 5v works | 18:06 |
| minute | abortretryfail: try to connect the motherboard to another computer with the internal usb-c and update sysctl firmware | 18:06 |
| minute | abortretryfail: curl -sfL mnt.re/update-pocket-sysctl | sudo bash - | 18:07 |
| abortretryfail | lol it says "error:syntax" for the version when I brought it in here | 18:07 |
| minute | abortretryfail: ok, sysctl not responding | 18:08 |
| minute | abortretryfail: maybe sysctl fw borked | 18:08 |
| minute | abortretryfail: prog switch off? | 18:08 |
| abortretryfail | It was before on teh other charger with a slightly higher 5v | 18:08 |
| abortretryfail | Yes, prog is off. | 18:08 |
| minute | abortretryfail: get 1 more 2mm jumper. restore internal usb jumpers but set to other position. connect to computer with usb-c. try to reflash | 18:08 |
| abortretryfail | Oddly this thing draws ~0.4W of power when off and she had issues with it self-discharging rapidly. | 18:08 |
| abortretryfail | Yeah I will try that and see if I can get it to program. Thanks :) | 18:09 |
| abortretryfail | I plugged it in to the charger that provides 5.35V and it correctly shows the FW version again. PREF1SYSR120240830normgL,0,0,0 | 18:14 |
| minute | abortretryfail: ok that's quite old, please update | 18:16 |
| minute | abortretryfail: also, in terms of batteries, tried un/replugging the batteries already? | 18:16 |
| abortretryfail | Yeah, we tried that weeks ago. One of the two cells is totally flat. | 18:16 |
| minute | abortretryfail: eek ok. it won't work with a flat cell | 18:17 |
| abortretryfail | It's still suspicious to me that this thing is using 0.07A off the 5V like this for only standby power | 18:17 |
| minute | abortretryfail: did you remove the cells already?> | 18:17 |
| minute | abortretryfail: where do you see the 0.07A? | 18:17 |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 18:18 | |
| abortretryfail | Right now the charger is totally disconnected. I am seeing 0.07A with a USB-C analyzer in between the power supply and the Pocket | 18:18 |
| minute | abortretryfail: what if you disconnect the keyboard? | 18:19 |
| minute | abortretryfail: what if you disconnect the cpu? anything else plugged in? | 18:19 |
| abortretryfail | I can pop the CPU module out, but iirc that didn't even have a connection to the standby3v3 rail | 18:20 |
| abortretryfail | CPU module out it's still around 0.06-0.07A | 18:21 |
| minute | abortretryfail: keyboard? | 18:21 |
| abortretryfail | I can tell it to power on and the blue LED turns on and current draw comes up to about 0.12A, but it never boots. | 18:22 |
| abortretryfail | Well, yeah the keyboard is connected. I can try removing that. | 18:22 |
| minute | abortretryfail: please do, otherwise you don't know where the current draw is from | 18:22 |
| abortretryfail | Both keyboard plugs out it's around 0.05A. I can't attest to the accuracy of this meter for small currents either. | 18:23 |
| abortretryfail | That still seems like too much for just a 3v3 regulator and a RP2040 | 18:23 |
| abortretryfail | I will try to find a jumper and do a firmware update and see if that changes anything. | 18:23 |
| - elb (QUIT: Remote host closed the connection) (~elb@2600:4041:6671:1300:e8f6:ebea:a95c:736) | 18:23 | |
| + elb (~elb@2600:4041:6671:1300:4e33:9196:e71c:f78e) | 18:25 | |
| minute | abortretryfail: there could be a short in the display cable also. what if you unplug the display cable from the mb? | 18:28 |
| - chomwitt (QUIT: Ping timeout: 265 seconds) (~chomwitt@2a02:85f:9a00:8300:1ac0:4dff:fedb:a3f1) | 18:58 | |
| josch | (upgrading to the branch snapshot-linux-only from reform-debian-packages worked -- now creating a system image of that with the "gnome" branch) | 19:22 |
| - mjw (QUIT: Ping timeout: 272 seconds) (~mjw@2001:1c00:2a07:3a01:e4e2:af5b:bbd7:5c24) | 19:26 | |
| josch | minute: i'm investigating why "dpkg-reconfigure gdm3" is necessary but because you have this workaround and because running "dpkg-reconfigure gdm3" is completely harmless, I suppose that fixing this "properly" is not very high on your list of priorities -- what thing *is* high on your priorities and somehow needs my input? I'll probably not be online tomorrow... | 19:43 |
| kfx | minute: should I wait for mb 3.0 to be available before I upgrade to rk3588 then? | 19:48 |
| + mjw (~mjw@178.230.17.0) | 19:54 | |
| abortretryfail | minute: I hadn't considered the display cable. I'll try that now (Sorry for the delay, I had to set this aside and attend work meetings) | 20:02 |
| + chomwitt (~chomwitt@2a02:85f:9a00:8300:1ac0:4dff:fedb:a3f1) | 20:04 | |
| abortretryfail | Hmm, when I remove the jumper between VCC and VCHG on pins 2-3 of the charger jack, current drops to 0. Maybe I can put a more accurate ammeter on there. | 20:05 |
| abortretryfail | I'm going to try and update the firmware first, this is difficult to do without cables that fit. | 20:09 |
| + josch-pocket (~josch-poc@195.12.232.234) | 20:16 | |
| josch-pocket | hello from pocket reform online via the quectel em06-e | 20:16 |
| josch-pocket | no ethertnet, no wifi so it must really be working or otherwise i could not compose these lines :) | 20:17 |
| josch-pocket | i also got a gnss antenna, so lets see whether i get geolocation as well (but not toda) | 20:18 |
| abortretryfail | The pocket has built-in GNSS? | 20:19 |
| abortretryfail | Oh, via the modem card | 20:19 |
| josch-pocket | the top plate ist off because the antenna is hanging out so offline again :) | 20:19 |
| - josch-pocket (QUIT: Client Quit) (~josch-poc@195.12.232.234) | 20:19 | |
| josch | abortretryfail: yes | 20:20 |
| abortretryfail | I am glad I did not just blindly run mnt.re/update-pocket-sysctl as root | 20:20 |
| abortretryfail | This would not work at all on my Archlinux desktop lol | 20:21 |
| minute | funfact: sleeping for 1ms in the lpc driver is not much worse than sleeping for 50ms | 20:28 |
| minute | but one has to sleep | 20:28 |
| minute | josch: while i'm doing some gsettings polishing, is it ok to just use the reform-tools branch setting in gitlab ci over the next days? because it doesn't make sense to cut releases for tiny adjustments there while there may be more coming | 20:30 |
| josch | minute: for building system images that you use to test or for building system images that you plan to distribute? | 20:32 |
| abortretryfail | minute: updating to 20250512 fixed the PD negotiation. Now it's coming up at 20V and not resetting. | 20:47 |
| abortretryfail | I'm going to reassemble this thing and see if it powers on now. :D | 20:47 |
| - jackhill_ (QUIT: Remote host closed the connection) (~jackhill@kalessin.dragonsnail.net) | 20:51 | |
| minute | josch: to install on a machine that is shipped | 20:53 |
| minute | abortretryfail: great | 20:53 |
| josch | minute: then you have to consider that use of a custom reform-tools branch will overwrite files from the installed "reform-tools" package with files from the downloaded tarball. This means that dpkg-verify will complain that files belonging to reform-tools contain incorrect content. It also means that the next time the user updates reform-tools, they will overwrite the custom files you put (which may or | 20:57 |
| josch | may not be what you want) | 20:57 |
| minute | josch: that's fine except for the dpkg-verify (will that show up on the next update of the package or when?) because it's only about gschema updates which are only relevant when creating a new user | 20:59 |
| josch | is there a better way? you could again put the settings in by calling "gsettings" in scripts like reform-setup.sh, you could also maintain your patch stack as a patch in reform-debian-packages instead of pulling in reform-tools | 20:59 |
| minute | josch: i think for now it's ok if these can somehow be added https://source.mnt.re/reform/reform-tools/-/merge_requests/124/diffs?commit_id=1c68cfef00afa82c05805956b1f33c0b0ef85b41 | 21:00 |
| josch | going via reform-debian-packages would mean that you drop your patches to the schema here: https://source.mnt.re/reform/reform-debian-packages/-/tree/main/reform-tools/debian/patches?ref_type=heads | 21:00 |
| josch | minute: the first one is already in as you can see above | 21:00 |
| minute | josch: yes, i need the second one also. the easiest for me would be if i could just replace the file instead of patching (because for patching i need to find the file to patch against, which involves nontrivial searching... well, i guess i need to write down the command to extract the source file from debian experimental (?) somewhere) | 21:02 |
| minute | josch: i guess it's late and i'm running out of steam a bit ^^ | 21:02 |
| josch | minute: i think pulling in unreleased reform-tools from git is okay *if* you do not add any files which will never get "corrected" by an actual upload of reform-tools. Should that happen, you have dead files on the user's system that belong to no package at all | 21:03 |
| minute | josch: yeah, i won't need to do that ^^ | 21:04 |
| josch | minute: the version in experimental is the latest git tag, so i just run this to generate the patches: | 21:04 |
| josch | git format-patch -2 | 21:04 |
| minute | josch: ahhh i see, thx | 21:04 |
| josch | here now: https://source.mnt.re/reform/reform-debian-packages/-/tree/main/reform-tools/debian/patches | 21:04 |
| josch | minute: is there something on my todo list for tonight? i'll not be online tomorrow | 21:05 |
| josch | hrm... or maybe i will... the thing has lte now :D | 21:05 |
| josch | (the thing => the pocket) | 21:05 |
| minute | josch: oh, thanks for the patch! no, nothing else right now, i think everything else is fine | 21:06 |
| minute | josch: then i can switch back to main and actually do a sysimage release i think | 21:07 |
| josch | \o/ | 21:07 |
| + mark_ (~mjw@178.230.17.0) | 21:12 | |
| - mjw (QUIT: Read error: Connection reset by peer) (~mjw@178.230.17.0) | 21:13 | |
| * mark_ -> mjw | 21:13 | |
| - mjw (QUIT: Ping timeout: 268 seconds) (~mjw@178.230.17.0) | 21:30 | |
| + mjw (~mjw@gnu.wildebeest.org) | 22:14 | |
| - gustav28 (QUIT: Quit: Quit) (~gustav@c-78-82-54-59.bbcust.telenor.se) | 22:15 | |
| - chomwitt (QUIT: Ping timeout: 265 seconds) (~chomwitt@2a02:85f:9a00:8300:1ac0:4dff:fedb:a3f1) | 22:44 | |
| + gustav28 (~gustav@c-78-82-54-59.bbcust.telenor.se) | 22:48 | |
| - gustav28 (QUIT: Client Quit) (~gustav@c-78-82-54-59.bbcust.telenor.se) | 22:52 | |
| + gustav28 (~gustav@c-78-82-54-59.bbcust.telenor.se) | 22:54 | |
| minute | at least on this machine i've achieved very stable spi communication now, but it was a grind (no checksums etc involved yet) | 23:49 |
| minute | also, at much faster speed than before | 23:49 |
| minute | 0 transfer errors in a 1000x test loop | 23:50 |
| - Ar|stote|is (QUIT: Ping timeout: 252 seconds) (~linx@149.210.74.47) | 23:51 | |
| + Ar|stote|is (~linx@149.210.73.109) | 23:56 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
