| abortretryfail | How's the burn-in happen? folks turn on their Reform and just leave the keyboard alone with the logo showing? | 00:08 |
|---|---|---|
| sigrid | basically | 00:09 |
| sigrid | I think one way to fix it is to have an inverted image displayed for some time | 00:10 |
| sigrid | (I might be partially joking) | 00:10 |
| vkoskiv | There are also some cases, for instance when you plug in a unit with low battery, that the keyboard controller may reset and present the logo. | 00:11 |
| abortretryfail | ACTION goes to take a peek at the keeb fw | 00:12 |
| jfred | Hmm. My Reform today has been having trouble staying associated to wifi APs. I did just apt-get upgrade recently so I wonder if there was a regression somewhere. Just cracked it open to see if one of the antenna cables had come loose but they seem fine | 00:21 |
| jfred | oh there are new kernel packages available | 00:25 |
| jfred | ACTION tries another upgrade | 00:25 |
| - mjw (QUIT: Ping timeout: 255 seconds) (~mjw@gnu.wildebeest.org) | 00:29 | |
| minute | jfred: someone mentioned a similar issue on community. i wonder if something is wrong with the kernel commandline. can you check if pci=nomsi is still there? | 00:32 |
| minute | in /proc/cmdline for example | 00:32 |
| jfred | oh, nope, it's not | 00:33 |
| jfred | jfred@reform:~$ cat /proc/cmdline | 00:33 |
| jfred | console=tty1 ro no_console_suspend cma=512M@3G cryptomgr.notests | 00:33 |
| minute | damn | 00:34 |
| minute | cc josch | 00:34 |
| minute | jfred: that's the issue then. you have graphical uboot, right? can you stop it (press space) and do: setenv bootargs pci=nomsi | 00:35 |
| minute | and then: boot | 00:35 |
| jfred | I do, will give that a try - one moment | 00:35 |
| jfred | minute: interesting, I set that boot arg from within uboot but it didn't end up in the kernel cmdline | 00:39 |
| minute | hmmmmmm | 00:39 |
| minute | can you pastebin your /boot/boot.scr? | 00:39 |
| jfred | https://bpa.st/3WYA | 00:41 |
| minute | jfred: also, do you have the file: /usr/share/flash-kernel/ubootenv.d/00reform2_ubootenv | 00:43 |
| jfred | yup, seems to exist | 00:44 |
| jfred | -rw-r--r-- 1 root root 546 Sep 28 09:03 /usr/share/flash-kernel/ubootenv.d/00reform2_ubootenv | 00:44 |
| minute | jfred: ok, thank you! can you check one more thing in u-boot's console for me? the output of: print board | 00:45 |
| minute | perhaps it's not nitrogen8m_som | 00:46 |
| jfred | Oh yep, it's listed as "MNT Reform 2.0" | 00:47 |
| minute | oh wow ok | 00:48 |
| jfred | (Also thanks, this has given me more of a feel for where things are in ARM-land haha - more used to x86 and GRUB) | 00:50 |
| minute | ok, so this might be a fix: | 00:50 |
| minute | jfred: can you replace the contents of the above mentioned 00reform2_ubootenv with this: https://source.mnt.re/reform/reform-tools/-/raw/main/flash-kernel/ubootenv.d/00reform2_ubootenv?ref_type=heads | 00:51 |
| minute | i added another if case for your board value | 00:51 |
| jfred | Out grabbing food real quick but I can try that when I get back to my Reform | 00:52 |
| minute | jfred: after that you only have to run `sudo flash-kernel` | 00:52 |
| minute | i think | 00:52 |
| minute | sure! | 00:53 |
| vkoskiv | I guess I ended up learning DBus, then. My program is now doing the right thing, listening for signals instead of polling properties like I initially wrote it | 00:54 |
| vkoskiv | I'm looking at periodic battery state updates from UPower 8) | 00:55 |
| vkoskiv | All sorts of goodies in here | 00:55 |
| minute | vkoskiv: cool! | 00:56 |
| jfred | minute: yup that did it, thanks! | 01:18 |
| - chomwitt (QUIT: Ping timeout: 255 seconds) (~chomwitt@2a02:587:7a24:b000:1ac0:4dff:fedb:a3f1) | 01:20 | |
| minute | jfred: cool, thanks for testing! | 01:46 |
| jfred | no prob :) | 01:49 |
| - mtm (QUIT: Ping timeout: 260 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 02:04 | |
| vkoskiv | I got deep into this apparently. Wrote an entire parser + callback dispatcher for dbus signals | 02:08 |
| vkoskiv | It now has a very satisfying API. | 02:08 |
| vkoskiv | Hook the system bus, hook a desired signal with a provided callback, done. | 02:08 |
| vkoskiv | I'm sure glib bindings would do something similar, but then I'd have to link in a bunch of libraries | 02:09 |
| vkoskiv | And I'd have to write less code | 02:09 |
| - jacobk (QUIT: Ping timeout: 260 seconds) (~quassel@utdpat241106.utdallas.edu) | 02:31 | |
| - nsc (QUIT: Ping timeout: 255 seconds) (~nicolas@250-97-142-46.pool.kielnet.net) | 03:10 | |
| + nsc (~nicolas@67-98-142-46.pool.kielnet.net) | 03:12 | |
| josch | "MNT Research Shop Shipment Notification" \o/ | 03:20 |
| josch | minute: your u-boot must be very old :D you changed the value for "board" in u-boot env from nitrogen8m_som to "MNT Reform 2.0" in May 2020: 7d4327e9938f6c3f3d016bbf0a89296a0536053c | 03:24 |
| josch | is anybody out there currently using such an old u-boot? | 03:25 |
| josch | also, thet "setenv bootargs" at the very top of 00reform2_ubootenv, should that not also have a ${bootargs} in it? | 03:25 |
| josch | you suggested jfred to set ${bootargs} in the interactive u-boot console but whatever is set there is currently ignored by our boot.scr as it is | 03:26 |
| josch | that should not happen, right? | 03:26 |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 04:09 | |
| - murphnj (QUIT: Quit: Leaving) (~murph@user/murphnj) | 04:58 | |
| - bkeys (QUIT: Ping timeout: 248 seconds) (~Thunderbi@199.226-24.cm.ptn.tftn.dynamic.friendlycity.net) | 05:05 | |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 07:22 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 08:35 | |
| + jacobk (~quassel@64.189.201.150) | 08:42 | |
| + chomwitt (~chomwitt@2a02:587:7a24:b000:1ac0:4dff:fedb:a3f1) | 09:17 | |
| - Gooberpatrol66 (QUIT: Remote host closed the connection) (~Gooberpat@user/gooberpatrol66) | 09:18 | |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 09:18 | |
| - chomwitt (QUIT: Ping timeout: 260 seconds) (~chomwitt@2a02:587:7a24:b000:1ac0:4dff:fedb:a3f1) | 09:22 | |
| minute | josch: correct. but also, if you look at jfreds boot.scr paste, you will see that there is a default old bootargs from somewhere, and then you'll have a clash with a bunch of redundant args | 11:32 |
| minute | josch: i guess i didn't understand how "board" works and can't look up that commit right now because i'm on phone | 11:33 |
| minute | josch: also i hadn't looked into the deb changelog commands yet (and was too tired last night) so i didn't make another packages build, but we should make one swiftly | 11:35 |
| minute | for reform-tools i mean | 11:35 |
| josch | minute: i can take care of reform-tools and triggering the reform-debian-packages pipeline | 11:35 |
| josch | minute: the "default old bootargs from somewhere" come from /etc/default/flash-kernel | 11:36 |
| josch | minute: we are not creating that file anymore with sysimage-v4 but existing installations will still have it | 11:36 |
| josch | since it lives in /etc it's not so easy to replace it with an upgrade of reform-tools | 11:37 |
| josch | maybe i should just add another test in reform-check that if reform-tools is new enough (i.e it includes the new 00reform2_ubootenv) then /etc/default/flash-kernel should be deleted | 11:37 |
| minute | josch: yea | 11:37 |
| minute | yes | 11:37 |
| minute | so lets override the whole bootargs as it is, it's more deterministic | 11:38 |
| josch | okay | 11:38 |
| josch | shipping better default bootargs for each board in u-boot itself will still be useful though for users of other distros than Debian | 11:38 |
| minute | and one can edit the ubootenv script to include a custom variable if needed | 11:39 |
| minute | that is true | 11:39 |
| josch | so i think there is some utility in making sure that each of the u-boot variants ships good values (even if they get overridden by boot.scr via flash-kernel) | 11:39 |
| minute | yes, agreed | 11:39 |
| josch | nice :) | 11:39 |
| minute | can you perhaps create a quick issue ticket for it in reform-tools? then it's not forgotten | 11:40 |
| josch | to add another test in reform check deleting /etc/default/flash-kernel? sure, will do | 11:41 |
| josch | at least it's not harmful to have it in there | 11:41 |
| minute | ah i meant to make sure that all uboots set good default bootargs also | 11:43 |
| josch | i think they currently do | 11:44 |
| josch | issue for reform-check so that i don't forget: https://source.mnt.re/reform/reform-tools/-/issues/3 | 11:45 |
| josch | i think our uboots already set the correct default cmdline values. you recently approved merge requests of mine that did this | 11:51 |
| josch | but let me check again -- i think there were some changes to 00reform2_ubootenv that are not reflected in some u-boot yet | 11:51 |
| josch | okay, ls1028a uboot was the only one were bootargs were not the same as the ones set by flash-kernel. MR here: https://source.mnt.re/reform/reform-ls1028a-uboot/-/merge_requests/2 | 11:56 |
| josch | okay, i merged https://source.mnt.re/reform/reform-tools/-/merge_requests/56 and will now trigger reform-debian-packages | 11:59 |
| minute | josch: great, thanks for checking this so quickly! | 12:00 |
| - klardotsh (QUIT: Ping timeout: 255 seconds) (~klardotsh@c-67-170-115-80.hsd1.wa.comcast.net) | 12:00 | |
| minute | i'm also now on a computer (it's by MNT) | 12:00 |
| + eery (~eery@77.137.73.95) | 12:10 | |
| + chomwitt (~chomwitt@2a02:587:7a24:b000:1ac0:4dff:fedb:a3f1) | 12:56 | |
| + mjw (~mjw@gnu.wildebeest.org) | 13:23 | |
| violet | MNT, ive heard of them | 13:31 |
| - Boostisbetter (QUIT: Ping timeout: 240 seconds) (4a410829d7@irc.cheogram.com) | 13:32 | |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 13:39 | |
| - mtm (QUIT: Ping timeout: 255 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 14:03 | |
| Boostisbetter | minute: Reform, Pocket, thin, or Rack? | 14:29 |
| minute | Boostisbetter: reform this time! | 14:54 |
| - jacobk (QUIT: Ping timeout: 255 seconds) (~quassel@64.189.201.150) | 15:02 | |
| Boostisbetter | minute: mosts excellent, me too. | 15:19 |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 16:09 | |
| vkoskiv | Oof, dbus introspection returns... XML. Now I'm going to have to link in an XML parser :( | 16:13 |
| vkoskiv | I was planning on just having a json parser, but it seems that binary bloat will be unavoidable if I want DBus support | 16:14 |
| minute | libxml is very common i think | 16:16 |
| vkoskiv | Yeah, I'm trying to stick to libraries that ship by default with Sysv4, and that is included | 16:16 |
| vkoskiv | So not too bad. | 16:16 |
| vkoskiv | I don't know what it is about xml, but any exposure to it makes computing slightly less fun for me :D | 16:17 |
| vkoskiv | Why can't the pkg-config name for a package just be... The package name | 16:21 |
| vkoskiv | In this case package is 'libxml2', but pkg-config wants you to guess 'libxml-2.0' | 16:22 |
| vkoskiv | But for dbus the package is 'libdbus-1-dev, and pkg-config wants just 'dbus-1' | 16:24 |
| vkoskiv | Am I supposed to just know that, or am I doing something wrong? | 16:24 |
| vkoskiv | I always just have to hammer in a bunch of guesses to find the right permutation for a given package | 16:25 |
| vkoskiv | I guess I could write some kind of script to just try everything, and find the name for me? | 16:25 |
| vkoskiv | (first example, package name is probably also 'libxml2-dev') | 16:26 |
| + nocko (~nock@user/nocko) | 16:26 | |
| vkoskiv | I guess I can tab-complete under /usr/lib/aarch64-linux-gnu/pkgconfig/, but couldn't pkg-config just go "did you mean <correct_name>"? :D | 16:28 |
| vkoskiv | Or better yet, have the same name for the .pc files as the name of the dev package | 16:28 |
| vkoskiv | Anyway, rant over, if I'm having trouble adding two packages to my tiny makefile, I can't imagine how much of a headache package maintainers have :D | 16:35 |
| minute | vkoskiv: yeah what i do is a combo of apt-file search and grep for the .pc | 16:36 |
| vkoskiv | Feels like a gap in tooling that everyone just scripts their way around, no? | 16:37 |
| vkoskiv | This is the kind of stuff that pushes developers towards single file libraries, and towards static linking as a result. | 16:38 |
| vkoskiv | I've done the same thing, I had no idea how to do any of this ~8 years ago when I was working on my first C projects | 16:39 |
| vkoskiv | Finding a .h, dropping it in, that worked. | 16:39 |
| vkoskiv | Far too often I also do `apt search <thing> | grep <thing>` | 16:40 |
| vkoskiv | I now know of --names-only, but I still get a bagillion results that I have to manually scan to find the one thing I was looking for | 16:40 |
| minute | i agree | 16:43 |
| - stites (QUIT: Ping timeout: 260 seconds) (~stites@130.44.147.204) | 16:43 | |
| + stites (~stites@2607:fb90:ad62:a207:f0a1:b4c8:6d1c:d465) | 16:44 | |
| - ethulhu (QUIT: Excess Flood) (ethulhu@nora.ethulhu.co.uk) | 16:49 | |
| josch | vkoskiv: do you think there is something i can help you with? i do not yet understand exactly what the confusion is | 17:02 |
| vkoskiv | Oh, I don't think there is an issue per se, just wondering if there would be ways to improve the discoverability of development packages somehow | 17:06 |
| josch | vkoskiv: you mean you need header foo.h and you want to find out what package you have to install on Debian to get it? | 17:07 |
| vkoskiv | Yeah, currently doing that for cJSON.h. I see that the .so is installed, but there is no .pc file, and I can't find a header | 17:08 |
| vkoskiv | 2015 me just copied the .c and .h files to my repo, but for this project, I'd like to do it the proper way for all libs. | 17:09 |
| josch | for shared libraries, the package name is *always* (except for zlib because of historical reasons) derived from the SONAME | 17:09 |
| josch | so your case you have libcjson.so.1 and that must be in the package called libcjson1 | 17:10 |
| vkoskiv | Aha, I'm missing libcjson-dev | 17:10 |
| josch | you could then run: apt-cache showsrc libcjson1 | 17:11 |
| josch | that would show you from which source package the binary package would be built and then in the Binary: field you would see that the source package also builds a binary package called libcjson-dev | 17:11 |
| josch | and that is the one you were looking for | 17:11 |
| vkoskiv | E: You must put some 'deb-src' URIs in your sources.list | 17:11 |
| vkoskiv | When trying that showsrc command | 17:11 |
| josch | yes, you have to add that | 17:11 |
| josch | do you know how? | 17:11 |
| vkoskiv | I'll look it up | 17:12 |
| josch | without a deb-src line in your sources.list, finding and installing development headers is not much fun :) | 17:12 |
| josch | the second way to find out that you need libcjson-dev is by using: | 17:12 |
| josch | apt-file search /cJSON.h | 17:13 |
| josch | and that would also list you libcjson-dev | 17:13 |
| vkoskiv | Ooh, I think that's what I was missing | 17:13 |
| josch | in general (but not always, so i always run apt-file search anyways) development header binary packages are named after their library names | 17:13 |
| vkoskiv | Ah, I don't have apt-file installed | 17:14 |
| vkoskiv | 'apt search apt-file' feels very meta :D | 17:14 |
| josch | so you need xml2, so install libxml2-dev. you need dbus-1 (according to pkgconfig), so install libdbus-1-dev, | 17:14 |
| josch | and libcjson-dev if you need cjson | 17:14 |
| josch | for *-dev packages, naming is not 100% uniform compared to shared libraries, so i always run apt-file instead | 17:15 |
| vkoskiv | same repo url for deb-src in sources.list, yeah? | 17:15 |
| josch | essentially, you copy the "deb *" line of your main repo and do s/^deb /deb-src / | 17:15 |
| vkoskiv | Cool, then apt-file update? | 17:16 |
| vkoskiv | It seems to work now, not sure if this warning I got is relevant: W: Skipping acquire of configured file 'main/source/Sources' as repository 'https://mntre.com/reform-debian-repo reform InRelease' does not seem to provide it (sources.list entry misspelt?) | 17:17 |
| vkoskiv | apt-file search works now, very useful | 17:18 |
| josch | vkoskiv: yes, you do not need a deb-src entry for https://mntre.com/reform-debian-repo because there are indeed no source packages in there | 17:18 |
| josch | you want a deb-src entry for deb.debian.org | 17:18 |
| vkoskiv | apt-cache showsrc still says the same error, though. | 17:19 |
| josch | vkoskiv: knowing about apt-file was the main puzzle piece you were missing. Now you can just search for a path or parts of in and find the package that provides it. :) | 17:19 |
| vkoskiv | Oh, I just copied the existing deb line there and made it deb-src | 17:19 |
| vkoskiv | josch: indeed, thanks a bunch! | 17:19 |
| josch | vkoskiv: after changing your sources, did you run "apt-get update"? | 17:19 |
| vkoskiv | Yeah. apt-cache still isn't happy, though. | 17:20 |
| vkoskiv | Confused, so should the deb-src line point to deb.debian.org or mntre.com, or both? | 17:20 |
| josch | what's the error? can you show your sources.list? | 17:20 |
| josch | only deb.debian.org | 17:20 |
| josch | it will not have any effect pointing at mntre.com | 17:20 |
| josch | (because no source packages are shipped in that repo) | 17:21 |
| vkoskiv | Aha, 'deb-src [trusted=yes] https://deb.debian.org/' ? | 17:21 |
| josch | leave out the [trusted=yes] | 17:21 |
| josch | because in contrast to mntre.com, deb.debian.org is gpg signed | 17:21 |
| josch | so no need to disable signature verification | 17:21 |
| + RPPeakt (~ERueanion@117.104.246.210) | 17:21 | |
| RPPeakt | https://supernets.social damn elon musk sucks dick | 17:22 |
| RPPeakt | RPPeakt stites nocko mtm Boostisbetter mjw chomwitt eery Gooberpatrol66 nsc sigrid Ar|stote|is colinsane doppler XYZ bleb rah marty_mcfly88 mark_ eschaton S0rin GNUmoon robin dozens henesy sknebel V swivel mhoye blast007 josch lexik minute leonardo ec0 truck laumann hl pinoaffe flowy deflated8837 plomlompom _nrb_ reformer sbp inhji jnerula kremlin XgF ndufresne5 skalk tarxvf brennen sundog2062 | 17:22 |
| - RPPeakt (PART: TRUMP WON IN 2016) (~ERueanion@117.104.246.210) | 17:22 | |
| vkoskiv | More errors, 404s with this config. | 17:22 |
| dozens | why | 17:22 |
| josch | i can show you my sources.list if you like | 17:22 |
| vkoskiv | Ah, I had reform on the debian.org line | 17:22 |
| vkoskiv | Probably want 'unstable main' instead | 17:22 |
| josch | yes | 17:22 |
| vkoskiv | Still 404s in apt-get update | 17:23 |
| vkoskiv | Err:6 https://deb.debian.org unstable Release | 17:23 |
| vkoskiv | 404 Not Found [IP: 151.101.246.132 443] | 17:23 |
| vkoskiv | Then at the end it says that one doesn't have a release file | 17:24 |
| vkoskiv | Which makes sense, I guess | 17:24 |
| vkoskiv | deb http://deb.debian.org/debian unstable main | 17:24 |
| vkoskiv | deb-src https://deb.debian.org/ unstable main | 17:24 |
| vkoskiv | deb [trusted=yes] https://mntre.com/reform-debian-repo reform main | 17:24 |
| vkoskiv | Oh, where thit that /debian come from | 17:25 |
| vkoskiv | It was there before, I'll add it for the deb-src too | 17:25 |
| vkoskiv | *now* it works :D | 17:25 |
| vkoskiv | apt-cache too works | 17:25 |
| vkoskiv | I chose cJSON because I'm familiar with the API, and the library ships by default in SysV4 | 17:27 |
| + ethulhu (ethulhu@nora.ethulhu.co.uk) | 17:29 | |
| josch | vkoskiv: yes, adding /debian was the missing part -- the rest looks correct. Good luck! :) | 17:31 |
| josch | oh one thing | 17:31 |
| josch | you don't need https | 17:31 |
| josch | since the repo is gpg signed, https does not add very much | 17:31 |
| josch | a case for https in Debian mirrors could be made if one argues that by snooping http traffic, an attacker could find out which security updates a Debian system has applied and which ones it is still vulnurable to -- but you are using unstable :) | 17:32 |
| vkoskiv | Yeah, encrypting that traffic for almost no reason does add overhead | 17:39 |
| - xktr (QUIT: Quit: leaving) (~xktr@2602:fe3d:c01:10ca:1050:1ace:0:b) | 17:54 | |
| + xktr (~xktr@2602:fe3d:c01:10ca:1050:1ace:0:b) | 17:57 | |
| vkoskiv | C designated initializers, my beloved | 18:05 |
| vkoskiv | https://pastebin.com/esYrnyQ6 | 18:09 |
| vkoskiv | Quick and dirty inline bitmaps, with nearly live editing as I work on things | 18:09 |
| vkoskiv | I have entr rebuilding and restarting the project whenever I :w, it takes less than a second to come back up | 18:09 |
| vkoskiv | I can stick these in named variables later for reuse, but for prototyping, this is excellent. | 18:10 |
| - Boostisbetter (QUIT: Ping timeout: 260 seconds) (4a410829d7@irc.cheogram.com) | 18:52 | |
| + Boostisbetter (571d69e5d6@irc.cheogram.com) | 18:58 | |
| - Boostisbetter (QUIT: Ping timeout: 260 seconds) (571d69e5d6@irc.cheogram.com) | 19:33 | |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 19:34 | |
| - stites (QUIT: Read error: Connection reset by peer) (~stites@2607:fb90:ad62:a207:f0a1:b4c8:6d1c:d465) | 19:53 | |
| + stites (~stites@130.44.147.204) | 19:53 | |
| - stites (QUIT: Read error: Connection reset by peer) (~stites@130.44.147.204) | 19:55 | |
| + stites (~stites@130.44.147.204) | 19:56 | |
| - XYZ (QUIT: Remote host closed the connection) (~XYZ@37-48-8-141.nat.epc.tmcz.cz) | 20:50 | |
| + XYZ (~XYZ@37-48-8-141.nat.epc.tmcz.cz) | 20:54 | |
| - XYZ (QUIT: Read error: Connection timed out) (~XYZ@37-48-8-141.nat.epc.tmcz.cz) | 21:13 | |
| + XYZ (~XYZ@37-48-8-141.nat.epc.tmcz.cz) | 21:31 | |
| + klardotsh (~klardotsh@c-67-170-115-80.hsd1.wa.comcast.net) | 21:58 | |
| + jacobk (~quassel@utdpat241106.utdallas.edu) | 22:22 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 23:16 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!