00:48:06 hi, does anyone know if Google Coral USB Accelerator works in FreeBSD? 00:49:03 uskerine: what does pciconf -lv say when you plug it in 00:53:08 uskerine: https://coral.ai/docs/accelerator/get-started/#2-install-the-pycoral-library that? 00:53:10 Title: Get started with the USB Accelerator | Coral 00:58:54 I do not have any, I am thinking in buying one 01:04:08 uskerine: in that case may reconsider 01:09:22 Unfortuantely FreeBSD support for GPU/TPUs is simply not there 01:19:30 looks like there's source code for the driver, but it's not clear how hard it would be to port 01:21:02 uskerine: want to do some kernel driver hacking ? 01:24:42 I would love to, but it seems out of my league 01:26:19 uh well 01:26:50 uskerine: could of just asked chatgpt to do it then submit it as your own 01:26:55 uskerine: come on man 02:48:58 i can't imagine how disastrous a chatgpt driver would end up 02:52:23 But it's look plausible. 02:52:31 It'd*. Like my typing, sometimes. 02:56:44 heh 02:58:14 had a local guy that was trying to do a custom mouse build; he was following along with an open source setup using arduino libs flashed to banana pi leaf s3, but he wanted to customize it with a different (higher resolution?) sensor 02:58:32 he asked chatgpt to write the arduino sketch he'd need to flash to work with the new sensor, and the result was... quite hilarious 02:59:28 it was absolutely confused, decided it would talk to the Banana Pi from an actual Arduino device and write serial out data from the other side and it just... didn't fit together in all the wrong ways 03:02:22 That's kind of wild. 03:11:39 yeah... i attempted to redirect him to a better way to port the existing sketch to his new sensor with some good ol' fashion pick 'n' pull (no time myself to try and get involved), hopefully that works out a little better 03:24:54 yikes 03:42:44 This article about the trainers of AI is really illuminating - there's a bit where they talk a little about how the model is "trained" for python and it's all so much less high-tech than everyone seems to assume: https://www.theverge.com/features/23764584/ai-artificial-intelligence-data-notation-labor-scale-surge-remotasks-openai-chatbots 03:42:45 Title: Inside the AI Factory: the humans that make tech seem human - The Verge 03:44:59 ah, the mechanical Turk strikes again :-) 03:47:46 this person did at least get bonus points for trying, and they had a healthy skepticism going into it but hoped it wasn't as bad as it ended up being 03:48:20 and to chatgpt's credit, it did at least advise on proper SPI hookup in the same answer 10:19:43 Hi all, 10:19:47 I couldn´t see any post about it on any of the mailing lists, so I thought it could be that you´re not aware of it: This past Monday, the 26th, Hans Petter Selasky, whom I believe many of you know, died in a car accident. 10:19:56 There will be a sermon in Norkirken in Grimstad, Norway this Saturday. 10:20:17 :( 10:21:52 I am sorry to hear. Since some months earlier I had started to notice Hans' name. :-( 10:32:46 (paywalled articles in the local news (in Norwegian): https://www.gat.no/nyheter/i/jlx7oe/bil-kjoert-av-veien-og-ut-i-vannet-ved-kaldvell https://www.agderposten.no/nyheter/i/4olA8o/hans-petter-41-doede-i-utforkjoeringen ) 10:32:48 Title: Omkom i ulykken: - Var en av støttepilarene - gat.no 11:22:08 magnus_reftel: thank you for sharing these sad news 12:04:59 Hello 12:05:07 I get this tar (child): Cannot connect to mtp: resolve failed 12:09:25 Phone set to the proper mode ? 12:23:12 How to make webcam be accessed by multiple programs? it seems it cannot be for now? 12:23:21 at the same time I mean 12:29:31 Don't think you can do that but that's more of a driver issue rather the program 12:30:03 you'd need some userspace program to multiplex v4l2 api calls, which i don't think exists (at least not for freebsd) 12:31:18 Think of moused vs direct x mouse usage , using moused you can have the mouse on console and x using the device directly in x you have no mouse on console as x is using it or locked to it 12:31:44 You would need someone similar for webcam 12:34:53 You could kind of get that with a multiplexer box kind of like the old days when you had a box for a printer and you would change a dial for the computer , they sell those but not sure if it will do what you need as it's still communicating in a 1 to 1 12:59:43 cpet debdrup I see so I cannot record with obs while having a google/zoom meeting 13:00:30 Corrwct 13:00:36 Correct 13:00:46 that is not the case with linux. right? 13:05:01 BSD and Linux are two different things work differently have way more support and funding so 13:05:51 ok 13:13:32 Anyone uses borg for backups? Speaking about personal workstation, not enterprise level solution. 13:20:41 I probably should have asked that question on #freebsd-desktop, haven't I? 13:33:23 Yes please use the correct hashtag channel 13:33:44 cdrmack: meena will get upset 13:34:03 cdrmack: can we have some more info though ? 13:35:44 Apologies, I moved my question to the -desktop channel. 13:36:10 cdrmack: no need to actually listen to cpet 13:36:47 incomplete list of channels, official and unofficial https://wiki.freebsd.org/IRC/Channels 13:36:48 Title: IRC/Channels - FreeBSD Wiki 13:36:50 cdrmack: nonned to actually listen to meena 13:37:26 cdrmack: just post errors and wait 13:38:00 :P 13:47:13 cdrmack: it's stupid so I always make it known but yeah post some errors 13:50:32 I do not have any errors. I'm new and looking for opinions about backup software for personal use. I did some research on my own and that's how I found borg. 13:50:44 cpet: come join me in #freebsd-bugs to see why it might be useful to compartmentalise certain topics 13:51:25 cdrmack: what's your use-case / requirements? 13:51:59 meena: yeah no 13:52:39 cdrmack: there's a few in ports backupme is what I used 13:52:48 I guess I want to backup my personal stuff on daily basis and probably some FreeBSD (kernel? full-system?) stuff before any upgrades 13:52:57 cpet: in that case, please, for love piss, and all that is golden, shut the fuck up about it. 13:53:02 Backupme should be fine then 13:54:02 meena: so you love piss and like golden things ok 13:55:09 cdrmack: zfs snapshots ? If you have some extra hw truenas is awesome for that 13:56:54 cpet: I have Synology DiskStation, still reading about ZFS but I have realized it makes snapshots before each upgrade already. Need to go through the docs to understand how these snapshots work. 13:57:18 thanks 13:58:44 zfs makes a snapshot and saves it to .snaps then you would use zfs to either restore them 13:59:19 Synology think has a network option to mount the share so you can backup was you need there better option if you have one of those 13:59:35 I used to use backupme with a mix of MySQLbackup 13:59:51 But now I just don't care after I got out to worry about anything so bleh 14:03:56 cdrmack: however remember of the disk dies you lose all the snaps 14:06:15 * V_PauAmma_V frowns at cpet's unwarranted assumptions about meena. 14:07:00 Need to do more reading but thanks for sharing. I do not have too much data so probably any solution will do. 14:07:16 Will read more about ZFS snapshots. 14:07:42 Yeap 14:12:52 cdrmack: which model do you have ? 14:13:44 * cpet frowns back at V_PauAmma_V 14:17:31 cpet: DS218j 14:17:47 Is that a 5 bay one ? 14:18:21 bay = disk slot? If yes then nah, there are 2. 14:18:29 2 bay 14:18:52 cdrmack: bay, slot same shit 14:21:21 cdrmack: should invest in one of those , I have a Lenovo small tower that has a mirror zfs that I updated to current and dealing with openssl cluster fucks but managed to fix most except for openldap having a Synology would of prevented downtime 14:29:31 noice 14:29:44 I have it but do not have automated backups in place. 14:29:47 so you still win, I guess :P 14:31:02 Configs are in a fossil repo 14:31:14 Back up those is pretty much all I backup 14:34:00 bt 14:34:23 oops irc doesn't yet show backtraces 14:35:47 Nope 14:43:45 paulf: dtrace is fun 14:46:05 cdrmack: a simple script that tars up etc and does a pkg list dump and home then copies it over to the Synology is all you really need doesn't have to be some overly complicated crap with a GUI and python 14:46:06 was just a window focus error, as usual, had a gdb session on the 2nd monitor 14:46:40 cpet's awful signal to noise ratio combined with inability to either stfu or get a clue makes this channel not worth reading. bye. 14:48:31 cpet: I will probably do the same, maybe also include current kernel just in case. 14:49:14 And we upset koston bye koston 14:49:39 cdrmack: not worth the noise to go edge though 14:50:50 cpet, *plonk* 14:56:20 V_PauAmma_V: https://github.com/dmentipl/plonk 14:56:21 Title: GitHub - dmentipl/plonk: Smoothed particle hydrodynamics analysis and visualization with Python. 15:02:10 cpet: you are probably right 15:08:10 cdrmack: it happens every now and then 15:30:43 •/nick magnus_reftel 15:33:43 Fail 15:36:38 is the quarterly pkg branch on a strict release timeline, or is it "roughly quarterly"? 15:38:41 https://wiki.freebsd.org/Ports/QuarterlyBranch 15:38:42 Title: Ports/QuarterlyBranch - FreeBSD Wiki 15:39:33 thanks, that's the doc i was looking for 15:39:56 Strict release, as far as I can tell. Next branch is due on July 1. 17:09:44 Freaky: https://freshbsd.org/freebsd/src/commit/a18b956b73cee784e5c422d20fd0e4dabebd7eee any chance to fix mention of multiple PRs? 17:09:45 Title: FreeBSD / src / a18b956 / libexecinfo: Enable backtrace_test.backtrace_fmt_basic on amd64 again - FreshBSD 17:10:12 Freaky: we probably have a few styles 17:45:52 Commit logs shouldnt really be that long , I did this change and that change k thanks bye 17:47:23 meena: it's actually not as complicated as you might think 17:48:42 You can setup git to have pre hooks that gives you a template when you commit , or setup formating and all that 17:49:20 meena: obviously this isn't exactly what's used anymore, but it should still roughly match: https://svnweb.freebsd.org/base/svnadmin/hooks/scripts/notify_bz.sh?revision=331884&view=markup 17:49:22 Title: [base] Contents of /svnadmin/hooks/scripts/notify_bz.sh 17:49:24 Or you can mess up and have a hat put on aka pointy hat to x 17:50:27 iirc there must be spaces, comma is optional 17:53:11 kevans, Got "installworld" completed for -CURRENT without issues; this time I ran "adjkerntz -i" (in single user mode) that I had not done earlier; will update the mail thread later. Thanks again for the pointers 17:53:14 If you look at the base dir and look at the . You can see the hooks and all that is done when you commit or do other things 17:54:15 not all of the hooks are published at the moment 17:54:18 only the client-side hooks 17:55:34 Never cared much to be a committer as I did it for fun and the praise on a commit logs was enough so never really cared about the internal bits but if do some magic I did have over 900 PRs which was enough for such 18:03:51 https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=54297 think this was my first one spending all that time to be told it's already there , cause I forgot to look but this was back then when we used gnats and it didn't do search while your to so it was easier to duplicate things. Fun times 18:03:53 Title: 54297 – New port fceultra 18:04:43 How do I get rid of xorg log too much frequency (EE) Wacom ^COne by Wacom M Pen eraser: Error reading wacom device : Invalid argument 18:04:49 its eating my dish 18:05:51 Edit startx and add in 18:07:18 cpet add what? 18:07:22 -logfile /dev/null 18:07:38 Or 18:07:40 the error shows in the middle of somewhere in the day. not related to start x i think 18:07:58 ok so there will be no logs 18:07:59 -logverbose 0 18:08:14 actually I can route the logs to /tmp.   /tmp is on my ram 18:08:24 how can I do that? 18:08:31 Do what ? 18:08:45 route the /var/log to /tmp. /tmp is on my ram 18:09:09 Instead of dev null ad in path to tmp file 18:09:16 yes, but how 18:09:26 Told you edit startx 18:09:55 parv: ah, nice 18:10:12 parv: adjkerntz without having to rebuild? 18:10:26 don' t you think I should mount /var/log  to /tmp/var/log in  /etc/fstab? 18:10:26 cpet ^ 18:11:17 No the fact that x bitches so much about Wacom means there's an insurance with the drivers 18:11:27 If you don't need the logs just nuke them 18:11:28 or maybe just for /var/log/Xorg.0.log 18:11:54 ok but I would need other logs 18:11:58 like dmseg etc 18:12:22 kevans, What do you mean? I started the sequence by cleaning out everything (just to be really sure sure) & then making buildworld in multi-user mode; boot in single user; ran adjkerntz; then making installworld 18:12:26 -logfile is a xorg arg to change the log file name 18:12:30 ahh 18:13:05 Ntpdate is probably a better bet 18:13:21 cpet wher is that `startx` script? 18:13:57 kevans, I had tried to correct the time via "date" but that did not seem to have done the trick previously. Oh well, adjkerntz works to be certain 18:14:01 which startx 18:14:16 If you use those dumb xdm things you need to edit that config 18:14:49 it's in xinit package, if that's what you are asking 18:15:00 cpet found it 18:15:08 Beladona: however note I don't log anything until I need to 18:15:10 where to add that? anywhere? 18:16:01 Should see a line that start with Xorg yada yada with a few - such as -notcp add it towards the end 18:16:28 This won't work if you use something like sddm 18:16:48 cpet https://imgur.com/fHt4hU1.png 18:17:03 I have to restart x. How can i enforce the log file path change right now? 18:17:18 Restart x 18:17:35 Imgur is blocked on my end using a personal squid 18:17:44 So God speed to your edit 18:17:58 any wy without restarting? 18:18:06 Nope 18:18:36 Think a -HUP will still kill the x session 18:20:02 However won't really fix your issue though as a tmpfs still has a limit null does not 18:20:39 May get away with using nullfs for car/log2 and changing it to that 18:21:11 ok 18:23:30 Or figure out why the Wacom spams start hacking the driver and remove that bit think you can also make syslogd not log the same thing over and over 18:25:04 cpet i don't think `-logfile` is correct syntax for startx file 18:25:21 it is correct for `Xorg -logfile /path` though 18:25:28 but Fatal server error: (EE) Server is already active for display 0  If this server is no longer running, remove /tmp/.X0-lock and start again. 18:27:14 Error is self explained 18:27:29 ya 18:28:45 I'm sure you'll figure it out 18:29:04 I am not. mayb :) 18:29:33 One of the nice thing about nix is the ability to modify things to suite you even if they are non std or frowned upon 18:29:58 ya.. 18:32:01 Firs thing I do when install freebsd is change periodoc.conf and change all instances of root to /var/log/periodic.conf then I sysrc sendmail_enable="off" then I edit newsyslog and change all log files to have a .log then I do the same for syslogd.conf then I edit the crontab to move backups to /var/backups some may think this is wrong but I don't care I still do it 18:32:28 hm 18:36:56 cpet ln -s /dev/null /var/log/Xorg.0.log warning: An error occurred while redirecting file '/dev/null'  open: Operation not supported 18:37:17 Yeah that's a no go 18:37:24 am. why 18:38:39 That's a dataset you may get away with symlinking the actual xorg log file 18:39:15 why it works in linux https://pastebin.ubuntu.com/p/fFMqySTPns/ 18:39:17 Title: Ubuntu Pastebin 18:39:21 So ln -s /dev/null /car/log/Xorg .0.log 18:39:39 make sure that reads var I'm on my phone and autocorrect is stupid 18:40:06 did you wrote car instead of var 18:40:09 I don't do pastebins so yeah 18:40:15 oh ok 18:40:30 cpet ln -s /dev/null /var/log/Xorg.0.log warning: An error occurred while redirecting file '/dev/null'  open: Operation not supported 18:40:47 what is the display manager in freebsd? 18:41:01 Which ever you chose when you install 18:41:16 Guess you can't symlink a file to a device 18:41:52 depending on the wm we have gdn xdm sddm 18:42:08 Ok bye 18:43:46 Backwards as well 18:44:25 meena: been on the TODO since forever 18:45:19 %r{\b((?:#{cats})/(\d+))}, %r{\b((?:PR)[:/#]?\s*)((?:#{cats})/)?(\d+)\b} 18:45:52 Ugly but works 18:47:23 the bright side is that it still gets the first one rather than breaking because the whole line doesn't match expectations 18:47:46 Neither do life experiences 18:53:25 oops 18:53:58 it mentions cats, so it must be good 18:55:35 had to kill -9 the webserver because I wedged it 18:55:57 band-aided it and neglected to update it after FreeBSD migrated to git 18:57:49 Yeap whole world nigratea to git let's do it too instill have issues doing the equivalent in svn svn diff -N 18:58:11 git <3, svn git is alright once you learn to accept the horrible UX 18:58:42 i will die on the hill that git has a horrible UX, though 19:00:02 I guess we can finally agree on one thing 19:00:46 i am very afraid of using svn because you can't mistakes when committing, with git i can at least still edit that ugly piece of text that are my commit messages (and add all the files i missed initially, and fix every other thing i tend to forget) 19:01:19 I don't like git , I actually export git to svn to work on it or import it to a fossil repo that shit just looks beautiful 19:01:35 grr overmind 19:01:41 Svn commit -m "edit file" 19:02:00 yuripv: that typo amused me way too much, I'm so dumb :D 19:02:04 kill -9 on one service should not terminate everything else in the Procfile 19:02:20 debdrup: shall I respond to that ? 19:02:39 cpet: If you can't act civilized, you should probably leave. 19:02:43 Freaky: you want HUP 19:03:02 I gave an option so I am 19:03:08 I want a regex engine that doesn't go quadratic if you look at it funny 19:03:18 and responds to SIGTERM, apparently 19:03:38 Freaky: that pretty much excludes all regex engines. 19:03:42 Turn your head the other way 19:03:57 And it's quadratic in the best case scenario, so far as I'm aware. 19:04:29 O(log²!) is probably more likely. 19:04:42 no, good ones are linear, they just happen to not be the ones you find in most scripting languages 19:05:17 Such as ? 19:05:19 https://freshbsd.org/freebsd/ports/commit/91473dddb1daaf7d393dd2782e5c818456ef9750 there, can render that classic again 19:05:21 Title: FreeBSD / ports / 91473dd / Update to r47441 of the FreeBSD Documentation Set, - FreshBSD 19:06:58 Freaky: you have gained the rank of "the freshbsd dewd" 19:07:09 I assume it was the result of vim misuse 19:07:28 i was going to ask what '["203274" repeated 18,535 times]' means, then i did git log; didn't know about that one 19:08:07 that's what you get for using svn 19:08:28 cpet: Rust's regex crate doesn't do this, don't think Go's does either 19:08:54 I like rust I don't like go cause it's Google related 19:10:31 Freaky: i still make sys admin programs in m3 cause no one even know wtf that is 19:11:02 Freaky: I got into that after I used cvsup for so long 19:11:15 Always liked the cde like iFace 19:12:57 oh yeah, it was written in Modula-3 19:13:39 I remember it being replaced with csup because it was C and so less annoying 19:15:56 debdrup: aren't you a committer ? 19:16:43 Freaky find it funny when I make a small program that does this and someone else goes what the hell is this made in 19:56:46 otis: i think you are missing at least the panic message itself in that nvme panic report 20:00:49 ah, yuri left before i caould write that i don't have any. i've left a tip session to that vm to capture the eventual output.