| + jacobk (~quassel@47-186-122-163.dlls.tx.frontiernet.net) | 00:30 | |
| - mjw (QUIT: Ping timeout: 246 seconds) (~mjw@gnu.wildebeest.org) | 01:09 | |
| - jomo (QUIT: Server closed connection) (~jomo@user/jomo) | 01:20 | |
| + jomo (~jomo@user/jomo) | 01:20 | |
| minute | Boostisbetter: etnaviv is the gpu, not the display engine. the gpu doesn't know anything about the display(s). are you using single or dual display mode? | 01:24 |
|---|---|---|
| - murph[m] (QUIT: Server closed connection) (~murphhope@2001:470:69fc:105::d564) | 01:57 | |
| + murph[m] (~murphhope@2001:470:69fc:105::d564) | 01:58 | |
| - mtm (QUIT: Ping timeout: 245 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 02:03 | |
| - nsc (QUIT: Ping timeout: 246 seconds) (~nicolas@195-48-142-46.pool.kielnet.net) | 03:12 | |
| + nsc (~nicolas@228-49-142-46.pool.kielnet.net) | 03:14 | |
| + bkeys (~Thunderbi@static-198-54-130-101.cust.tzulo.com) | 03:35 | |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 04:09 | |
| - anzu (QUIT: Server closed connection) (~anzu@melkki.cs.helsinki.fi) | 05:17 | |
| + anzu (~anzu@melkki.cs.helsinki.fi) | 05:17 | |
| - ex-parrot (QUIT: Server closed connection) (~fincham@user/ex-parrot) | 05:38 | |
| + ex-parrot (~fincham@user/ex-parrot) | 05:39 | |
| - lifning (QUIT: Server closed connection) (~lifning@user/lifning) | 05:47 | |
| + lifning (~lifning@user/lifning) | 05:47 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 05:59 | |
| josch | would anybody like to be my guinea pig for the reform.debian.net stable apt repo? | 06:34 |
| - robin (QUIT: Ping timeout: 246 seconds) (~robin@user/terpri) | 06:35 | |
| josch | if yes, follow these steps: https://reform.debian.net/ | 06:35 |
| ex-parrot | keen | 06:44 |
| ex-parrot | my reform is apart at the moment but I can set this up | 06:45 |
| ex-parrot | exciting | 06:45 |
| ex-parrot | I was going to try get an out-of-the-box Bookworm going | 06:45 |
| ex-parrot | once I have it back together | 06:45 |
| Boostisbetter | minute: I would think single as I have never changed it. | 07:10 |
| - Boostisbetter (QUIT: Ping timeout: 240 seconds) (4a410829d7@irc.cheogram.com) | 08:14 | |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 08:21 | |
| - qbit (QUIT: Quit: WeeChat 3.8) (~qbit@2602:ff16:3:0:1:3a0:0:1) | 09:22 | |
| + robin (~robin@user/terpri) | 09:32 | |
| + robin_ (~robin@user/terpri) | 09:39 | |
| - robin (QUIT: Ping timeout: 246 seconds) (~robin@user/terpri) | 09:39 | |
| ex-parrot | going to try and safely remove the serial number sticker from my damaged base plate... | 10:03 |
| ex-parrot | Ok I warmed it up, peeled it off and stuck it on some adhesive laminating film to make it in to a tag. Now I’ll just secure that inside the chassis for all eternity. | 10:13 |
| ex-parrot | https://img.hotplate.co.nz/K7Joe9YMuu | 10:13 |
| josch | i still haven't put my sticker XD | 10:15 |
| ex-parrot | I'm not sure I recommend putting it on the outside of the bottom lid :P | 10:15 |
| ex-parrot | now I have a new lid I can do some Science on the old lid and see if the cracks would have been fixable by injecting resin | 10:17 |
| - deflated8837 (QUIT: Server closed connection) (~deflated8@50.53.206.61) | 10:27 | |
| + deflated8837 (~deflated8@50.53.206.61) | 10:27 | |
| ex-parrot | the cracks have continued spreading in the old lid so it definitely needed something done | 10:33 |
| - klardotsh (QUIT: Ping timeout: 246 seconds) (~klardotsh@98.97.115.75) | 11:02 | |
| + eery (~eery@77.137.76.80) | 11:23 | |
| josch | uff number #000037 -- that's impressive! :D | 11:24 |
| + Booster[m] (~boosterbo@2001:470:69fc:105::3:3d99) | 11:45 | |
| + mjw (~mjw@gnu.wildebeest.org) | 13:15 | |
| minute | josch: exciting @ debian reform repo | 13:32 |
| minute | i'm currently travelling to paris for FSiC2023 and only have my pocket reform with me | 13:32 |
| josch | that's also exciting! | 13:34 |
| josch | i started my switch to 100% reform also with a trip -- portugal to my partner's family for 3 weeks and all i had with me wath the reform and a screwdriver :) | 13:35 |
| josch | i was running kde plasma on the reform back then | 13:35 |
| josch | minute: i also have the hdmi situation fixed -- was there anything else that needed to be taken care of? | 13:35 |
| josch | c-keen[m], violet: i think i have fixed the hdmi situation but i cannot do a final test because i do not have a hdmi monitor -- would either or both of you like to test the fixed kernel? | 13:41 |
| - eery (QUIT: Remote host closed the connection) (~eery@77.137.76.80) | 13:57 | |
| - grubaroni (PART: !!unknown attribute: msg!!) (~user@user/grubaroni) | 13:58 | |
| minute | josch: oh great, thanks a ton for fixing it. i will test it after coming back on tuesday | 14:01 |
| minute | yesterday i improvised and watched a movie with my partner using pocket's hdmi and usb audio | 14:01 |
| - mtm (QUIT: Ping timeout: 250 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 14:03 | |
| josch | i just canceled a scheduled run of reform-debian-packages because it would've run a job in the middle of another pipeline i had running for the cdns-fix branch -- breaking both pipelines as a result | 14:20 |
| josch | i then looked for a way to force gitlab to only start jobs from a new pipeline *after* all older pipelines are done | 14:21 |
| josch | it seems there is an API call that allows setting process_mode to oldest_first that achieves exactly that: https://stackoverflow.com/a/74883391/784669 | 14:21 |
| josch | curl --request PUT --data "process_mode=oldest_first" --header "PRIVATE-TOKEN: <your_access_token>" "https://gitlab.example.com/api/v4/projects/1/resource_groups/production" | 14:24 |
| josch | minute: is this something you can look into next week? | 14:24 |
| - Boostisbetter (QUIT: Ping timeout: 260 seconds) (4a410829d7@irc.cheogram.com) | 14:38 | |
| minute | josch: my flight is delayed so i can click around in the web interface in a few minutes (or adjust settings via ssh on the build server) | 14:50 |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 14:53 | |
| josch | here is the gitlab upstream issue: https://gitlab.com/gitlab-org/gitlab/-/issues/16548 | 15:05 |
| minute | josch: you have root acccess to the builder, right? could you add this key to authorized_keys? http://dump.mntmn.com/id_rsa_reform.pub.txt | 15:10 |
| minute | (realized that my account on my pocket is not yet authorized on that server) | 15:11 |
| josch | minute: hrm... i'm not sure it's very responsible to give somebody ssh access if i cannot be sure whether the account named "minute" here is really Lukas or not -- do you have a gpg key that you can sign your message with? | 15:13 |
| minute | josch: ah, i thought my ownership of the domain might be enough :3 | 15:14 |
| minute | i can also put it on mntre.com | 15:14 |
| minute | i also don't have my gpg private key with me :3 | 15:15 |
| minute | i think i can also gain control over the server by using the remote console, lets see | 15:16 |
| josch | minute: true -- unlikely that anybody else wan do this | 15:16 |
| josch | one sec | 15:16 |
| josch | minute: done | 15:16 |
| josch | does it work? | 15:16 |
| minute | thanks, one moment | 15:17 |
| minute | ah, i realize i am also not opped here | 15:17 |
| * ChanServ changed mode (+o, minute) | 15:17 | |
| minute | josch: hmm, ssh access doesn't work for me. not sure what's wrong | 15:18 |
| minute | (i.e. it's asking for a pw) | 15:19 |
| minute | josch: you added me to root@ ? | 15:19 |
| josch | the username is mntmn | 15:19 |
| minute | oh | 15:19 |
| Boostisbetter | a password on the cert right? | 15:19 |
| josch | no, to /home/mntmn/.ssh/authorized_keys | 15:19 |
| minute | josch: alright, thanks | 15:19 |
| minute | works | 15:19 |
| josch | nice! :) | 15:19 |
| minute | i also have the sudo pw with me for some reason, so i have root access, thanks | 15:20 |
| + qbit (~qbit@h.suah.dev) | 15:20 | |
| minute | ok so the runner has builds_dir = "/ramdisk/builds" | 15:21 |
| minute | and executor is docker | 15:21 |
| minute | josch: not sure i understand how different build's data should affect each other | 15:21 |
| minute | do we need to disable the cache? | 15:21 |
| minute | disable_cache = false | 15:22 |
| josch | minute: i explained the problem here yesterday | 15:22 |
| josch | essenitally the problem is the shared build space | 15:22 |
| minute | ok, but why would it not create a different subdirectory for each build | 15:22 |
| minute | actually there is no /ramdisk/builds | 15:23 |
| minute | there is only /ramdisk/docker/volumes/ | 15:23 |
| minute | these are called sth like: | 15:23 |
| josch | the git repository for which the build is done always gets put into /ramdisk/builds/reform/reform-debian-packages | 15:23 |
| josch | but that is a path from inside the docker image | 15:23 |
| minute | huuuh | 15:23 |
| josch | the path outside might be different | 15:24 |
| minute | josch: well, docker containers can only write to their own temp storage, to their cache and to volumes mounted into the container | 15:25 |
| minute | we used to run the builds in a shell executor that used a shared /ramdisk/builds, but we don't do that since a longer time anymore | 15:26 |
| josch | let me write you a small script that demonstrates the behaviour | 15:26 |
| minute | ok | 15:26 |
| josch | minute: consider this .gitlab.yml: https://source.mnt.re/reform/reform-debian-packages/-/blob/cidemo/.gitlab-ci.yml?ref_type=heads | 15:30 |
| minute | josch: yep | 15:31 |
| minute | if "demo exists", i would consider that a bug | 15:31 |
| josch | minute: in the first pipeline run, the file "demo" doesn't exist: https://source.mnt.re/reform/reform-debian-packages/-/jobs/1499 | 15:31 |
| josch | in the second pipeline run it exists: https://source.mnt.re/reform/reform-debian-packages/-/jobs/1500 | 15:31 |
| minute | wtf | 15:32 |
| minute | this should def. not happen! | 15:32 |
| josch | so state is shared between multiple pipelines | 15:32 |
| josch | i agree | 15:32 |
| minute | not sure why this happens | 15:32 |
| josch | me neither -- i have little to no clue about gitlab unfortunately | 15:32 |
| josch | i just see the effects | 15:33 |
| minute | i just took away the old "builds_dir" setting, maybe that caused this... can you rerun your test pipeline? | 15:35 |
| josch | minute: maybe that was it! https://source.mnt.re/reform/reform-debian-packages/-/jobs/1501 | 15:36 |
| minute | josch: did you run it twice already? | 15:37 |
| josch | yes | 15:37 |
| josch | this is the third time | 15:37 |
| josch | or did you clean it up before? | 15:37 |
| josch | let me run it a fourth | 15:37 |
| minute | great. i did not do any manual cleaning | 15:38 |
| minute | but i did restart the runner after removing that setting | 15:38 |
| josch | it exists :( https://source.mnt.re/reform/reform-debian-packages/-/jobs/1502 | 15:38 |
| josch | so restarting the runner resets the state | 15:38 |
| minute | crap | 15:39 |
| minute | why would it behave like that.... | 15:39 |
| minute | josch: normally only /cache should be persistent | 15:40 |
| josch | i'm sorry i cannot be of more help and for having unearthed this frustrating issue :( | 15:42 |
| minute | well it's important to know | 15:43 |
| minute | i just removed more options from the config.toml and restarted | 15:43 |
| minute | i also set cache to disabled | 15:43 |
| minute | can you try a few more test builds? | 15:43 |
| josch | i'll schedule two more | 15:43 |
| minute | i am also looking at docker ps to see if docker is used at all :D | 15:44 |
| minute | (so far it's quiet, but maybe i'm not seeing the jobs with a watch interval of 1s) | 15:44 |
| josch | should i put a sleep into the job? | 15:44 |
| minute | yes please, maybe sleep 2s | 15:45 |
| minute | sleep 2 | 15:45 |
| josch | pushed | 15:45 |
| minute | on a positive side, working on remote admin stuff with pocket reform works well :D i just need to work on battery life | 15:45 |
| minute | ah, saw the container! | 15:46 |
| minute | it popped up briefly in docker ps, so docker is def. used :D | 15:46 |
| josch | now running a second time | 15:46 |
| minute | 779ce52029eb dc29b27ee12f "sh -c 'if [ -x /usr…" 1 second ago Up Less than a second ru | 15:46 |
| minute | nner-rp9pp3ds-project-67-concurrent-0-184ded908bce0abb-build-2 | 15:46 |
| josch | ++ echo 'demo doesn'\''t exist' | 15:47 |
| josch | aha! | 15:47 |
| josch | that looks good | 15:47 |
| minute | :0 | 15:47 |
| minute | i also don't see any leftovers in /ramdisk/docker/volumes | 15:48 |
| josch | then let me trigger the normal pipeline again to see if that still works | 15:48 |
| minute | yip | 15:48 |
| minute | while the build is running, there is a volume for it: | 15:49 |
| minute | runner-rp9pp3ds-project-67-concurrent-0-93e000d956e21e8c-cache-c33bcaa1fd2c77edfc3893b41966cea8 | 15:49 |
| josch | pipeline is running: https://source.mnt.re/reform/reform-debian-packages/-/pipelines/1080 | 15:49 |
| josch | lets hope that some state survives between the jobs of the same pipeline :D | 15:49 |
| minute | soo i think what was happening is that the entire build was happening inside a cache volume that was reused?! | 15:50 |
| minute | josch: ah, do you need to have state survive in the same pipeline? | 15:50 |
| josch | yes, or otherwise splitting the pipeline into multiple jobs isn't doing much | 15:51 |
| minute | ah :D | 15:54 |
| josch | minute: well at least you were very successfully in eliminating shared state :D https://source.mnt.re/reform/reform-debian-packages/-/jobs/1506 | 15:54 |
| josch | maybe the right way to do this is to use job artifacts to get the data from one job to the next? | 15:55 |
| josch | i'll have to investigate how that is done | 15:55 |
| minute | josch: thanks, otherwise lets undo the split perhaps | 15:55 |
| minute | if it's too complex | 15:56 |
| minute | but i think taking artifacts from another build is the way to go | 15:56 |
| josch | it seems that how i've done the split was incorrect | 15:56 |
| josch | because it only worked because of the (wrong) shared state situation that you fixed | 15:56 |
| josch | and in fact i need to declare artifacts... lets see... | 15:57 |
| minute | lol! | 15:57 |
| minute | we _can_ share state via volumes but i don't think it would be very clean | 15:57 |
| josch | yes, encoding what is getting passed on this in the .gitlab.yml sounds like a better idea | 15:58 |
| josch | okay, new pipeline is running. I have no idea whether a job can depend on multiple jobs writing into the same directory and whether the contents of that directory are then merged... but i guess we'll see. :) | 16:04 |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 16:08 | |
| josch | Downloading artifacts from coordinator... ok id=1510 responseStatus=200 OK token=64_852Ne | 16:09 |
| josch | nice :) | 16:09 |
| minute | nice | 16:19 |
| minute | i think you can make complex pipelines that depend on artifacts built on multiple builders in parallel etc | 16:20 |
| Boostisbetter | minute: the Amiga shot on the website is also great! I can't help but wishing the 8-bit guy would do some brightening on your 3000 though. Do you know I never got to experience the Commodore or Amiga. I was only exposed to the x86 world of the PC. | 17:06 |
| Boostisbetter | A real shame. | 17:06 |
| Boostisbetter | It really makes me wonder how the inferior processing might of those products didn't win out. | 17:30 |
| Boostisbetter | just goes to show you how things live or die based on the intensity and quality of their marketing campaigns | 17:30 |
| minute | Boostisbetter: not a fan of 8 bit guy | 18:51 |
| minute | Boostisbetter: commodore and amiga was pretty cool back in the day! | 18:52 |
| Boostisbetter | minute: really? You just don't like his personality, or is it something more serious? I don't like all of his content, but I really like Planet X, and I think the X16 seems interesting as well. | 19:12 |
| Boostisbetter | I also really appreciated the history series on the Commodore and Amiga products. For someone who never used even one of them, I thought it was very interesting. | 19:15 |
| minute | Boostisbetter: i got really turned off the politics, open carry advocacy, stunts with assault rifles in public etc | 19:23 |
| minute | by the | 19:25 |
| Boostisbetter | minute: I understand. I very much support open carry, not so much stunts with assault rifles in public though. | 19:28 |
| minute | :S | 19:29 |
| Boostisbetter | differences in opinion are one of life's only certainties it seems. I will say that I genuinely wished we lived in a world where all weapons could be destroyed now and forever. | 19:30 |
| Boostisbetter | but such talk is inappropriate here I think, so my apologies for bringing it up. | 19:31 |
| Boostisbetter | My Reform bricked again today. I need to get the lead out and upgrade my LPC and keyboard firmware. | 19:31 |
| Boostisbetter | I think tomorrow I am going to book some time in the afternoon, if the heat isn't killing me, to knock it out. | 19:32 |
| minute | i just arrived at charles de gaulle airport and trying to make my way to paris. | 19:32 |
| josch | still on the pocket -- nice! :) | 19:33 |
| Boostisbetter | minute: indeed! Being on the pocket is awesome. All the best in Paris. Hope you find some amazing food while you are there! | 19:44 |
| minute | no, now i am on iphone lol | 19:45 |
| minute | sorry to disappoint | 19:46 |
| minute | i did charge the pocket in the plane though | 19:46 |
| minute | they had a 60W usb-c outlet | 19:46 |
| - jacobk (QUIT: Quit: http://quassel-irc.org - Chat comfortably. Anywhere.) (~quassel@47-186-122-163.dlls.tx.frontiernet.net) | 20:12 | |
| Boostisbetter | noice! | 20:43 |
| - deflated8837 (QUIT: Read error: Connection reset by peer) (~deflated8@50.53.206.61) | 20:46 | |
| + deflated8837 (~deflated8@50.53.206.61) | 20:49 | |
| - bkeys (QUIT: Ping timeout: 246 seconds) (~Thunderbi@static-198-54-130-101.cust.tzulo.com) | 20:51 | |
| + bkeys (~Thunderbi@static-198-54-130-101.cust.tzulo.com) | 20:55 | |
| josch | minute: you based your a311d branch on top of the pocket reform branch which i now cleaned up and rebased on top of the fixed ci branch | 21:30 |
| josch | minute: since pocket reform needs kernel 6.4, would like me to create another branch (i'll not mess up your a311d branch) containing your a311d patches on top of 6.4? | 21:30 |
| - Gooberpatrol66 (QUIT: Quit: Leaving) (~Gooberpat@user/gooberpatrol66) | 21:37 | |
| - bluerise (QUIT: Ping timeout: 264 seconds) (~bluerise@user/bluerise) | 22:26 | |
| + bluerise_ (~bluerise@p5b211f1e.dip0.t-ipconnect.de) | 22:26 | |
| + amospalla (2a2f8bc48e@telesto.hot-chilli.net) | 22:29 | |
| amospalla | hi, just configured a IRC client besides the webchat. | 22:33 |
| - buckket (QUIT: Quit: buckket) (~buckket@vps.buckket.org) | 22:40 | |
| + buckket (~buckket@vps.buckket.org) | 22:41 | |
| ex-parrot | morning | 22:52 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 23:03 | |
| pandora[m] | <josch> "maybe the right way to do this..." <- It is! That makes things also more reproducible without running everything again :) | 23:07 |
| josch | minute: | 23:08 |
| josch | error: arch/arm64/boot/dts/amlogic/meson-g12b-bananapi-cm4-cm4io.dts: already exists in index | 23:08 |
| josch | error: arch/arm64/boot/dts/amlogic/meson-g12b-bananapi-cm4.dtsi: already exists in index | 23:08 |
| josch | 6.4 already seems to include lots of stuff | 23:08 |
| + amospalla93 (~amospalla@212.231.228.113) | 23:15 | |
| - amospalla (PART: !!unknown attribute: msg!!) (2a2f8bc48e@telesto.hot-chilli.net) | 23:15 | |
| * amospalla93 -> amospasdfsdfsdf | 23:16 | |
| - amospasdfsdfsdf (QUIT: Client Quit) (~amospalla@212.231.228.113) | 23:16 | |
| + amospalla18 (~amospalla@212.231.228.113) | 23:16 | |
| - amospalla18 (PART: !!unknown attribute: msg!!) (~amospalla@212.231.228.113) | 23:16 | |
| josch | minute: 42 files changed, 70 insertions(+), 3479 deletions(-) | 23:30 |
| josch | minute: so rebase onto 6.4 allows dropping 3.4 thousand lines of patches | 23:30 |
| + amospalla (2a2f8bc48e@telesto.hot-chilli.net) | 23:32 | |
| - amospalla (PART: !!unknown attribute: msg!!) (2a2f8bc48e@telesto.hot-chilli.net) | 23:39 | |
| + amospalla (2a2f8bc48e@telesto.hot-chilli.net) | 23:40 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!