00:00:02 but it doesn't ask "do you have a commit bit?" so hard to know how they'd interpret the answers 00:00:14 maybe check if people put an @freebsd.org email address in the contact bit :-D 00:00:19 yeah 00:00:25 my thought exactly lol 00:00:47 jbo: so we shall see.. it did pretty well.. no more raspbery pi server.. will havve an offisite backup and plenty of power to try things 00:01:20 then again, if i was committer and wanted to complain, i probably wouldn't put a freebsd.org email address in there... 00:01:42 i have not built a computer since 1999.. back in the day when pentium slot cartridges were the bees knees 00:02:15 and abit still existed. 00:07:29 my system dates back to 2012 00:08:53 "Tired of seeing ZFS get all of the attention, ext4 developers introduce a data corruption issue of their own." - emaste@ burn of the month 00:09:11 I just use UFS still 00:12:48 this is debian 00:13:16 freebsd is on 00:13:20 Why is my /sysinfo script so....non-verbose? 00:13:22 Client: HexChat 2.16.1 • OS: FreeBSD 14.0-RELEASE • Storage: 0 bytes / 0 bytes (0 bytes Free) 00:13:51 AumShivaya: impressive that you managed to fit the base system into 0 bytes 00:13:55 Client: HexChat 2.16.1 • OS: Debian trixie/sid • CPU: AMD FX(tm)-8350 Eight-Core Processor (4.11GHz) • Memory: Physical: 30.5 GiB Total (17.4 GiB Free) Swap: 954.1 MiB Total (913.1 MiB Free) • Storage: 735.1 GB / 1.9 TB (1.2 TB Free) • VGA: NVIDIA Corporation TU116 [GeForce GTX 1660 SUPER] @ Advanced Micro Devices, Inc. [AMD] Family 15h Processor Function 3 • Uptime: 1d 7h 29m 30s 00:14:29 :`( 00:14:36 mine does not even say uptime!!!! 00:14:39 ilythia ~> uname 00:14:39 FreeBSD 00:14:43 ? am i doing this right 00:14:55 00:14 -!- Irssi: Unknown command: sysinfo 00:15:04 oh sysinfo is a hexchat script 00:15:17 never heard of it 00:15:19 I think irssi has one too, tho 00:15:24 which is not on bsd 00:15:35 (hexchat is not a terminal irc client) 00:16:09 no they dont give you the sysinfo script 00:16:20 in hexchat 00:17:00 I think I may have had to compile it in as an option or install it somehow, but it still is not comprehensive/verbose 00:18:54 meena: hey did you see i got in the ports commit log! though it's funny, because the committed patch is completely different from the (one-line) patch i actually submitted... but still... it has my name on it, so it's going on my CV 00:20:19 unixwitch: I don't religiously follow ports commits, like I do src commits 00:25:15 Also bofh just pushed a million commits removing all expired ports 00:25:49 oh no, not xf86-video-sunffb! 00:25:59 how will i upgrade my SPARCstation 20 now 00:31:32 that channel is hopping 00:35:36 unixwitch: there's always NetBSD 00:37:16 pcbsd too 00:55:41 jbo: why does your blog pretend to need js? 00:57:59 nimaje, hmm, not sure. might be for image gallery or search crap. 00:58:08 nimaje, it's a hugo thingy 00:59:01 nimaje, I've never been happy with this theme but also didn't care too much. I welcome any suggestions. 01:00:11 unixwitch: congrats! 01:00:17 on the CV commit thingie 01:02:33 jbo: well, it has some white "preloader" as position: fixed over the content, so you can't read it without js removing it or using devtools to remove that shit 01:04:41 nimaje, that's stupid. I'll look for a different theme soon (tm) 01:09:00 well, lots of stupid broken stuff on the www, at least it doesn't need js to load static text from a web server and display that as other blogs do for some reason 01:13:51 jbo: https://themes.gohugo.io/themes/blank/ ? 01:13:53 Title: Blank | Hugo Themes 01:17:59 nimaje, a bit of simple non-bullshit CSS wouldn't hurt :) 01:18:39 nimaje, I do like that it has a search functionality, but that is just a simple index that hugo builds during generation and then a simple JS script on top of that which is not needed to actually read the blog - I am totally fine with that sort of thing 01:21:20 nimaje, I was looking at this theme the other day - does that work on your end without shenanigans? https://neonmirrors.net/ 01:21:21 Title: Neon Mirrors 01:23:13 that seems ok 01:35:53 unixwitch, you still up mate? 01:37:24 * meena will only put one of her many FreeBSD commits on her CV: https://github.com/freebsd/freebsd-src/pull/664 written on my phone, while driving in a car (passenger seat) 01:37:26 Title: apic: prevent divide by zero in CPU frequency init by igalic · Pull Request #664 · freebsd/freebsd-src · GitHub 01:47:32 jbo: a little bit 01:48:55 unixwitch, I'm "evaluating" sanoid/syncoid right now. Having a hard time finding proper documentation on the config file 01:49:12 unixwitch, is there anything other than the github wiki doc which properly explains all options in depth? 01:49:19 I got no man page either :( 01:49:55 jbo: You could take a whack at writing your own. 01:50:28 mason, are you suggesting I am writing my own documentation for a project I am currently evaluating for use? 01:50:46 jbo: No, I'm suggesting you write your own snapshot manager. 01:50:58 mason, that is where I am coming from. 01:52:23 jbo: i honestly can't remember what i read when setting it up but i think the config file was fairly well commented... maybe something on the github repository? 02:04:14 * unixwitch wonders what happened to window(1) 02:04:45 Oh, forgot all about that. 02:07:27 ah, moved to misc/window in 2009 02:08:21 unixwitch, does syncoid require/imply use of sudo? can't do doas? 02:08:58 jbo: i currently have it log in as root, so i'm not sure. does zfs allow work? 02:09:41 unixwitch, I setup zfs allow on the target host. then tried 'doas syncoid ...' from the source host and got this: 02:09:46 CRITICAL ERROR: ssh connection echo test failed for jbo⊙111 with exit code 255 at /usr/local/bin/syncoid line 1714. 02:10:26 if it's running as a dedicated user, does it need the ssh host key accepted or something like that? 02:11:12 ah shit, doas syncoid ofc makes my ssh key being invisible I guess (because then it does stuff as root?) 02:11:19 jargon libera is corrupt 02:11:21 on the source host that is 02:12:00 surprised it doesn't print a more useful error though... 02:12:24 well it does, I just didn't show you that for silly reasons :> 02:12:30 jbo⊙111: Permission denied (publickey). 02:12:30 jbo⊙111: Permission denied (publickey). 02:12:30 CRITICAL ERROR: ssh connection echo test failed for jbo⊙111 with exit code 255 at /usr/local/bin/syncoid line 1714. 02:12:34 ah 02:13:10 i don't use doas, but i guess it's wanting root to have the key instead of the user, or vice versa 02:13:13 libera is corrupt 02:15:24 unixwitch, syncoid seems to have a switch that might (?) help here? --no-privilege-elevation Bypass the root check, for use with ZFS permission delegation 02:15:39 not sure what "zfs permission delegation" is tho 02:15:48 jbo: man zfs-allow 02:16:07 excellent - that is what I already set up. 02:16:17 jbo: i think that means it won't check if it's root on the remote host, not related to logging in in the first place 02:16:32 > cannot create snapshots : permission denied I guess my zfs-allow didn't work :p 02:29:20 7.01KiB 0:00:00 [ 125KiB/s] [=====================================================================================================================================================================] 115% 02:29:33 49.7KiB 0:00:00 [4.87MiB/s] [=====================================================================================================================================================================] 105% 02:29:41 funky progress bar values 02:29:51 apparently I have more than 100% backup now. 02:36:31 /31/60 02:36:47 Is compression involved and different on the two ends? 02:52:55 reasonable question 04:31:05 hrm.. have to install NIC drivers.. realtek what a pain 04:31:19 at least freebsd 14 is operational, minus the nic card 07:15:25 Is there not any other way using an mDNS service to let hosts living the in the same LAN resolve and see each other based on their .local host names? 07:17:27 I used to set up an NSD on the router plus unbound on each local servers (physical, virtual, anything,) and whether or not via DHCP, tell hosts to use it for resolving the LAN zone. 07:17:47 But am thinking if there's any better way (haven't tried mDNS stuff, yet, btw.) 07:18:16 RFC6762 (Multicast DNS) defines .local 08:31:57 Is mDNS supported/enabled in FreeBSD by default, or I shall use a third-party port? 08:32:31 jbo: "if xmj is right" :-> 08:32:32 OK, it ain't. 09:36:49 mdns cool? 10:24:33 what's rpath? 10:25:55 polyex: a hardcoded runtime (shared) library path in an executable. it allows the executable to run even when its shared libraries aren't in the system library path 10:26:34 for example: "cc prog.c -o prog -L/opt/mylib -l mylib -Wl,-R/opt/mylib" allows prog to load mylib.so.whatever from /opt/mylib 10:27:34 kinda like dep injection? 10:28:09 not exactly, aiui dependency injection would be more like LD_PRELOAD, where you load a library that application doesn't link at all 10:29:14 k sweet ty 11:03:04 polyex: mDNS is convenient for certain things and you probably want it if you use Apple HomeKit devices or w/e it is called 11:03:16 ew apple 12:11:12 polyex: depends on the type of product; for smartphones I prefer Apple over anything Google-adjacent as I used Android phones since October 2011 and until February of 2022 12:11:45 and strongly rejected iPhones haha 12:35:42 how would I go about connecting airpods ? 13:34:33 Does FreeBSD have an equivalent of Linux's timerfd? 13:35:51 jemius: timerfd(2) 13:36:14 (added in 14.0) 13:36:18 Why couldn't the posix people just include that 13:37:38 wouldn't kqueue be the FreeBSD analog? 13:39:07 yeah, the native version is EVFILT_TIMER... i assumed the question was from porting a linux app but if not, kqueue would be the better approach 13:41:26 It seems to me that everyone basically agrees that timers through signals are a bad approach 13:41:50 though they're the only portable timers in posix, afaik 13:42:17 well, the alternative is to implement your own system system as a layer on top of kqueue/epoll/select/whatever, which is what most applications do (if they don't use a newer api) rather than signals 13:42:23 s/system system/timer system 13:49:16 I like "system system". I'll call my next system that ^_^ 13:50:13 unixwitch, you mean those use poll()'s timeout parameter and then calculate which timer fired and how the timeout has to be modified? 13:51:05 The thing is that you often want to monitor sockets and nevertheless do something periodically. If that something is more than one single thing, you basically only can use timerfd, poll()-timeout-magic or signals 13:51:36 Signals are troublesome because they could corrupt your state if the signal handler accesses something that the other poll() functions were in the process of modifying 13:51:58 jemius: yes, you keep some kind of sorted list of your upcoming timer events, then when calling poll() (or whatever) you set the timeout to fire when the timer is due. then, in the event loop, check for pending timers and dispatch them somehow 13:52:16 this is how libev does it, for example ... or at least it did last time i looked, i'd be surprised if it doesn't support at least timerfd nowadays 13:52:45 I mean, we can use timerfd ourselves, we don't need a lib for it 13:53:06 i know, i'm just mentioning another way to do this without timerfd or signals 13:53:16 hmm 13:53:23 * jemius is unhappy 13:54:29 Ah, screw it, I'll just use timerfd. If I have to port to Apple's systems one day I'll have to think of something. But that's a problem for future me 13:56:53 you know, i'll laugh if this all gets standardised in the upcoming C++ network extensions before there's a standard way to do it in POSIX 13:57:09 ... which is almost certainly going to happen because POSIX seem to have no interest in doing it 13:58:40 posix seems to be on vacation since 2008 13:58:42 jemius: doesn't Mac OS have kqueue 13:59:11 Remilia, no idea. I don't use it yet, but consider it likely that we'd need it (iOS) in a few years 13:59:22 oh iOS 13:59:24 it does yes, but i don't know if iOS does (although i see no reason it wouldn't) 13:59:46 iOS probably has its own higher level stuff for developers 14:00:28 yeah i actually don't know if you even get libc on iOS. i guess at least some parts must be included because it was ObjC before Swift came along 14:02:45 There are probably very few app developers who even do anything in low level languages. On Android there at least is NDK 14:06:21 so i got this machine 99% operational and now i am doing to the dang realtek nic card.. i see here: https://www.freshports.org/net/realtek-re-kmod that there is a module to get the NIC card operational. my question is: if i do a /usr/ports install of this do i then have to work about pkg updates "messing" with it? I ask as this NIC is not part of base install 14 14:06:22 Title: FreshPorts -- net/realtek-re-kmod: Kernel driver for Realtek PCIe Ethernet Controllers 14:07:21 voy4g3r2: are you using a non-GENERIC kernel? the pkg version should work the the GENERIC that the pkgs were built for 14:07:37 when i bought up the machine, the kernel does NOT find the card 14:08:00 but i could be going down the, i did not know something path. 14:08:26 https://forums.freebsd.org/threads/realtek-rtl-8125-2-5gbps-lan-controller.79710/ <-- this mentions what i am experiencing, so i am walking through it 14:08:26 well, if you're using pkgs anyway, i would start by just installing the pkg and following the instructions (/boot/loader.conf) and see if it works 14:08:27 Title: Realtek RTL 8125 2.5Gbps LAN controller. | The FreeBSD Forums 14:08:31 voy4g3r2: is your NIC not working with built-in re? 14:09:17 let me boot the beast up, but that is correct.. right now it is not loading the re module.. https://www.gigabyte.com/Motherboard/B760M-DS3H-DDR4-rev-10/sp#sp 14:09:19 Title: B760M DS3H DDR4 (rev. 1.0) Specification | Motherboard - GIGABYTE Global 14:09:37 wait, you mean you have a custom kernel? 14:10:16 doesn't GENERIC have interfaces built into the kernel 14:10:16 the instructions on that thread seem to be about installing the package, not installing it from ports, which is also how i would suggest doing it 14:10:17 i just did a base freebsd install, 14, on the box. it detect everything except the NIC 14:10:53 yeah.. then i head over to this: https://www.freshports.org/net/realtek-re-kmod which mentions a package.. but i have no url to get it.. that i can see 14:10:54 Title: FreshPorts -- net/realtek-re-kmod: Kernel driver for Realtek PCIe Ethernet Controllers 14:11:49 i am just surprised, that after 15+ years of NOT buildin a computer.. it started up first time and my only issue is this 14:11:52 does dmesg mention anything at all? 14:11:55 voy4g3r2: http://pkg.freebsd.org/FreeBSD:14:amd64/quarterly/All/realtek-re-kmod-198.00_3.pkg 14:12:47 fwiw, base re(4) is kind of broken and i would always prefer the pkg driver even if it is detected by base... there's some undocumented thing that at least some of the ICs need to work properly and the freebsd driver doesn't do it 14:13:00 huh 14:13:05 as expected from Realshit 14:13:15 yeah, it is interesting, i found multiple youtube videos on it 14:13:21 redefining the low-end 14:13:28 and they said.. your response is common Remilia :) 14:13:35 just buy an intel nic card they say 14:13:48 i went into microcenter expected to spend 300 i spent 850 14:13:51 so trying NOT to buy more stuff 14:14:05 i concur with that, Intel is good. or Mellanox or Chelsio 14:14:06 i did a search, in the store, and said realtek works with re module. 14:14:20 when I was choosing my bothermoard I specifically looked into those with Intel NICs 14:14:21 so i go, sweet, moving on 14:14:33 fwiw, i used realtek-re-kmod on one system here (before i put an Intel card in it) and it does work fine, i didn't have any issues. just one extra step when installing 14:15:00 tbh the best thing you can get is a quad port Intel PCIEx4 card 14:15:03 hrm.. can't find the re module 14:15:05 ports also needs working network to fetch the distfiles, so if you can use ports to install it you should be able to install it via the normal pkg install realtek-re-kmod too 14:15:06 * Remilia hides 14:15:33 haha 14:15:38 Remilia: yea that was a 300 buck card 14:15:43 it does look sexy 14:15:45 also 2.5 Gbps is generally dumb 14:15:54 go either 1 or 10 14:16:06 2.5Gbps is (sadly) gaining quite a bit of traction in the home market... which annoys me because i agree 14:16:12 2.5 means you are locked into 'gaming routers' for ethernet 14:16:23 yeah well it is staying 1 gig on this network 14:16:34 and dang nab it, my dumba** did not install ports when i did this install 14:16:39 i knew that was going to bite me in the a** 14:16:46 (i do think there's one valid use for 2.5G/5G, which is running wireless APs over existing Cat5e structural cabling... but that's pretty niche) 14:18:12 voy4g3r2: just copy the realtek-ke-kmod package i linked above and the pkg package onto a usb stick or something and install with pkg add (i think this works with /usr/sbin/pkg, right?) 14:18:39 doing the move now 14:18:56 the land of thumbdrives in freebsd partition format and apple formats.. no fat ones.. 14:19:06 pkg package: http://pkg.freebsd.org/FreeBSD:14:amd64/quarterly/All/pkg-1.20.9.pkg 14:19:07 at least the usb enclosures are gone from my network now :) 14:20:06 what did you replace those with? a new motherboard with more ports? 14:20:29 a whole new system.. 14:20:40 ah 14:20:46 hope that works out better 14:20:50 me too 14:20:51 (well, aside from this issue) 14:21:07 and i can then have the offiste backup complete as that is also sata stuff 14:21:46 i was more worried i was going to mess up the computer build 14:21:59 and i swear that spring for the cpu.. that thing hurts if your fingers are in its way 14:22:12 i was expecting a small "pop" nope.. damn thing flings open.. fast and hard 14:23:50 voy4g3r2, if you care: my current home NAS is a HPE MicroServer Gen10+ running 14.0-RELEASE 14:24:09 I wish it would have more bays but whatever - I guess raidz2 is good enough 14:24:15 heh, i used to have a microserver years ago... probably one of the first generations. nice that they're still around 14:24:30 that was the biggest struggle.. find a machine that had enough bays 14:24:42 this one has 6 2 2.5" and 4 3.5" 14:24:58 unixwitch, glad to hear that I got something that was already reasonable years ago :) 14:25:02 this board even has nvm.e so in theory i could eliminate the 2.5" 14:25:07 voy4g3r2, what did you end up with? got a link? 14:25:15 custom build 14:25:29 ah, so just shopping for chassis and then whatever fits inside of it? :D 14:25:35 or did you get a barebone? 14:25:43 i use a Corsair Obsidian 750D for our file server... it's not too expensive since it's a desktop/gaming case but you can put 9 (!) 3.5" disks in it plus a bunch of SSDs 14:25:44 gigabyte motherboard, 32 gig of ram, 3 4tb toshibas, 1 2.5" generic some case.. 14:25:52 jbo: i went around and found parts :) 14:26:53 I was actually looking around to upgrade my current workstation. it's an i7-8086 with a quadro p5000 and 64 GB DDR4 RAM. I got that machine in 2018 I think. 14:27:08 anyway, when I was looking for parts online it seems like the only choice these days is gaming parts 14:27:14 oh and an i3 3.4ghz quad core 14:27:23 everything hast stupid RGB and probably costs 50.- more because the PCB is black 14:27:28 not sure your location but i went here: www.microcenter.com 14:27:37 there is a store by philadelphia area 14:27:48 jbo: have a look at Fractal, they make well designed cases that don't look like gaming bling and don't have RGB or transparent sides 14:27:53 you find a sales person, they walk around ith you and confirm compabitibility 14:28:11 an old timer is like.. yeah we do nto have prebuilts with bays of drives.. the rage is nvm.e.. you gotta go custom 14:28:13 unixwitch, my non-rack cases are all fractal r5 :) 14:28:35 he did laugh when i said, last time i built a machine the pentium slot was the cpu rage of the day 14:28:46 unixwitch, I was mostly upset by mainboards I guess 14:28:47 i miss those slots 14:28:49 p2 400 on an abit bx6v2 board 14:28:58 inwin q500 case 14:29:11 that was my high school graduation gift.. that machine lasted till 2008 14:29:14 i used to have a DEC system with the CPU on a slotted daughterboard... you could switch between i386 and Alpha by just replacing the cpu card 14:29:29 damn you exfat partition 14:30:27 unixwitch, I am also very much on the fence regarding intel vs. AMD. I never had an AMD system before but ryzen 9 7900 looks... reasonable 14:30:34 jbo: ah yeah. my desktop motherboard has RGB but fortunately i can't see it. i did manage to find a decent ASRock B450 board with no RGB on it for another system 14:31:08 asrock is chinese... I don't feel like trusting their firmware "currently" tbh. 14:31:27 oh wait, they are taiwanese 14:31:44 i actually like this ASRock board, i'd consider them if i end up doing another build 14:32:02 I have an asrock board in another build. it seemed like great price/value back in the day 14:32:11 it generates fewer acpi(4) errors than my MSI board which i take as a good sign :-) 14:32:17 unixwitch, any opinions on the intel vs AMD story? 14:32:49 the exact cpu is core i3-13100 raptor lake 3.4ghz quad-core lga 1700 14:32:56 jbo: not really. i've been using AMD since the Athlon MP because the price/performance always seemed better at the times i needed to buy something. my only complain is they are pretty starved for PCIe lanes, i think Intel might be better there 14:33:29 (it doesn't help that consumer motherboards distribute the lanes as like, 2 x16 slots and 4 x1 slots, and no actually useful x4 slots) 14:33:51 yeah, I was actually considering just getting a supermicro board for my next desktop 14:34:06 but they don't seem to have AMD socket offerings which made me worry about my potential pick of AMD for the next machine 14:34:08 i noticed the supermicro are doing AM5 now, not sure if that's for desktop too or just server 14:34:19 they do? 14:34:24 https://www.supermicro.com/en/products/motherboard/h13sae-mf 14:34:25 Title: H13SAE-MF | Motherboards | Products | Supermicro 14:34:41 aye 14:34:49 only noticed that board because someone linked it in another channel, i don't know if they have anything more suited to a desktop 14:35:11 https://www.supermicro.com/en/products/motherboards/desktop-gaming-boards 14:35:13 q___q 14:35:15 Title: Desktop & Gaming Motherboards / Boards | Supermicro 14:35:21 there's your RGB boards - now from SuperMicro 14:36:19 unixwitch, I see what I did wrong when I looked for SM boards: I only checked the "workstation" category 14:45:06 well looks like the module is on the machine 14:45:09 time to reboot said machine 14:46:12 woohoo!! it finds the card 14:47:24 sweet! 14:53:04 unixwitch, is there a reasonable way of telling syncoid not to mount datasets on the remote host? I see that it copies mountpoint properties (which is reasonable) but I don't want the remote host to actually mount them ever. 14:53:14 jbo: i don't know off hand, sorry 14:54:03 aye 14:57:46 and i can ssh in, thanks channel for the help 14:57:51 realtek crap engaged! 14:59:15 now time to figure out how to move the zfs to this new one, 14:59:54 voy4g3r2: zfs send | zfs recv! 15:00:04 (or: zfs send | ssh newhost zfs recv) 15:03:08 yeah.. these 3 drives.. i am NOT doing raidz0+1 this time 15:03:10 Happy New Year! 15:12:08 my head hurts.. with 3 drives (4tb each) i can do raidz or raidz2.. no other option is available to me? 15:12:20 you could also do a 3-way mirror 15:12:57 which caps me at 4tb - minus the overhead of the filesystems. 15:13:15 yes, the usable capacity would be the same as a raidz2, i.e. a single disk's worth 15:13:46 buyers remorse.. they did not have 6tb drives there.. only 10tb 15:13:51 and my heart leaped.. 15:14:04 raidz it is and having this backup going.. so i can minimize pain, thank you unixwitch 15:14:22 3 disks is kind of an awkward size, i would prefer 4 disks and do raidz2 (or maybe mirror-stripe) so you get 50% usable capacity 15:14:49 haha.. yeah, they only had 3 drives left and the majority of this build was the hard drives.. 15:14:53 although stripes over 3-disk mirror or raidz2 is not uncommon 15:16:56 yeah.. when i get a nvm.e drive.. i can make it a 4.. 15:17:04 now i remember why i just bought macs for the last 18 years 15:17:12 this stuff gets you in trouble and a thin wallet 15:17:33 hmm, i feel fairly certainly if you'd bought a system of the same config from Apple you'd have ended up paying a lot more... 15:18:29 that is the thing, i would not because i would spend a lot more.. but those g5 macs back in teh day were damn good looking 16:43:00 ah damit... located a machine that is still running 13.1-RELEASE :s 16:43:12 can I directly jump to 14.0 or should I do the intermediary 13.2 update first? 16:47:44 jbo: update to latest 13.2 first because of the freebsd-update bug tripped by something in /usr/include 16:47:53 ack, thanks! 16:48:14 unixwitch, I played around with sanoid/syncoid since you recommended it here yesterday and overall I like what I am seeing. 16:48:31 just trying to figure out how to properly prevent the destination host from mounting when mountpoint properties get replicated 16:48:55 I guess an ugly option is to set canmount=off on the parent dataset on the destination host 17:25:50 ruuun little king 17:26:06 err, wrong chat sorry 17:30:01 your contribution has been noted 18:52:22 Hello! I am running FreeBSD 14 and several versions of openssl are available as packages. `openssl32' is considered an alpha version and so discouraged in production. Which one is it recommended to use for a stable operation? 18:52:47 rockyh: i would recommend the version in base (3.0, i believe) unless you have a particular reason to use a different version 18:54:17 unixwitch: so if it is in base I guess there's no even need to install the package 18:54:28 right 18:55:35 thank you! 18:57:47 rockyh: you only need extra versions is some software depends on it, or you're developing new software against new versions of it 18:58:10 and in the first case, pkg takes care of it 19:04:49 meena: thanks, in my case the openssl in the base system is enough 19:05:30 hurray 19:05:42 :) 19:25:19 is there some way to run some arbitrary application (like say a pdf reader) in a capsicum sandbox? i realise this wouldn't be as secure as modifying the application to support it properly 19:28:58 https://arxiv.org/pdf/1909.12282.pdf looks relevant ("CapExec: Towards Transparently-Sandboxed Services") but the software it's talking bout doesn't seem to exist anywhere 19:32:04 "Once set, the flag is inherited by future children processes, and may not be cleared.", so write a wraper that enters capability mode, but no idea how restricted the child would be if it knows nothing about capsicum 19:32:40 right, but you need something to provide the OS services that the application now can't use 19:58:41 hm, write an email to one of the authors? I probably would first try that @freebsd.org email 20:24:19 nimaje: i mailed the first author (on the basis that that's usually the person who actually wrote the paper :-) so we'll see what they say 21:19:07 unixwitch: while you're at it, you can complain that they don't provide nice LaTeX that can be converted into nice HTML https://ar5iv.labs.arxiv.org/html/1909.12282 21:19:08 Title: [1909.12282] Untitled Document 21:19:37 meena: https://www.arxiv-vanity.com/papers/1909.12282/ ! 21:19:38 Title: CapExec: Towards Transparently-Sandboxed Services (Extended Version) – arXiv Vanity 21:20:18 nice 22:04:32 hm, the latex source looks ok (download in "Other Formats"), but well it uses tikz, no idea how well latex to html converters can handle tikz 22:07:15 ah, that arxiv-vanity render is truncated at the first tikzpicture 22:07:45 oh i didn't notice that 22:17:22 that looks like something that could be converted to SVG 22:20:34 i think the authors should put their paper in a github repo so we can all submit PRs to fix it 22:20:39 open science! 22:21:38 meena: can you even put SVGs in LaTeX or do you have to convert it to EPS/PDF first? 22:23:44 there is some package (svg), that shells out to inkscape 22:27:45 ｍａｎ，　ｍｉｃｒｏｓｏｆｔ　ｓｗｉｆｔｋｅｙ　ｉｓ　ｒｅａｌｌｙ　ｓｔｒｕｇｇｌｉｎｇ　ｗｉｔｈ　ｔｈｉｓ　ｂｕｇ　ｒｅｐｏｒｔ 22:28:20 (but I think meena meant the other way, render that tikz stuff to svg for the html) 22:28:56 yes, that's what she meant 22:30:29 we secretly replaced meena's teletypewriter with a double-width character version, let's see if she notices 22:32:53 There was an error creating your Issue: body is too long, body is too long (maximum is 65536 characters). 22:33:04 ahahahaha 22:34:31 unixwitch: i had to switch to my other keyboard, https://im.eena.me/uploads/7e5ba8e6cde5d6ff/Screenshot_2023-12-31-22-33-53-29_ffb2f5e1b976ff98cfc94f359fbce8de.jpg ｂｅｃａｕｓｅ　Ｓｗｉｆｔｋｅｙ　ｗａｓ　ｐｏｏｐｉｎｇ　ｉｔｓｅｌｆ 22:35:41 Github also not accepting my bug report, because the LaTeX output is probably a few megabytes 22:42:00 i'm doing some work on 'toot' (mastodon client) but it's reminding me why i dislike python so much... wonder if there's a C/C++ alternative 22:45:13 my guess is: no 22:45:52 for some reason i still follow the toot repo, even tho i don't have it installed rn 22:47:18 is there something special you have to do for an encrypted swap in 14-release? i have /dev/nvd0p3.eli in fstab, but it doesn't actually create it/use it (nothing shows up in swapinfo). if i try swapon /dev/nvd0p3.eli, i get Invalid parameters error 22:49:41 bluetooth on FreeBSD looks like fun 22:50:12 bcarson: have you read https://docs.freebsd.org/en/books/handbook/disks/#swap-encrypting 22:50:13 Title: Chapter 20. Storage | FreeBSD Documentation Portal 22:50:35 unfortunately, I do not seem to be able to get it to recognise my card: an Asus AX-3000 with intel chip ax200 wifi+bluetooth 22:50:40 AumShivaya: bluetooth is never, ever fun. or a good idea. 22:50:43 wifi works fine, bluetooth ..otoh 22:50:46 meena, yes, i added .eli to the device name 22:51:17 bcarson: what's dmesg has to say? 22:51:43 nothing :( 22:52:19 apparently I should manually load firmware to my card....which is fine, except I only have instructions for how to do it to usb card..mine is pcie 22:52:25 https://www.davidschlachter.com/misc/freebsd-intel-8265-bluetooth 22:52:26 Title: Intel 8265 Bluetooth on FreeBSD 22:53:35 https://xkcd.com/2055/ 22:53:36 Title: xkcd: Bluetooth 22:56:28 still no new what-if xkcd 22:57:17 if i try with bde instead of eli, it thinks about it for awhile and then returns: swapon: gbde (attach) error: /dev/nvd0p3.bde 22:57:29 (and still nothing in swapinfo) 22:58:31 hmm 22:58:44 i think this is an "nda" vs "nvd" problem 22:58:58 my fstab has nvd, but my dmesg shows nda 22:59:44 hm, i accidentally 'doned' a task on github and now i can't work out how to un-'done' it 23:01:01 oh, found it, there's a secret unlabelled button you need to click 23:01:17 lol, ok, that fixed it. i guess i missed that change. 23:04:19 I thought that was one of the big ones announced 23:06:13 * unixwitch wonders if there's something like Mp3tag.app in ports 23:13:22 Happy New Year from Denmark! 23:37:10 someone suggesting 12.x should be maintained until OpenZFS "is proven", apparently missed the freebsd-specific ossl(4) that destroyed encrypted zfs pools 23:38:18 i missed that 23:38:58 https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=275306 23:39:01 Title: 275306 – 14.0-RELEASE: ossl(4) causes data corruption on encrypted ZFS filesystems/volumes 23:39:19 (my point only being that all software has bugs... staying on ancient software forever isn't really a solution) 23:41:35 Thanks for reminding me I need to upgrade my 12.4 VMs. 23:53:28 unixwitch: they also missed the somewhat important part which is that the race condition was introduced back before OpenSolaris.. was. It dates _very_ far back, and just wasn't possible to hit until I/O offset operations were implemented, and only then if very specific behaviour is done by userland tools (which was enabled by a new set of GNU coretools). 23:55:23 If I understand the userland behaviour, it also seems to me to go against how userland programs should behave when it comes to caching? 23:57:24 ok I have googled 23:57:51 come to the conclusion that no one has a pci-e wifi+bluetooth combo card like I do 23:57:54 I am spesh 23:58:39 happy new year 23:59:08 i've never even heard of a PCIe wifi+bluetooth card 23:59:14 is that an amazon special? 23:59:19 happy new year to all 23:59:59 now you have