01:56:27 well freebsd never claimed to be a desktop os and not really what it is made for 01:56:33 bit rich coming from #linux 01:59:32 bombuzal: I do not understand 01:59:55 As I said, a serverside os 02:00:27 the fact there is thunderbolt support at all is more suprising 02:01:24 Unlike GNU/Linux, which was written for desktop use - like BeOS.... 02:03:30 agree 02:05:15 lol 02:05:21 stupid 02:05:58 bombuzal: not much of a people person eh? 02:06:17 thats ok 02:21:28 bombuzal: You are not quite right. First, there's "Chapter 8. Desktop Environments" of the Handbook. Second: https://github.com/FreeBSDFoundation/proj-laptop 02:22:38 * checkpoint uses FreeBSD on laptops without any claims ;) 02:23:08 well for all practical purpose - he is right 02:25:06 well, ahtough there was not any official claim (how that claim would it look like?), it just works on desktops and laptops very well. 02:25:38 yes I agree. And then there are the areas, where linux also took a long time to get it right 02:25:39 * checkpoint recently installed FreeBSD on old MacBook, which brought it to life. 02:26:07 and linux is maybe not even there completely yet 02:26:40 I know I have not have much success with running Freebsd On my monitors via thunderbolt 02:26:53 on my laptop directly - works 100% 02:27:22 * checkpoint thinks Linux is going towards nowhere... Linux guys are trying all possible directions at once, which makes it a mess. 02:28:05 it does make it a mess - it also grows innovation 02:28:15 but it is not pretty 02:29:52 what's good about being sencond on the run is that you can learn from those how are running first and try to avoind making same mistakes. 02:30:05 absolutely 02:30:13 checkpoint: it was BarnabasDK saying FreeBSD wasn't for desktops/laptops 02:30:19 I was being sarcastic 02:30:42 bombuzal: oops, sorry. 02:30:42 bombuzal: well in that case so was I of cause .. 02:31:20 :3 02:31:32 * bombuzal hugs checkpoint and BarnabasDK 02:32:15 one thing I do not understand is the drive to introduce non standardized languages into the kernel in linux 02:32:23 * checkpoint thinks that Linux is kind of doomed, their intention to rewrite the whole world in Rust will kill Linux. 02:32:27 so many applications REQUIRE that 02:32:29 lol 02:32:35 death to Rust 02:32:42 yeah! 02:32:43 :) 02:32:44 automotive, space, military etc etc etc 02:33:11 I actually like rust - the rust people - not so much 02:33:41 BarnabasDK: there are special purpose OSes for that case, I cannot see why would they need Linux on flight control. 02:34:31 saving money? 02:34:53 what OS do a Tesla use? I have no idea 02:35:09 ElonOS 02:35:10 besides, Rust itself does not solve any issues, esp bad coding attitude and logical errors. 02:35:46 MaskOS X ? 02:36:30 checkpoint: no, you still have to use your head, but some of the traditional memory allocation issues are solved - at the price of more memory usage ofc 02:38:17 Okay -- I don't code, but if I did -- I would probably consider Zig 02:38:18 * checkpoint has never tried Rust, its syntax is too complex for old fart like him ;) 02:38:22 more memory usage? how? 02:38:23 and some of the rust re implementations just make no sense what so ever. Like re implementing UUtils 02:38:44 hodapp: because of the default immutable model 02:39:29 hodapp: exactly, more $$$ to DDR5 producers 02:41:10 I mean - its a tradeoff isn't it 02:41:29 the main problem is that there still very few Rust programmers, who till support all that stuff they are trying to rewrite ? 02:42:05 and to what standard was it written? whatever the git state of the date was 02:43:10 but yeah, code and dissapear 02:43:21 leave the rubble to someone else 02:44:00 lets take UUtils, it must be the best tested set of tools at all 02:44:11 in 2026 02:45:04 I cannot possibly see the idea in a re implementation just to have it in your "favorite" language 02:45:40 there's some reason though - get money from US gov/mil 02:46:11 so just to cash a government check? 02:46:14 those are obsessed with proven code and proven languages 02:46:23 then why rust 02:46:37 unproven, and non standardized 02:46:50 and yes I still like it 02:47:06 but as opposed to c/c++ not a comparizon here 02:48:08 from what I know, gov/mil guys love Ada and that soft of stuff, Rust is pretty close 02:48:49 but seriously - I had the idea that linux was deep inside automotive, aerospace and medical etc 02:49:21 maybe I was mistaken 02:49:54 Linux is in a lot of places including embedded. But FreeBSD is actually in a lot of places used by companies too 02:50:03 I worked with Ada and that is DEFINATLY standardized up its wazoo 02:50:24 It has the advantage of the MIT license, where these companies can take FreeBSD do whatever they want, and not have to give anything back at all, even recognition 02:51:03 I realize the use of FreeBSD, I am one of the ones that do 02:51:19 famously - network devices 02:51:20 actually not, Linux is heavily used in appliances like car radios etc. all the real stuff (ECI) work on RTOses and a like. memory shortage and execution time prediction is still an issue. 02:51:38 As to why Rust, and not Ada, Zig, Go, or something else -- No idea, Rust has a cult like following though 02:52:17 SponiX: yeah, I do not understand the cult thing, eventhough I like the language 02:52:29 it is just another programming lang 02:53:11 * checkpoint likes pure C and maybe golang 02:53:14 I agree that encouraging memory safe code for security purposes is important. But I don't think Rust is the only way to do this 02:53:28 checkpoint: I like c too, thats where I started 02:54:29 * checkpoint started with assembly long time ago, but C compiler is just a good macro assembler :) 02:54:57 yeah, I actually started with lattice c on the mc68000 02:55:16 amiga 02:55:28 before that - on the c64 those things made little sense 02:56:10 it supported amigaos quite well in fact 02:57:37 or I suppose - workbench 02:57:44 actually my first lang was Fortran on big iron in late 80th, but I lost interest in big iron pretty soon and switched to micros (Atari, BK-0010) 02:58:34 yeah, for me it was the other way around, I had to move backwards to make money, so I started coding cobol on VAX / acms decset for LEGO 02:59:11 now there is a truely terrible language 03:00:42 must be up there with php or perl 03:01:26 at the place I worked in 80th PL/I was quite popular 03:01:30 and since I realized we might as well code in pascal or c in acms, it just made it the more bitter 03:01:59 * checkpoint still uses Perl for web stuff 03:02:12 perl is ok if you have to hack stuff 03:02:14 and for some automation 03:02:19 just dont build large systems around it 03:02:44 it is very fast at parsing large files, which was the original purpuse 03:02:59 I still don't like it 03:03:24 ooh, neat, have never really looked at Linuxulator and didn't realize it's pretty turn-key to get a Linux userspace in something like Debian on top of FreeBSD 03:03:41 * checkpoint has had experience building a number of large systems in Perl5 (ISP billing and Soft-PBX) 03:03:59 all this talk of webdev reminds me that I need to write a fastcgi server for something 03:04:00 hodapp: I think you can even make a linux jail, I haven't tried though 03:04:24 but the handbook says so 03:06:53 anyone aware what is the state for Docker in FreeBSD ? some ppls needs it badly 03:07:16 I do sort of miss the terminal days though, no matter the system. They where nigh on indestructable 03:07:49 checkpoint: docker and kubernetes tie close to the linux kernel afaik 03:08:41 BarnabasDK: yes, I know. there's docker package, just wonder what is it. :) 03:08:46 docker-18.09.5_30 Open-source application container engine 03:09:03 so close that microsoft did not re implement the handles in the windows core, but added a linux sub system, which is really a linux core booted in a virtual machine afaik 03:11:12 * checkpoint recalls in days of big iron all OSes and event distinct programs had to be run in VMs (at least on IBM S/370) 03:12:35 but where they vms? 03:12:41 or just partitions 03:13:27 HP-UX had a great system for dividing expensive hardware 03:13:35 now it is non supported 03:14:45 BarnabasDK: the underlying system was a sort of hypervisor, an operator starts a many VMs as programs he/she wants to run 03:14:59 ic 03:15:48 but within the same os and arch though? 03:15:59 BarnabasDK, first VMs (370 onward) if you were runing the hypervisor, then partitions (LPARs) IIRC from ES/9000 onwards. 03:16:09 for me Docker looks pretty similar to what we had on S/370 03:16:09 sweet 03:16:33 you know much more about this stage of computing than me 03:17:06 * V_PauAmma_V cut his mouthpalps on APL under VM/CMS. 03:17:27 but still same os and etc 03:17:34 mouthpalps 03:17:45 you could not run a bsd unix in a partition for argument 03:18:24 I never really understood why we needed to do hw level abstraction tbh 03:18:35 in production 03:18:43 for development - yes ofc 03:19:58 LXGHTNXNG, yes, mouthpalps. I'm a giant crab, as my name would indicate. 03:20:25 I was aware of that. 03:20:38 * checkpoint thinks it's bizarre how bad solutions tend to repeat again and again (VMs, now Docker) 03:20:53 checkpoint: yes 03:21:13 seems to be a 7 year cycle 03:21:15 ... are jails a bad solution (being essentially lean VMs)? 03:21:39 they do not emulate hw 03:21:50 does Docker? 03:22:12 no, but I like that too 03:22:29 I am thinking of the vmware stuff 03:22:33 see also: vnet jails, which create separate network stacks, thus emulating network hardware 03:23:19 jep 03:23:23 I know 03:23:28 is there something close to vnet in Linux, btw ? 03:23:33 netns 03:23:41 yeah something like that 03:23:51 but jails do not let you run a macosx in a sub system 03:24:28 because jails are not a hw emulation layer 03:24:52 I think that entire idea was wrong in production 03:24:58 bhyve ? 03:25:31 * checkpoint still has not tried bhyve 03:25:34 bhyve is a alternative to wmware isnt it 03:25:43 yes 03:25:49 rather KVM 03:25:50 bhyve fits into the same class as kvm 03:25:56 Jails just let you virtualize the FreeBSD kernel. Linux can do it too, with namespaces. 03:26:03 Bhyve is qemu, vmm is kvm 03:26:34 the bhyve binary compares to the qemu binary, the vmm kmod compares to kvm 03:26:46 yeah, like I run kvm here on my linux lappie in order to have a windows 10 03:28:07 * checkpoint uses VirtualBox occasionally for Windoze stuff 03:29:42 I think I have a win 10, a freebsd and a haiku os pt on my lappie 03:29:54 deployed via zfs 03:30:12 If bhyve could do intel igpu sr-iov, I would use it yesterday 03:30:20 I've used it before at work 03:31:20 bhyve cannot share GPUs ? 03:31:49 It can passthrough the entire device, but SR-IOV lets you create multiple virtual instances of a single GPU and pass them through independently 03:31:56 And the host can maintain its own device 03:32:24 amigan: got it, thanks. 03:32:26 FreeBSD can do SR-IOV for ethernet controllers but it does not seem that i915/xe are supported 03:32:58 Even on linux you have to use a janky out of tree kernel module 03:33:44 yeah 03:34:25 hw virt has stil not ever brought anything to me in production, for local dev yes 03:34:45 I know a lot of my ms friends disagree 100% 03:35:52 why not just build the sw to your platform? in place 03:52:09 My use case is a windows VM, it's the best way to get a desktop on your phone for quick stuff 03:52:16 RDP is unmatched 03:52:44 I daily linux but this is one place windows shines in 2025 03:53:08 FreeBSD on all the network and app infra, of course 04:02:33 * checkpoint proclaims 2026 to be year of FreeBSD on laptops 04:03:03 * checkpoint has some more laptops to upgrade to FreeBSD 04:08:22 ix 04:08:27 ic 04:09:17 o\ 09:22:56 Happy new year :) 13:54:47 hny 13:55:06 i think this weekend i'm going to really sit back and think about the network setup i have going on here heh 13:55:53 i have an old xeon 5560 1u that i think i'll use as a router just to have the 10gbit 15:53:28 * rtyler waves 16:34:00 Quick question about a.out(5) file format. It's documented that the struct is based on `unsigned long` ... does that mean that on 64-bit, the file's header is thus larger? (because of LP64) 16:34:14 Or is the manpage just very old, and I should read the fields as `uint32_t`? 16:36:27 The manual page is indeed old. As in, here for historical purposes only, since a.out format support went away in FreeBSD 11 (IIRC). 16:37:13 Yeah, I am aware. I need this for mostly-historical reasons, too ^^ 16:38:45 (though I'm also considering ste^H^H^H adopting it for my own purposes --- not intended to be natively executed mind, it's just that using an existing format as a base saves me on a lot of documentation, assuming it's a good fit ofc) 16:38:46 DarkUranium: the manpage is wrong, look sys 16:39:15 `uint32_t`, I see; thanks! 16:41:43 it may be academic since i don't think a.out was ever used on a 64-bit platform 16:42:12 True. 17:14:50 we were builidng and running a.out binaries in the past, whole FreeBSD OS was a.out-based 17:16:17 we didn't care at that time, at least me 17:16:31 yes, but ELF replaced a.out in 3.0, and 64-bit platforms didn't arrive until 5.0, so i don't know if an a.out format was ever defined for a 64-bit platform 17:17:04 actually maybe Alpha was 4.0 17:48:37 I think my server is compromised 17:49:55 wsky_ what makes you to think that? 17:50:54 Because whenever I connect a machine to the server’s vpn it gets compromised 17:51:04 British services are after me 17:51:07 Im screwed 17:51:15 I see... 17:51:46 British and polish services