2025-11-04.log

- plomlompom (QUIT: Remote host closed the connection) (~plom@139.59.138.73)00:01
+ plomlompom (~plom@139.59.138.73)00:02
minutei have a very unorthodox workaround for the waybar icon issue... but it could be much more unorthodox00:05
minutejosch: i just pushed a new tool00:29
minutejosch: so thanks for updating reform-tools/changelog, but there's now reform-waybar-icon-wedge as well00:30
minutejosch: which i'm testing on target now00:30
joschthen we put that into the next release00:31
+ TechnoWizard (~TechnoWiz@user/TechnoWizard)00:50
minutejosch: how do you like my latest horror https://source.mnt.re/reform/reform-tools/-/blob/main/bin/reform-waybar-icon-wedge?ref_type=heads00:51
- TechnoWizard (QUIT: Remote host closed the connection) (~TechnoWiz@user/TechnoWizard)00:55
+ TechnoWizard (~TechnoWiz@user/TechnoWizard)00:56
- TechnoWizard (QUIT: Remote host closed the connection) (~TechnoWiz@user/TechnoWizard)00:56
elbI just updated my pocket reform, and it acts like I'm sitting on the spacebar when it reboots00:59
elb(and I can't decrypt the disk, possibly for this reason)00:59
elbdoes that sound familiar to anyone00:59
minuteelb: hmmm, try oled menu -> reset keyboard01:00
elbdid nothing, neither did power cycling01:00
elbI'm opening it up to unplug the keyboard USB right now01:00
elbbut like ... they keyboard worked fine before rebooting01:00
minuteelb: huh. by power cycling do you mean standby switch toggle?01:01
elband the oled menu works normally and fine01:01
elboh no I mean hyper+enter 0 then 101:01
minuteelb: try standby switch toggle, then you don't need to unplug01:01
minuteelb: keyboard will be powered off by doing that01:01
elboh I meant to make sure the keyboard is even sending anything at all01:01
elbbut yeah let me try that, good call01:01
minuteah gotcha01:02
minutelike, switch or gpio failure01:02
elbstandby switch did nothing01:02
minutehmm, mechanical problem with one of the space keys?01:03
elb ok, the keyboard is unplugged01:04
elband it's still doing it01:04
elbso it's teh kernel or something01:04
minuteanything else connected? HDMI?01:04
elbnope01:04
elbnot even usb pd01:05
+ TechnoWizard (~TechnoWiz@user/TechnoWizard)01:07
elbgoing to put a system image on microsd01:08
minuteelb: i recently added gpio-keys for the wakeup gpio, with linux,code = <57>01:08
minuteelb: i wonder if it's that01:08
elbI think last I updated was about a week ago01:09
elbthis addition was in the linux kernel?01:09
minuteelb: it's in the devicetree, so in the kernel package01:09
minuteelb: merged oct 2401:10
elbhmmm not surew hat my previous kernel was01:11
elbI'll find out as soon as I get this image written01:11
minutei'm gonna comment this out of the device tree in any case, because if the gpio is not initialized by system controller, i'm not sure about the initial state01:12
minutepushed (but will take a while until that's rolled out)01:14
elbit will apparently take a while until this image writes, too ;-)01:17
minutebtw the gnome session is kind of broken (i think due to gnome-settings-daemon or similar) after logging out of a sway session 01:18
elbI normally log in at boot and then shut down when I'm done, so that feels survivable for me, personally01:19
minuteyeah... i don't have the energy to look into fixing this now... but i think the way that the sway session is started needs to be improved at some point01:19
elbI find that a lot of things about sway need to be improved; -)01:19
elbbut I'm not sure that's a reform  problem01:19
minuteyeah...01:20
elbhmmm this system image is 11 GB uncompressed but compresses to 201:21
elbas someone who lives always on the edge of running out of storage for reasons of personal shortcomings, that's a lot of zeroes (or whatever) to store ;-)01:21
elbI'm writing by piping gzip -dc to dd01:22
minuteelb: the best way to write the image is:01:22
minuteelb: bmaptool copy https://image/url /dev/sdX01:22
schalken1i'm having the same problem as elb, oof01:22
elbbmaptool is new to me01:22
minuteelb: it will then do sparse copy and decompression with http range requests01:22
minuteschalken1: same solution applies!01:23
elbinteresting01:23
elbwell for now I'm using a pipe01:23
elbI'll look at this for next time01:23
minutei.e. you don't need any temp storage, it will be streamed from the web to your disk :D01:24
elball of this because I was going to install typst and try to build slides with pdf accessibility built in01:24
elbbut typst is %!@#!@ rust01:24
minuteoh and what's the problem with rust? 01:24
elbit's a toolchain nightmare01:25
elband they're deeply unserious about security and reliability01:25
minuteelb: hmm. i would recommend "apt install rustup"01:25
elbyes, that's what I do01:25
minuteah01:25
elbthen I rustup update rustup and all its crap01:25
minuteoh01:25
elbbut the _recommended_ method is friggin' curl | sh01:26
minutetrue01:26
elbwhich makes me want to either cry or scream01:26
elbdepending on mood01:26
minuterustup update rustup not working on debian sounds like a debian bug tho01:26
elboh it works01:26
minutebecause rustup do recommend using apt on debian...01:27
elbbut then you transition from a system rust to an unmanaged rust in ~01:27
minuteyeah that's what you want tbh...01:27
elbI mean this is just how software is, now01:27
minutei guess the debian devs can't chase that much of a moving target01:27
elband rust has to be that way because every package I ever look at demands that i install a newer verison of the compiler01:28
minutefor python3 i recently started using "uv" to01:28
minutetoo01:28
minuteyeah, the debian rustc is usually old01:28
elbuv is also new to me01:28
elbI try to just keep my python3 needs moderate01:29
sigridwith uv I can actually run python scripts for once01:29
minutei'm not very good at python but uv was surprisingly painless01:29
elb3.8 or so added the type checking stuff that's important to me, personally, so past that I have little opinion01:29
- TechnoWizard (QUIT: Remote host closed the connection) (~TechnoWiz@user/TechnoWizard)01:35
elbok, confirmed that the system image works fine01:46
minuteuh oh01:47
minuteelb: btw... which one did you image?01:47
elbI used the "known working" one from 2025-07-2401:47
elbI'm guessing the non-booting kernel is 6.17.5-mnt-reform-arm6401:51
elbthe previous working terminal was 6.16.901:51
elbno, maybe .1201:51
elbyup, running 6.16.12 fine business again01:54
+ TechnoWizard (~TechnoWiz@user/TechnoWizard)01:59
- TechnoWizard (QUIT: Remote host closed the connection) (~TechnoWiz@user/TechnoWizard)02:02
- enwu (QUIT: Ping timeout: 256 seconds) (~enwu@user/enwu)02:03
+ enwu (~enwu@user/enwu)02:07
- chomwitt (QUIT: Ping timeout: 256 seconds) (~chomwitt@2a02:85f:9a5f:900:1ac0:4dff:fedb:a3f1)02:07
- enwu (QUIT: Ping timeout: 244 seconds) (~enwu@user/enwu)02:12
+ enwu (~enwu@user/enwu)02:14
- vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50)02:16
minuteelb: hmm i merged 6.17 patchset earlier because i interpreted josch's comments as "tested" on his devices. so it's actually broken?02:16
+ TechnoWizard (~TechnoWiz@user/TechnoWizard)02:16
minutei also tested several new system image builds a few hours ago and they worked (on reform rk3588 though, not pocket)02:17
elbit's certainly broken for me, at least along with whatever else happens in early boot02:18
elbit really acts like it's geetting continuous keyboard input, the cursor just keeps moving to the right02:18
elb(but it did it even with no keyboard attached)02:20
minuteok, i'm installing 6.17 update on my reform now02:25
minute> Error! Bad return status for module build on kernel: 6.17.6-mnt-reform-arm64 (aarch64)02:27
minuteFailed command:02:27
minutemake -j8 KERNELRELEASE=6.17.6-mnt-reform-arm64 -C /lib/modules/6.17.6-mnt-reform-arm64/build M=/var/lib/dkms/reform2_lpc/1.76/build02:27
minutetrying > sudo apt install linux-image-mnt-reform-arm64 reform2-lpc-dkms02:29
minuteok, that fixed it... reboot!02:31
minuteok, works for me on classic reform rk358802:33
- TechnoWizard (QUIT: Ping timeout: 264 seconds) (~TechnoWiz@user/TechnoWizard)02:36
- dok (QUIT: Ping timeout: 256 seconds) (~dok@2a01:e0a:12:d860:3696:e5bc:8411:9bc5)02:36
minuteelb: works for me on pocket rk358802:41
minute> Linux pocki3000 6.17.6-mnt-reform-arm64 #1 SMP PREEMPT Debian 6.17.6-1+reform20251104T001425Z (2025-11-0 aarch64 GNU/Linux02:42
+ dok (~dok@2a01:e0a:12:d860:3696:e5bc:8411:9bc5)02:44
elbok mine is 20251101, so this is your updated kernel with the new dtb?02:46
elbupdating now02:46
minuteelb: yes, very likely, the time checks out02:49
minuteelb: i really need to sleep, it's almost 3am. fingers crossed it works for you02:49
- mjw (QUIT: Ping timeout: 250 seconds) (~mjw@gnu.wildebeest.org)02:50
elbgood night :-)02:53
elbI'll let you know when I get it back up02:54
elbit's updating, my wi-fi from this location is very slow02:54
- paperManu_ (QUIT: Ping timeout: 244 seconds) (~paperManu@107.159.15.124)02:54
elbminute: confirmed 6.17.6-1+reform20251104T001425Z works properly here03:01
grimmwareI’ve not looked at uv yet. Work is predominantly python-based so I know my way around a virtualenv and have been using pipx for installing command line utilities03:03
grimmwareI’m not terribly keen on python to be honest. The idioms are syntax maximalist, especially list comprehensions.03:04
- kop316 (QUIT: Remote host closed the connection) (m-6f6zq6@static.138.159.90.157.clients.your-server.de)03:06
elbI put python in the "ok" department03:07
elbI strongly preferred ruby, but the language kind of lost to python and RoR03:07
+ kop316 (m-6f6zq6@static.138.159.90.157.clients.your-server.de)03:08
- kop316 (QUIT: Remote host closed the connection) (m-6f6zq6@static.138.159.90.157.clients.your-server.de)03:08
+ kop316 (m-6f6zq6@static.138.159.90.157.clients.your-server.de)03:08
- paperManu (QUIT: Ping timeout: 252 seconds) (~paperManu@107.159.15.124)03:28
+ paperManu (~paperManu@107.159.15.124)03:33
+ plomlompom0 (~plom@139.59.138.73)04:01
+ iank_ (~iank@fsf/staff/iank)04:02
+ wakest_ (m-a7d6fe@67.219.99.225)04:04
- plomlompom (QUIT: *.net *.split) (~plom@139.59.138.73)04:07
- colinsane (QUIT: *.net *.split) (~colinunin@97-113-142-187.tukw.qwest.net)04:07
- rwa_ (QUIT: *.net *.split) (0a82deb4eb@2a03:6000:1812:100::41b)04:07
- iank (QUIT: *.net *.split) (~iank@fsf/staff/iank)04:07
- jn (QUIT: *.net *.split) (~quassel@user/jn/x-3390946)04:07
- wakest (QUIT: *.net *.split) (m-a7d6fe@67.219.99.225)04:07
- qbit (QUIT: *.net *.split) (~qbit@user/qbit)04:07
+ rwa_ (0a82deb4eb@2a03:6000:1812:100::41b)04:07
+ jn (~quassel@user/jn/x-3390946)04:07
+ qbit (~qbit@user/qbit)04:07
* wakest_ -> wakest04:09
+ colinsane (~colinunin@97-113-142-187.tukw.qwest.net)04:18
- paperManu (QUIT: Ping timeout: 244 seconds) (~paperManu@107.159.15.124)04:27
+ chomwitt (~chomwitt@2a02:85f:9a5f:900:1ac0:4dff:fedb:a3f1)07:10
joschminute: my comment was indeed only the "booted" part, not that I tested much more than that. That's why I had the Linux 6.17 MR still marked as "Draft".07:26
- aloo_shu (QUIT: Ping timeout: 252 seconds) (~aloo_shu@85.51.18.212)07:34
+ TechnoWizard (~TechnoWiz@user/TechnoWizard)07:40
- TechnoWizard (QUIT: Remote host closed the connection) (~TechnoWiz@user/TechnoWizard)08:04
+ gidzit (~gidzit@gidzit.org)08:05
- gidzit (QUIT: Ping timeout: 256 seconds) (~gidzit@gidzit.org)08:30
joschminute: about reform-waybar-icon-wedge -- do you know if there is already an open waybar issue about this problem? Usually when I write ugly "hacks" I make it under the condition that the software with the problem has a bug about it and then I can reference that bug in my "hack" to proof to others: "see, this hack is necessary because upstream doesn't want it any other way" or "see, this hack is 08:38
joschnecessary because fixing the problem properly is much harder or too difficult"08:38
joschit at least gives the chance of the hack not staying forever or for those who hate the hack sufficiently to put in the work and fix it upstream08:38
+ gidzit (~gidzit@gidzit.org)09:14
minutejosch: there's at least a gtk bug about this (wontfix) iirc. someone requested and env var for icon theme10:29
minutejosch: and yeah there's something among the 1000 issues of waybar too, would need to dig it up10:30
joschminute: no worries, i can search for it too. I was just asking in case you had looked it yesterday or something.10:32
joschthe file /usr/share/xdg-terminal-exec/sway-xdg-terminals.list also sounds like something that should be going into a different package than reform-tools -- there might be a file conflict waiting to happen10:38
minutejosch: probably in sway, where i see a very low probability of that happening10:40
joschyes. if somebody complains we can move it10:51
grimmwareelb: yeah I prefer ruby, I find you can write much more terse but still readable code10:53
grimmwarebut I do think it's mostly that all the method chaining and shit like that fits the way I think10:54
grimmwareI'd prefer it if they dropped some of the syntactic sugar stuff and just stuck with parentheses10:55
+ aloo_shu (~aloo_shu@85.51.18.212)11:21
- RandyK (QUIT: Remote host closed the connection) (~RandyK@user/randyk)11:27
+ RandyK (~RandyK@user/randyk)11:27
MartiniMoeminute: I really don't mean to push and its totally fine if you answer "can't say", but I wanted to ask if you can somehow guess if my pocket reform might arrive before 39c3. Because otherwise I'll look for a different laptop to bring to the event, which would also be totally fine ^^ again, I really don't wanna push and its totally fine if it won't be shipped until then :D11:36
- chomwitt (QUIT: Ping timeout: 250 seconds) (~chomwitt@2a02:85f:9a5f:900:1ac0:4dff:fedb:a3f1)12:08
+ paperManu (~paperManu@107.159.15.124)12:18
+ mjw (~mjw@gnu.wildebeest.org)12:20
+ tisiphone (~quassel@user/tisiphone)12:35
+ gustav25 (~gustav@c-78-82-52-240.bbcust.telenor.se)13:02
minuteMartiniMoe: order date?13:09
MartiniMoeminute: 20th of september13:17
- svp (QUIT: Quit: Gateway shutdown) (~svp@2002:4f07:f0bd:0:69b1:b463:d245:e861)13:28
+ chomwitt (~chomwitt@2a02:85f:9a5f:900:1ac0:4dff:fedb:a3f1)13:46
- cwebber (QUIT: Quit: crossing the threshold of Lugonu's domain) (~Christine@user/cwebber)14:16
+ Serve0925 (~Serve0925@2a0a:ef40:da7:b301:11d9:40ba:e142:b53)14:16
- Serve0925 (QUIT: Remote host closed the connection) (~Serve0925@2a0a:ef40:da7:b301:11d9:40ba:e142:b53)14:35
+ Serve0925 (~Serve0925@2a0a:ef40:da7:b301:11d9:40ba:e142:b53)14:36
minutejosch: ok i'm back to trying to release the system image14:39
minute(for flashing here)14:39
minutelast night some gitlab glitch happened, so rerunning14:40
+ josch-pocket (~josch-poc@194.213.108.8)14:48
josch-pocketminute: would you like another reform-tools release with your recent changes for that?14:49
josch-pocketminute: are you flashing rk3588 units? Maybe this is a good opportunity to merge my outstanding u-boot MR and tag a new release of reform-rk3588-uboot? Then that release tag can be part of the reform-tools release as well14:50
minutejosch-pocket: ok both yes14:54
+ paperManu_ (~paperManu@modemcable141.205-200-24.mc.videotron.ca)14:55
+ cwebber (~Christine@user/cwebber)14:57
minutejosch-pocket: meh, uboot pipeline failed after merge15:00
minutebizarro > WARNING: Failed to pull image with policy "always": open /ramdisk/docker/tmp/GetImageBlob3135916763: no such file or directory (manager.go:203:1s)15:00
minutejosch-pocket: i'll assign the arm runner15:00
josch-pocketin the past, the x86 runner was necessary because there was a x86 binary blob run during u-boot compilation --it turned out that that proprietary blob did nothing so it's removed now15:05
josch-pocketand that's why it should work fine on the arm runner now even if it did not before15:05
minuteah nice. hope it'll work15:06
minuteotherwise: box64 :D15:06
josch-pocketi didn't get the blob to work with box64 only qemu workey15:07
josch-pocketbut it doesn't matter the blob is gone15:07
minuteyes, uboot build worked15:09
josch-pocketwith the new build.sh it should be very simple to do a release15:09
josch-pocketyou no longer have to edit a file with the name of the git tag15:09
josch-pocketit should be sufficient to just tag it15:09
minuteok, just make a tag?15:09
josch-pocketyes, that should work, lets see :)15:10
josch-pocketand then i copy the checksum of the binaries produced by the gitlab job into reform-tools machine.conf15:10
minutetagged15:10
josch-pocketgood the new pipeline is running15:10
josch-pocketi'll flash that to my pocket and see if it looks right15:11
- josch-pocket (QUIT: Quit: josch-pocket) (~josch-poc@194.213.108.8)15:23
minute:015:23
+ wielaard (~mjw@2001:1c06:2486:a800:158:adb1:a1e7:3c3f)15:25
- mjw (QUIT: Ping timeout: 264 seconds) (~mjw@gnu.wildebeest.org)15:27
+ svp (~svp@host-79-7-240-189.business.telecomitalia.it)15:33
joschit doesn't boot :(15:39
* wielaard -> mjw15:46
joschoh no it *does* work15:55
joschjust wonky emmc init i guess15:56
josch(i'm at work right now so i cannot have a look via uart)15:56
joschminute: coudl you change the gitlab runner for reform-tools to the arm64 machine as well? https://source.mnt.re/reform/reform-tools/-/jobs/1521816:13
joschi'd like to see a green pipeline result before i tag this16:13
joschbut unless you have anything else to add, the release should be ready16:13
joschoh, there is one problem with reform-waybar-icon-wedge which was not spotted because the CI didn't run16:18
minutejosch: ah sorry, one moment16:18
minutejosch: what's the wedge problem?16:19
minutejosch: arm64 ci pointed to reform-tools16:20
minute(now i mean)16:20
joschminute: man pages are automatically generated by calling the tool. Calling it with --help just starts another waybar instead of showing --help output. I have it fixed locally.16:21
joschAnother problem is that the makefile is not creating usr/share/xdg-terminal-exec/ in $(DESTDIR) -- also fixed16:22
joschit helps to have CI :)16:22
minutejosch: ah thank you! i didn't really check those last night, just did a best effort :D16:25
minutejosch: where did you put the fix though?16:25
joschno worries, easy to fix up16:25
joschi'm pushing in a bit16:25
joschi'm cleaning up the git history16:25
minuteah alright16:25
joschpipeline is green, tag 1.81 is created16:30
joschnow uploading to Debian16:30
- aelius (QUIT: Remote host closed the connection) (~aelius@user/aelius)16:38
- gidzit (QUIT: Ping timeout: 246 seconds) (~gidzit@gidzit.org)16:38
+ aelius (~aelius@user/aelius)16:40
joschand uploaded16:42
joschlets hope that nothing else was missed16:42
joschbut if it was, we'll do another release :)16:42
minuteah, uploading to debian... how can i build a system image with this version, or one that can be upgraded to debian later?16:53
minutei mean, how can i build an image now?16:53
minutejosch: is an image with reform-tools branch set, and using make install, installing a .deb that can be updated to a release version?16:53
minutejosch: ah, this looks like it's using reform-tools 1.8 from our git or.... ? https://source.mnt.re/reform/reform-system-image/-/merge_requests/14516:55
minuteyeah ok https://source.mnt.re/josch/reform-system-image/-/jobs/15210#L387116:55
- RandyK (QUIT: Ping timeout: 272 seconds) (~RandyK@user/randyk)16:55
+ RandyK_ (~RandyK@user/randyk)16:55
minuteso i'll update our factory installer to job 15210 and test that now16:56
* RandyK_ -> RandyK16:57
- Serve0925 (QUIT: Remote host closed the connection) (~Serve0925@2a0a:ef40:da7:b301:11d9:40ba:e142:b53)16:57
+ Serve0925 (~Serve0925@2a0a:ef40:da7:b301:11d9:40ba:e142:b53)16:57
minutejosch: ah woops i ended up in your repo16:59
minute1521616:59
joschminute: that is reform-tools 1.80 but the test version is 1.8117:00
minuteuff17:00
minuteok what do i have to do... my brain is cooked already17:00
minutei need an image right now... 17:01
joschminute: if you want to build a system image with reform-tools 1.81 package *now* then you need to adjust reform-debian-packages to build a *.deb from sources instead of pulling it from Debian17:01
minuteok i'm not capable of doing that atm17:01
joschsorry, i have had classes the whole afternoon (until 17:30) so my availability is also spotty17:01
minuteok i'll start with the old version and see17:02
joschi could've prepared this this morning if you had told me a bit earlier17:02
minuteyeah, somehow i didn't remember that this has to go through debian17:02
joschthe reform-debian-package machinery could be extended to build a .deb from scratch17:02
joschit doesn't have to17:02
minutejosch: so the "make install" thing does not cleanly install a package, but just installs the reform-tools content on the target fs?17:02
minutejosch: i mean if i specify a reform-tools branch in the CI variables17:02
joschbut if it's not going through debian first, then somebody else has to do the work of maintaining the version which is temporary on the MNT repo17:03
joschand the upgrade path between the two in case there are descrepancies17:03
joschyes, the "make install" only copies files17:03
minuteok, thanks for clarifications17:03
joschusing the "make install" method might be an acceptable hack if it's okay that dpkg-verify will show you problems until the user updated to 1.81 properly17:03
minuteok we'll abort here then and i'll just do some tests with the version with 1.8017:03
minuteand we'll try again tomorrow17:04
joschi have a system image here with 1.81 properly installed but i cannot upload it because uni wifi, sorry :(17:04
minuteok no problem... it would be great to make it possible in the future that i could act independently to make in-between versions17:04
josch(and you probably should not distribute what i build here to customers :D)17:04
joschminute: we'd have to talk about how you'd like the tooling to work for that17:05
joschbecause it would be an extra layer of complexity17:05
joschwe should've talked about that last Thursday :)17:05
minutejosch: yeah... i guess i wanted to fix too many bugs last minute17:05
minutebut it feels like some things can't be fixed later because there's no mechanism to import fixes from /etc/skel17:06
joschwell, there is always a way -- it depends on how hacky is still acceptable17:06
minutejosch: hmm but reform-debian-packages already builds a reform-tools .deb... can't i just rerun reform-debian-packages pipeline and yield 1.81?17:14
minutehmmm no it builds 1.80... ah, so that's what you meant by changing reform-debian-packages17:21
- mjw (QUIT: Ping timeout: 245 seconds) (~mjw@2001:1c06:2486:a800:158:adb1:a1e7:3c3f)17:31
joschso, the core of reform-debian-packages is build_patched.sh, the rest is just "bolted on"17:36
joschwhat build_patched.sh does is to download source packages from the debian archive, patches them and then builds them17:36
joschreform-tools is built like that right now17:36
joschnow we could special-case reform-tools to be built differently, that instead of pulling it from Debian, it pulls the packaging source from elsewhere and builds that17:37
joschbut then somebody needs to maintain that packaging source17:37
joschi mean you could pull the packaging source from salsa.debian.org but then you would again be dependant on me which you want to avoid17:37
joschso instead you have to start maintaining your fork of my packaging work and change it in lock-step with the changes I perform17:38
joschthis is extra complexity and extra work for you and it's up to you whether you would want to do this or not17:38
+ mjw (~mjw@gnu.wildebeest.org)17:48
- chrcav (QUIT: Ping timeout: 264 seconds) (~chrcav@user/chrcav)17:57
- mjw (QUIT: Ping timeout: 252 seconds) (~mjw@gnu.wildebeest.org)18:07
+ mjw (~mjw@2001:1c06:2486:a800:158:adb1:a1e7:3c3f)18:18
- paperManu_ (QUIT: Ping timeout: 260 seconds) (~paperManu@modemcable141.205-200-24.mc.videotron.ca)18:18
joschminute: this upload is actually a good example. Because things needed to be done quickly, I uploaded without testing. This means that I have a bug in what will now be in Debian. Somebody has to put in the work and fix the packaging changes which accumlate over time. (in this case, this was the fix: https://paste.debian.net/hidden/9e80487e/ )18:29
joschminute: sorry for the wall of text. I mean no disrespect. I know you have an extremely busy schedule and while I'd like to help, I 100% agree that it would be great if I was not a single-point-of-failure. I remember the last time you wanted a new reform-tools Debian package *now* I think it was a few months ago. I have performed over 80 reform-tools releases so far and I think the vast majority of them 18:31
joschwas without much urgency. I wonder if the extra complexity and the extra things you have to learn to then only really need them every 3 months or so is worth it?18:31
joschthis is your project and you get to decide how it's run. I would like to supply you with a better way forward than the status quo but cannot come up with one which would be an overall win but of course my perspective might be different from yours. Maybe somebody else has an idea. :)18:32
minutejosch: ideally what i want is that either reform-tools builds a .deb, and reform-system-package pulls in that deb18:33
minutejosch: or reform-debian-packages builds the reform-tools .deb18:33
minutejosch: but not from debian sources but from mnt reform-tools git18:33
joschright, but that either has to be coordinate with what is in Debian or you do a complete fork independent from Debian because otherwise the package versions will be incompatible (think of the upgrade scenario)18:34
minutejosch: why would they be incompatible? i would build 1.81 now, the same that you upload to debian...18:38
joschno, what i'm uploading to debian right now is 1.81-218:38
joschit's slightly different from 1.81-118:38
minutejosch: it sounds like we just have to make sure that the version in mnt repo is not numerically ahead of a newer version you upload to debian18:39
minutejosch: so that one can always upgrade to the debian version without losing functionality18:39
joschno, it's not about which version is larger because the apt pinning rules already prefer the MNT repos indpendent on which version is larger18:40
minutejosch: ok but 1.80 was built by reform-debian-packages and will go to the mnt repo, right? so that means users won't be able to install 1.81 by default until reform-debian-packages reruns with debian as reform-tools deb source?18:40
joschthat is correct18:41
joschbut that's because how reform-debian-packages is set up. This can be changed.18:41
joschanother example to illustrate what i mean when weighing complexities against each other:18:43
joschyou asked for an easier way to quickly put changes into reform-tools without editing a patch file, so I added the ./reform-tools directory to reform-debian-packages to make this easier18:44
joschuntil now, I have made 24 commits in the ./reform-tools directory (just to keep the reform-tools version in sync with Debian) and there are 2 commits from you18:44
joschso there is extra complexity and extra work for me for just two commits from you18:45
joschagain, please do not see this as any disrespect. I have no problem with helping18:45
joschbut to me, this sounds like a similar situation where more complexity is added, more work for me, which will only be used very rarely by you18:45
joschas in: it will be used in situations like today but these are "rare" no?18:46
joschi hope you underrstand what i'm trying to say and that i do not come across at wanting to remain the single point of failure, i don't -- i just don't have a good idea myself18:47
joschthis would be easier to talk about than to write about :)18:48
- chomwitt (QUIT: Ping timeout: 250 seconds) (~chomwitt@2a02:85f:9a5f:900:1ac0:4dff:fedb:a3f1)18:48
joschsorry, i realize this is frustrating for you right now because you had to scrap plans you had. Sorry :/18:53
minutejosch: ah, so i could just add the 1.81 patch to reform-debian-packages?18:55
minutejosch: i.e. the diff from 1.80 to 1.81?18:55
joschoften but not always18:56
joschunfortunately the Debian patch system has some limitations18:56
minutejosch: yeah but i mean today18:56
joschyes, try it and if it breaks i can explain you why it didn't work today :)18:56
minutejosch: i don't have the mental bandwidth to work on future proof solution today because i'm a bit sleep deprived, it would just be great to have a fix for this release which doesn't break the update to the next version(s)18:57
minutei just had to make a 6 page document pointing out cutting and sewing issues in the reform next sleeve 18:58
joschPlease try it out. I hope it works but no promises. I'm on my way home right now and will be back later (but by then, reform-tools 1.81 will already be in Debian unstable)18:59
josch(currently sitting at the station for wifi)18:59
joschfingers crossed and ttyl19:00
minuteah ok. then i'll just wait19:01
minutei stared at reform-debian-packages/reform-tools and don't know how and what patches to put in there anyways19:01
+ vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50)19:10
- pomel0 (QUIT: Ping timeout: 246 seconds) (~pomel0@user/pomel0)19:16
joschminute: maybe like this: https://source.mnt.re/reform/reform-debian-packages/-/merge_requests/14719:24
+ Murphy (~Murphy@76-223-253-78.lightspeed.frokca.sbcglobal.net)19:26
MurphyOne of my keys are stuck down after upgrading packages. I think it's spacebar. I tried resetting the keyboard and using the power switch. Any suggestions?19:28
- Serve0925 (QUIT: Remote host closed the connection) (~Serve0925@2a0a:ef40:da7:b301:11d9:40ba:e142:b53)19:30
+ Serve0925 (~Serve0925@2a0a:ef40:da7:b301:11d9:40ba:e142:b53)19:30
+ pomel0 (~pomel0@user/pomel0)19:31
minuteMurphy: uhm, that sounds like the issue elb had last night19:37
minuteMurphy: do you have any possibility to check which kernel version was installed, i.e. can you do uname -a ?19:37
minuteMurphy: (i guess hard with space key flood)19:37
minuteMurphy: when did you do the upgrade?19:37
MurphyI updated last night from a pretty old version on an SD card19:39
Murphyif I switch to my local version its fine, so it seems software related19:39
minuteMurphy: at which time exactly?19:39
Murphymust have been around 7ish pacific19:40
Murphyit took an hour or two to run an I think I started around 6pm pacific time19:40
Murphylet me check /boot on the SD card and see the kernel version19:41
minuteMurphy: ok, then you got the faulty dtb i think. can you do another update/upgrade of linux-image-mnt-reform-arm64?19:41
MurphyI can try, otherwise no big deal since it's an SD card install. if its a known issue and it was fixed, no worries19:42
minuteMurphy: i see, but in any case thanks for confirming it19:44
Murphyya I'll see if I can get the kernel info just incase it helps19:45
- mjw (QUIT: Ping timeout: 246 seconds) (~mjw@2001:1c06:2486:a800:158:adb1:a1e7:3c3f)19:45
MurphyIve been busy so I haven't had a chance to use my pocket. I picked up a MT7921AUN wifi/bt m.2 and im curious if it will work19:48
Murphyminute: its 6.17.619:56
Murphybtw is gnome-firmware supported yet? I don't see any new firmware since last time I was around19:57
minuteMurphy: thx. and yes, gnome-firmware is supported20:00
MurphyI ordered the new expansion board yesterday and upgrade board. wanted to make sure im on the latest version20:00
+ mjw (~mjw@gnu.wildebeest.org)20:14
- Serve0925 (QUIT: Remote host closed the connection) (~Serve0925@2a0a:ef40:da7:b301:11d9:40ba:e142:b53)20:16
- Murphy (QUIT: Ping timeout: 250 seconds) (~Murphy@76-223-253-78.lightspeed.frokca.sbcglobal.net)20:16
+ Murphy (~Murphy@76-223-253-78.lightspeed.frokca.sbcglobal.net)20:36
Murphyugh firmware-gnome crashes on first open then won't open again20:43
joschMurphy: could you open it from a terminal? Maybe there is useful output on stdout/stderr20:45
* plomlompom0 -> plomlompom20:46
Murphyfwupd_client_build_report_devices: assertion devices != NULL failed20:47
Murphyif I do verbose last message before that line is no report metadata20:48
Murphywonder if running from sdcard is the problem20:51

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!