07:31:14 holly crap DDR4 server memory has gone through the roof 07:57:55 why? 07:58:10 what happened 07:59:30 some sort of spare parts bin explosion? 08:32:25 what's happened? 09:03:38 prices more than doubles in the last week 09:13:55 ivy: do you mean literally that some warehouse caught on fire and wiped out a substantial part of world supply of memory? 09:15:09 seti_: no, i was just wondering what sort of event would cause memory to go through the roof 09:15:19 i see i see 09:17:27 prep for black friday, so we can be offered substantial reductions lmao 09:18:02 calling malloc too often (or with a size that is too big) 09:18:50 missed the beginning so I didn't see if this is userland or kernel 09:20:23 paulf: the conversation was about the price of memory rather than its usage :-) 09:21:16 well in that case it's probably just the usual semiconductor cycle 09:22:17 new node comes out, prices rocket until fabs switch to that node and then there is a glut and prices fall until the next node ... 09:22:18 and how would that 'cycle' be? 09:24:08 about every 3 years 09:24:22 it's not like ddr4 is the new hotness 09:24:53 no but the 2nm node will be 11:33:58 I'm trying to install mozilla syncserver, but (of course) it fails to build the rust crate "grpcio-sys". 11:34:37 Is there an easy(?) way to find ports that may be pulling in the same crate, so I can figure out what those are doing to make it build? 13:29:08 Ltning: i only see py-grpcio related ports, no crates :\ 13:30:02 can freebsd run ollama and large language models and all on gpu? 13:31:01 response from gpt-oss: No – there is no supported way to run Ollama on FreeBSD with GPU acceleration. 13:31:21 https://forums.freebsd.org/threads/ollama-working-with-deepseek-r1-deepseek-coder-mistral-nvidia-gpu-and-emacs-with-gptel-on-freebsd-14-2.96500/page-2#post-719399 13:33:48 i like gentoo too.. but somehow like freebsd too.. can i asc why you specialy use freebsd and why not gentoo if is so customizable? 13:34:28 iRobbery 13:34:30 thaks 13:34:31 thanks 13:36:18 I use freebsd because its not such a mess as linux is, no systemd, MIT license, ports system, zfs native, ipfw firewall, and more if i really will think about it 13:37:54 i use openrc 13:38:06 my gentoo is special 13:38:15 use like nothing of memory 13:39:30 gentoo users are special ;) 13:39:34 here are packages of one server, they use target x86_64 so is for all computers prepared...: http://gentoobins.grandekos.com ... :) 13:41:25 i like to make a team.. like we was when we was jung.. on ircnet.. similar but with more responsability.. anyone like to think and join and work on projects? programming? msg me.. *** 13:42:07 even electronics are our projects.. drones printers similar 14:19:02 ivy: misordering also messes up dtb package creation i saw 14:19:35 ketas: i've never seen that and i build for arm64 fairly regularly, could you file a PR with logs etc? 14:22:00 i did put it to current ml once, had to run make packages twice to get a dtb package... unless this is 13 make that misorders this, it's bug... and i don't see .order for that either 14:22:23 :/ 14:22:29 me or bug? 14:22:46 i have no idea, this is why i asked to see the logs :-) 14:25:06 although if it only affects freebsd 13, that's pretty low priority for pkgbase issues 14:26:31 as a build host, i was building 16 14:28:02 logs at https://lists.freebsd.org/archives/freebsd-current/2025-September/008958.html 14:37:32 .ORDER: create-kernel-packages-flavor${flavor:C,^""$,${_default_flavor},} create-dtb-package 14:37:35 ? 14:37:38 i never tried 14:38:16 ketas: this should be working fine, there's already a .ORDER to run stage-packages (which installs dtb) before create-packages. i need to leave in a minute but i'll see if i can reproduce it later 14:38:45 hmm 14:38:54 yeah 14:39:46 "i also noticed "buildworld buildkernel installworld installkernel distribution" installs world before building kerne" 14:39:52 cursed anyway 14:40:37 maybe i should check how much make changes 14:40:43 does it ever change? 14:41:06 make itself does that? 14:41:10 :/ 15:04:25 any ideas why "pkg update" inside 13.5 jail on a 14.3 host would Segfault 15:04:54 I thought as long as your host is newer kernel than the jail there should be "No Problems"(TM) 15:05:39 does pkg-static update also segfault? 15:26:17 anyone running wayland/x11 in qemu = 15:38:06 Anyone able to assist with this? My FreeBSD server can't hit jailed items from their public IPs. For example: I have traefik running in a jail and have all requests on 80 or 443 sent to it. From there, it distributes to other jails. I have keycloak running in a jail and gitea running in another. Keycloak is accessible from my local machine at sso.domain.tld, but gitea can't connect to 15:38:08 sso.domain.tld, and neither can the host machine. Any ideas on how I might resolve this? 15:40:55 check the resolver from within the jail first 15:42:08 divlamir: resolves fine 16:16:52 phlux: and they're different public IPs? 16:33:36 ah I see. Inside jail: "pkg update -f" no problem 16:34:15 Outside jail (14.3 host): sudo pkg --jail 135 update -f .....Fetching data.pkg, Processing entries: 100%, Child process pid=53344 terminated abnormally: Segmentation fault 16:34:29 odd 16:36:00 ohh 16:36:51 hmm 16:37:09 I have a 14.3 jail too, checking pkg on host --jail 143 update -f 16:37:27 pkg --version? 16:37:43 no crash. ok will wait for potential pkg upgrades in the future. 2.2.2 16:38:24 hmm, ok. i'm running the same and can't reproduce 16:38:33 what version of pkg is in the jail? 16:38:56 2.2.2 apparently it's 2.3.1 on 14.3 16:40:41 aic: might be some kind of 'old' package installed it cant quite solve for. do a spotcheck for old versions of python, etc 16:40:48 is it?? thanks divlamir I don't quite understand how that's possible because pkg.conf doesn't have versions it has $ABI 16:41:27 no, `the pkg --version` 16:42:04 you said 2.2.2, I thought you meant that 16:42:09 there's some problems I can see with how pkg does -j, but none that you'd normally run into with a normal `update` 16:42:36 e.g., plugins would seem like they'd cause problems, but you're probably not running ny 16:43:00 aic: pastebin.com pkg info 16:43:17 exactly yes, I am apparently stuck at "2.2.2" if 2.3.1 exists I just didn't understand that's possible when pkg configuration doesn't contain a version but references ${ABI} variable 16:43:29 pkg -j probably needs to actually attach to the jail and re-exec a version of pkg inside the jail 16:44:49 could be that there's some differences in the pkg databases between the versions... 16:45:23 pkg maintains backwards compatibility for things like that, especially with versions this close together 16:45:26 if I forgot to update a string in pkg.conf "pkg+https://pkg.FreeBSD.org/FreeBSD::13/quarterly" then I would know I made a mistake 16:45:47 kevans: that seems a sensible approach, delegate to jail's pkg 16:46:20 except in some very stripped down jail without pkg lol 16:49:13 pkg --debug update says: Request to fetch pkg+http://pkg.FreeBSD.org/FreeBSD:14:amd64/quarterly/meta.conf 16:49:21 so I am definitely getting the 14 package list back 16:49:38 but because it claims "2.2.2" is the "latest" version still. then my metadata is probably corrupt in some way, right? 16:50:05 if we could get a backtrace that'd be really helpful 16:50:38 I'll get to the bottom of it, this is really newbie stuff and if you're not even just freebsd users but developers then I am even more embarrased 16:50:49 to waste your time 16:51:18 (it's not a waste of time to help investigate) 16:52:05 aic: 2.2.2 might latest for the quarterly branch. also: pkg info to pastebin.com 16:52:38 or |nc paste.fbsd.dev 9999 16:52:40 :-) 16:53:33 uh, pkg is not a port how does it update itself actually? 16:53:40 pkg is a port 16:53:44 oh 16:53:57 there's pkg(7) in base and pkg(8) in ports 16:54:18 pkg(7) fetches pkg(8) if it's not installed, then execs it 16:55:51 why then: pkg which /usr/sbin/pkg --> /usr/sbin/pkg was not found in the database 16:55:57 because that's pkg(7) 16:56:05 /usr/local/sbin/pkg is pkg(8) 16:56:57 the shim in base only knows how to fetch the real pkg and verify signatures 16:57:39 divlamir: pkg(7) isnt a port is the tl;dr here. only pkg(8) is 16:57:57 oh, ok, right. got a bit confused here 16:58:23 it's ok, someone put some pkg in your pkg for pkg purposes 16:58:28 pkg. 16:58:56 i wrote a function in the pkg shim in base, pkg_is_pkg_pkg, that I was quite proud of 16:59:11 love it 17:00:04 owo 17:00:27 woo? 17:01:40 kevans: are you also the one who named "make package-pkg"? 17:01:54 nope 17:02:11 if it were me, that would've been `pkg-pkg` 17:02:48 name humor 17:03:31 looks like gjb is package-pkg 17:03:47 pkg-pkger 17:07:16 packeption! 17:08:21 how many packages can pkg pack if pkg can package packages 17:08:53 is it pkg or pkg pkg, or bird or plane 17:09:28 help, I'm being packaged 17:13:46 you sickos, this convo could make a doc PR for pkg_is_pkg_pkg XD 17:19:51 phlux: does it work if you make an entry in /etc/hosts pointing directly to your internal jail ip 17:20:40 tcpdump the connection attempt, see what happens 17:22:38 phlux: vnet jail(s) involved? 17:22:40 i was wondering if the phlux thing was a DNS issue. I had that issue with getting from one jail to another jail's gitea instance. I had to git clone ssh://internal_loopback_IP/project.git, or whatever. I couldn't use the normal hostname resolution. Overriding it in /etc/hosts didn't work 17:26:40 hi everyone, this is my first time using freebsd, and i am trying to install vscode, what i did: move to /usr/ports/editors/vscode .. and type "make install clean", but for now, the system are compilling and download a lot of another things, it's normal ? 17:27:18 s1lversurfer: yes, those are vscode's dependencies 17:27:34 wow, thank u, it's taking a lot of time hehe 17:28:26 You can use pkg instead of ports to download and install pre-built packages instead of building them. 17:28:49 yeah, but vscode wans't available on pkg search 17:30:35 but for while, i am enjoying the OS, i'am a linux user since 2004, but BSD it's very f good \m/ 17:31:55 pkg search vscode --> vscode-1.104.1 17:32:27 also `make install-missing-packages` could save a lot of time 17:33:02 divlamir: s1lversurfer is probably using quarterly, vscode is missing https://www.freshports.org/editors/vscode/ 17:33:41 yeap, only on latest 14/15/16 according to freshports 17:33:49 https://www.freshports.org/editors/vscode/ 17:34:28 oh sorry, need to thibk about getting glasses maybe 17:36:20 wow i came back, lost connection.. but i didn't found the vscode packet on pkg... someone talks about quarter 17:37:36 i thought we went over this, the electron package is build is broken on 14, which breaks vscode 17:37:44 it should exect for 15 17:38:55 rtprio: understoood 17:39:59 ty guys for help, i'll wait the system build vscode, and i have to install touchpad driver too, is not working 17:40:45 you are gonna wait looong time, switch to "latest" pkg branch methinks 17:42:15 oh, it finally built 17:42:30 the last dude who asked, it was broken at the time 17:43:42 the community that helps with code of freebsd have few people ? 17:43:55 and if it's really first time using freebsd, starting with a mix of ports and packages will end up with some curious errors some day 17:44:55 i don't program in C, but i would love to help the project 17:45:28 divlamir: i've downloaded the larger iso 17:46:27 it doesn't matter how you installed, pkg defaults to the "quarterly" branch 17:46:52 s1lversurfer: ports always needs more people and you don't need to know C, but it's pretty tedious work 17:47:47 so i did halfassed boot and now my armv7 system is in translation fault panic loop eh 17:48:10 s1lversurfer: also docs is very short handed right now 17:48:11 eh 17:48:39 wow, serius ? i program in Rust, Perl and Python.. i think that i need some extra work hehe i have no project in mind 17:48:51 if i can help with that.. 17:49:27 also, something useful anyone can help with right now is test the 15.0 prereleases and report issues 17:50:01 to report issues i use the mailling list ? 17:50:30 there's bugs.freebsd.org 17:50:34 mailing list or bugzilla 17:51:37 nice! 17:51:37 unfortunately going back to the previous comment, there are no docs for the major new feature in 15.0 yet :-d 17:51:55 that is? 17:52:05 divlamir: pkgbase 17:52:13 lol, my nightmare 17:52:50 maybe it's my nightmare, _because_ of the lack of docs exactly 17:53:47 i think the code is mostly done now so i am probably going to be writing that this weekend 17:57:53 thanks, i was thinking to stick to freebsd-update until the whole pkgbase thing makes it way through my skull 17:58:49 but it's the end of the road for freebsd-update as of 16, right? 18:00:47 yes, it'll only be supported for 15 18:00:56 you can keep using installworld though 18:01:56 yes, that will never die :) 18:03:40 we will probably switch it to use the pkgbase scaffolding at some point but that won't affect the user-facing behaviour 18:09:03 though, I can't see how a regular user could help with that sort of documentation. Without being familiar with the internals I mean 18:39:36 FreeBSD 15 is still progressing in ALPHA stages 18:45:18 divlamir: documentation like that practically requires a lot of input and/or editing from regular users 18:45:55 because documentation written by people familiar with the topic very often misses points they consider obvious 19:40:05 It also needs, ideally, input from people familiar with doc-writing. 19:40:46 but in an ideal situation we wouldn't be able to release a beta without any docs 19:42:27 AI will help 19:45:52 hey guys, im here again with a problem durnig compilation of vscode hehe, the vscode ports look for a perl5.40.2 but the freebsd 14 contains in his port folder a perl5.40.3 19:46:47 and now, im trying a something very newbie hehe, i just exec a mv /usr/local/bin/perl5.40.3 to perl5.40.2 lol.... and it works 20:11:24 Number of packages to be installed: 6 20:11:24 Number of packages to be upgraded: 432 20:11:25 The operation will free 4 MiB. 20:11:31 these new anti-packages with a negative size are great 20:11:47 wow 20:11:47 if we could add one that's -800MB or so, you could install freebsd without using any disk space at all 20:15:45 800mb from 432 packages to be upgraded ? 20:34:08 ivy: heh 20:45:18 anyone using 15 alphas? 20:46:28 FreeBSD freebsd15.le-fay.org 15.0-ALPHA4 FreeBSD 15.0-ALPHA4 stable/15-n280536-15d367160d00 GENERIC amd64 20:46:39 noice 20:46:53 anyone know what that means? "===> Compilation failed unexpectedly. Try to set MAKE_JOBS_UNSAFE=yes and rebuild before reporting the failure to the maintainer. *** Error code 1 20:47:05 did you install from alpha4 media, or upgrade from earlier alphas? 20:47:29 s1lversurfer: not sure. what are you building 20:47:48 i am building the "electron34", to compile vscode 20:48:27 but im trying again, after a "make clean" 20:48:30 jaredj: this is my stable/15 dev vm, it started on 15.0-CURRENT and just updated from source since then 20:48:44 mm ic. 20:49:19 do it like that s1lversurfer it is probably equivalent of make -j1 20:49:24 but, i do install 15.0-ALPHA from media several times a day, if you have any questions 20:49:36 i installed alpha3 from media with pkgbase and now i'm trying to upgrade to alpha4 20:49:39 then if it fails to compile again, it will give you a cleaner explanation of why it failed 20:49:48 ivy: hmmmm ty 20:49:49 first i have to find the alpha handbook 20:50:02 i found it and lost it again 20:50:24 aic: not exactly the same as -j1, because -j1 disabled make's compatibility mode even though 1 job is the default, MAKE_JOBS_UNSAFE doesn't pass -j at all. this matters for some ports like vim that depend on compatibility mode 20:51:08 jaredj: "pkg update; pkg upgrade" should work. there isn't much/any actual documentation on this aside from freebsd-base(7) 20:51:23 then it sounds like your compilation log s1lversurfer should already have a pretty clear error message where it failed and why 20:52:19 aic: after make clean and make install clean it's working per now.. i will try again 20:52:42 ok so now it compiled and no error? never mind then 20:54:08 you getting close to that vscode working? how many more dependencies to compile? :) 20:54:40 ivy: it doesn't *complain*, but it also doesn't move me to alpha4. also i was interested to find that i couldn't install a new base package because pkg couldn't chflags 20:55:18 vscode need a lot of dependencies hehehe but i was getting an error with Perl5 that i solve with a newbie trick, and now was stoping on electron34 compiling proccess 20:55:34 jaredj: how did you determine it didn't upgrade to alpha4? 20:55:57 let me test on a system that uses pkg.f.o 20:56:05 it didn't install anything new and my uname still says ALPHA3 20:56:38 ultimately remember that you are on the quarterly updates branch so if you can't compile all dependencies you can always switch to the latest updates, then install vscode from pkg 20:56:41 IGNORE ME 20:57:43 s1lversurfer: here is an explanation between "quarterly updates" and "latest updates": https://www.siberoloji.com/how-to-enable-the-latest-package-repository-branch-on-freebsd-operating-system/ 20:58:14 i ran it in the wrong place. i'm trying to build a router in a jail, so handing my network interface into the jail, and i must have run pkg inside the jail 20:58:23 jaredj: i shall ignore you if you insist, but fwiw, i just tested on a system running 15.0-PRERELEASE and pkg upgrade gives me 15.0-ALPHA4 20:58:30 ah, that would explain it 20:59:26 "[228/367] Fetching FreeBSD-clang-dbg..." etc etc 21:00:20 aic: oh tyyy 21:00:25 oo how exciting this is my first pkg-base upgrade 21:00:51 it's only fetching hundreds of things instead of tens of thousands :D 21:00:52 note that you're not actually getting ALPHA4, you're getting 15.0-STABLE, which is a little newer than the ALPHA4 media 21:01:14 but uname will say ALPHA4 since that's what it does during the release cycle 21:01:38 oh. ?? 21:02:00 the alpha releases are not really "releases", they're just snapshots of stable/15 at a particular time 21:02:03 the first thing i looked for was the repo config, and the only versiony thing i saw in there was ${ABI} (literally) 21:02:31 mmkay, and so the pkg-base repos are getting rebuilt against stable all the time? 21:03:01 currently, yes... for reasons... this will change in BETA1 but the new infrastructure isn't ready yet, which is why we're getting ALPHA5 instead of BETA1 21:03:36 once BETA1 is out you won't get any updates until BETA2 is out, then pkg upgrade will give you BETA2 21:03:45 makes sense 21:04:43 that's no problem certainly, i expect alphas to be gooshier, i just wanted to understand. oh heck i didn't make a boot environment 21:05:39 am i asking faq's here? i don't want to waste anyone's time 21:06:05 no, there are probably like 5 people in the entire world who either know or care about this right now. see above re: lack of documentation... 21:06:41 aw. WELL I CARE o/ and i'm glad you do too 21:10:32 woohoo! my crackpot system configuration manager didn't completely crash! 21:14:55 still a bit of a conundrum though. i have one rule that says "install doas," and more rules that say "hand the network interface into the jail thus becoming entirely dependent on it for networking," and that's before the said jail exists to be depended upon 21:17:19 gotta bring occambsd back into the picture. that'll fix it 21:18:55 unfortunately i got the same error: ===> Compilation failed unexpectedly. Try to set MAKE_JOBS_UNSAFE=yes and rebuild before reporting the failure to the maintainer. *** Error code 1 Stop. make[2]: stopped in /usr/ports/devel/electron34 *** Error code 1 21:22:03 well, either the port is broken or electron is 21:22:35 yeah.. i tryed to set the MAKE_JOBS_UNSAFE=yes.. a last try, if didn't work i will use Sublime Text, it's ok 21:23:12 you're not building with `make -j 8` or anything? 21:23:17 i will not die just for vscode, i a programmer since 2004, and his even exists hehehe 21:23:24 if you're really desparate for vscode you could try running it under linux emulation 21:23:51 rtprio: no no 21:24:24 hehe it's fine, i was a sublime text user, and use vscode only for the last year 21:25:04 i think we also have ports for some of the JetBrains IDEs, if that's of any interest 21:25:19 last time i tried CLion it didn't actually work though 21:27:52 for me it's fine, i even work wih Frameworks, except Flask in Python, i am cave man programming hehehe Sublime Text will help a lots :) 21:34:40 s1lversurfer: just wondering do you have the last several lines of output handy 21:34:45 where there's the compiler error 21:36:17 rtprio: the compiler is running again, if got error i will send it here alright ? :) ty 21:40:54 * kevans hacks on codebases of all shapes and sizes in vim only 21:42:34 emacs rep-a-zent! 21:43:24 kevans: for large C++ codebases, i miss live clang-tidy in vim and have never found a version of this that seems usable 21:43:47 wooy, my head is swimmy. foods time 21:44:08 i always heard emacs was the only thing anybody ever needed 21:44:20 * Macer walks backwards into the bushes 21:45:08 (clang-tidy include the static analyzer, so it's quite helpful to get those while writing code) 21:46:11 'miss' as in you can't do it anymore, or that you prefer other !vim solutions because they have other benefits that outweigh the lack? 21:47:19 i can't do it anymore, since i haven't found any implementation i like for vim. although if you have suggestions, i'm interested, especially if they work on src 21:47:49 oh, ok, I misread that the other way around: you had a solution in vim at one point 21:48:01 no suggestions here, i don't do anything fancy 21:48:26 my .vimrc uses autocmd to configure different style for different code-bases, and that's about as fancy as i get 21:49:28 yeah, i just have the style(9) stuff and a few sets for vim 21:50:11 i've recently started executing commands from within vim sometimes, to save myself from another tmux pane 21:50:51 i tried to start using :make at one point and didn't like it, but i don't remember why. maybe i should try again 21:51:25 i mostly just use vim as vi(1) with better indenting though, i ignore most of the other fancy features 21:51:36 yeah, see, we're pretty much on the same page here 21:52:22 I ended up working around my networking issue from Jail -> Jail by just spinning up haproxy inside the jail that needed to reach out and adding a /etc/hosts entry...was the only way I could figure out how to properly forward gitea to keycloak 21:55:40 kevans: fwiw, i found tmux a lot more usable once i found ctrl-left/right to switch tmux windows and cmd-left/right to switch terminal tabs 21:55:47 s/found/bound 21:56:05 i guess the advantage of :make is it can transport you to the next error 21:58:14 i don't actually mind it all that much, just wanted a little more width; i normally operate with a pretty trivial split window, top pane is vim and bottom pane is `make check` or `kyua test` or whatever output, so the third pane for `make all install` naturally limits one of them 21:58:30 well, sometimes both panes are vim 21:58:35 oh that's fancy 21:58:50 i just use three windows, 1 is a root shell for pkg upgrade, 2 is vim, 3 is build.sh 23:10:27 s1lversurfer: paste the actual clang or gcc errors on www.paste.rs nobody can tell what the problem is from only: "Compilation failed unexpectedly" 23:47:13 ivy: i use one window, a tmux session with irssi and 13 other windows