02:14:28 It's quiet in here 03:50:17 freebsd is perfect and has zero issues. therefore, there is nothing to discuss. 04:03:43 fair enough 06:08:21 I believe none of the BSD's like my Ryzen 5 + Radeon desktop PC though. FreeBSD can't recognize more than 1 monitor, and generally behaves funny on it, whereas OpenBSD's sndio crashes after playing audio for a random amount of seconds, with the only fix being to reboot the PC. Didn't try NetBSD on it though. 06:08:52 On the other hand, all BSD's work very well on all my ThinkPads and PowerBooks. 06:17:30 remiliascarlet: BSD doesn't work great on myy OptiPlex 3050 too, sound doesn't work :/ 06:17:38 even the beep command does nothing. 09:44:29 schizophrenic: Which BSD is that? 09:44:59 I had much less issues with sound on FreeBSD compared to Linux when I used that. 09:54:54 remiliascarlet: have you tried installing drm-kmod, xf86-video-amdgpu and loading amdgpu? I have Ryzens in a ThinkPad and Lenovo's desktop PC (ThinkCentre or something) and the only issue with multiple screens is that Xorg needs a restart when connecting/disconnecting additional screens. 11:33:30 erk: It's FreeBSD. (we are in #freebsd lol). 11:35:11 the sound card is detected by the kernel however beep command doesn't work 11:35:13 https://termbin.com/a8xx 11:40:40 schizophrenic: I am apperently not awake enough at 11:44 after a holiday. 11:45:06 erk: take a coffee or something 11:46:02 schizophrenic: Do you get anymore stuff in your dmesg if you do something like "dmesg | grep hda" 11:46:13 Just got some lunch so I am already better. 11:47:47 yeah, I found bunch of other devices 11:47:49 https://termbin.com/v4md 11:51:37 It detected Realtek Audio Device 11:51:43 It just looks like the driver is not getting loaded, otherwise there should be some info about the pinout. 11:52:00 pinout? 11:52:13 You could try to add `snd_hda_load="YES"` to loader.conf. 11:52:42 So you could see how each output is set up. 11:53:26 you need to set a tunable for that, driver is loaded as it's reported in dmesg 11:54:00 added snd_hda_load="YES" to loader.conf 11:54:14 yuripv: what is tunable? how to configure it? 11:54:44 schizophrenic: now remove it from loader.conf, snd_hda is builtin in GENERIC 11:55:00 ohk 11:55:04 * schizophrenic removing it 11:55:11 schizophrenic: see man snd_hda, search for pindump 11:55:24 ohkk 11:55:32 * schizophrenic opening man page 11:56:39 Does it only sayt that it is loading the driver if you have that set? 11:58:44 regis: Yes, I did. 11:59:00 And the desktop PC is a custom build, so not a ThinkCentre. 11:59:40 remiliascarlet: Doesn't really matter when it comes to CPU 12:00:06 remiliascarlet: Is amdgpu module loading properly? DRM? 12:00:21 erk: no it was even saying that previously when I didn't added it 12:00:27 another way to see if sound works cat /dev/dsp 12:00:32 ikt should make a sound of some sort 12:00:49 yuripv: I found "dev.hdac.%d.pindump", the man page says I have to set it to nozero value 12:01:11 voy4g3r2: echo $RAND > /dev/dsp 12:01:54 voy4g3r2: it didn't make *any* sound, even `beep` command isn't working 12:02:00 remiliascarlet: You wrote about "BSD" in general so FYI: I'm on FreeBSD 15-CURRENT but was using drm-kmod on 14-CURRENT way before it landed in ports. It's been literally years since Renoir GPU in Ryzen CPU works well for me. I have no experience with other GPUs than Renoir though 12:02:17 regis: Not sure. I didn't boot into FreeBSD on that machine for a long time now. And it's meant to be used as a gaming rig anyway, so. 12:02:52 Never heard of a Renoir GPU. 12:03:25 how could I change value of dev.hdac.%d.pindump? 12:03:25 This is the post-Vega GPU family, packed in at least 4000 and 5000 Ryzen series 12:03:26 https://en.wikichip.org/wiki/amd/cores/renoir 12:06:59 schizophrenic: man page says it's sysctl, so sysctl dev.hdac.0.pindump=1 12:07:42 yuripv: https://termbin.com/vrp6 12:07:43 (just don't ask me about the output, that black magic is above my mortal skills :D) 12:08:01 lol I was just about to ask you what the output means 12:08:05 (some sysctl values are read-only or configurable only at boot time) 12:08:06 schizophrenic: now check dmesg, the output should be there 12:08:45 sysctl -d may shed some light on what's the output. Unlike in Linux, FreeBSD's sysctl is well-documented and sysctl -ad is a trove of information 12:08:49 yeah I can see the output there 12:09:05 however all the GPI%d pins are disabled 12:11:05 regis: syctl -ad gave me what I already knew "dev.hdac.0.pindump: Dump pin states/data" 12:16:00 schizophrenic: Yeah, that's -d. I'm just saying that sysctl -ad in FreeBSD is a great overview of configurables with basic explaination for every single tunable. When I was looking for it in Linux, most tunables had no descriptions what so ever 12:17:07 I agree, BSD in general has good docs than Linux 12:31:29 Linux documentations are specifically designed to be so horrible, you have to depend on search engines, communities, and StackOverflow. 12:32:42 Linux documentation is written by people who don't know how to write 12:33:00 and used by people that don't know how to read. 12:33:08 Kinda works ou 12:33:09 t 12:33:21 and a lot of the documentation that exists is third-party - written not by the developers, but by people from Debian.org 12:34:26 Or GNU. 12:34:36 GNU doesn't write man pages 12:34:55 they write for that horrible monstrosity called info(1) 12:35:13 That's extra fucked then. 12:35:47 i thought Elmur Fudd wrote that stuff 12:35:50 fwiw, I can't write either. (I was a C-student in English) 12:36:01 flunked out of college because ENG099 was too much for me 12:36:37 I have written some man pages for my own project. they're roughly in imitation of the BSD style 12:36:41 On one hand, Linux's modularity is fun because you can replace bloat with less bloat (or more bloat), but on the other hand, this is Linux's greatest weakness, because every configuration ends up in a mismatch. 12:36:57 GNU itself is bloat 12:38:15 The GNU guys are seemingly incompetent. I tried to follow the Linux From Scratch guide, first tool I had to compile was GNU's binutils, and the latest few versions literally cannot be compiled at all, so I had to resort to 6 or so versions older for it to finally compile. 12:38:40 And then the next step was compiling GCC, and I thought "meh, fuck this shit". 12:42:01 I will say I have had better luck rolling gcc than clang 12:42:12 but gcc is a pile of flaming dog waste 12:53:01 yes but building gcc is pretty much a question of running one script to get the dependencies then run configure then make 12:53:09 clang is a nightmare 12:53:57 half of the clang web pages describe building with cmake/make and the other half with cmake/ninja 12:54:59 really the clang doc is "how to build clang if you've been working on the code base for years and already have a working setup" 12:55:15 if it were up to me there'd be a basic C compiler written strictly in C, BSD-licensed, and set up to build with just a generic make tool 12:55:23 Much that I have autotools it simple and it works 12:55:27 have=hate 12:55:47 ./configure && make && su -c 'make install' 12:55:52 done 12:56:34 ideally, it should just be make && su -c 'make install' 12:56:38 but eh 12:56:59 I don't have the mad skillz to write a compiler 12:57:31 If it were up to me I'd mandate using languages that can afford a decent amount of memory safety 13:03:24 also the clang people are quite hostile to outsiders 13:22:15 I haven't had any issues with llvm folk. Sometimes slow to answer, but that's always the case for open source projects 13:29:45 Hallo. I seem to remember recently learning about a tool that re-symlink libraries installed in base, important with upgrading major versions or in thin jails, but I can't remember what it is. Have I described it well enough? Anyone know what I'm talking about? 13:38:23 nope 13:48:48 Creating random symlinks in directories that are ports/package managed is a fools game. 13:49:03 That's what libmap is for. 13:52:26 random: I wonder how hard it would be to hack pdksh or one of its derivs into a ksh93... 13:54:04 inb4 "but ksh93 itself is already open source", the fewer licenses in base the better 13:57:03 Licca: i'm pretty sure ksh isn't in base. maybe you're thinking of the tests that depend on ksh? 13:57:54 I'm not talking about FreeBSD 13:58:57 a project of my own which needs to derive from one or another BSD will require "ksh93" in base 13:59:04 but ksh93 is a mess 14:00:00 like what the hell is that build system? 14:01:27 weird 14:01:33 aix has ksh93, only place i use it 14:01:42 Debian has it, lol 14:02:04 not by default? 14:02:07 it's in apt 14:02:23 if you install "ksh", it's ksh93 14:03:20 out of the box I think you only get dash and bash 14:04:33 Licca: Well written code should be compilable with any C compiler. 14:04:36 ^ 14:08:13 there's also 2 variants of ksh93 14:10:07 most unofficial Korn shells are ksh88 14:10:10 including bash 14:12:14 I have a very low opinion of pdksh 14:12:49 The original author became a usenet troll AND a Micro$soft employee 14:12:53 I've thought of writing my own Bourne shell, but I'm way out of my league 14:12:54 how low can you go? 14:14:21 what's the most minimal shell out there 14:14:27 I am using /bin/sh btw 14:14:44 there is osh (v6 shell) 14:14:58 then there's msh (Minix v7 shell) 14:15:19 do we have that in FreeBSD? 14:15:45 I got msh originally from some one-disk FreeBSD fork from 25 years ago called "PicoBSD" 14:16:08 wow you are using computer for so long 14:16:49 I wonder if there's any letter left to give to another non-POSIX Shell. We already have ash, bsh, bash, csh, dash, zsh, ksh, msh, osh, tcsh... 14:17:51 I have what I call "nash", but it's just NetBSD's /bin/sh recursively backported 14:18:01 (since it has elements of dash) 14:18:08 and fish 14:18:12 and oilsh (for texans) 14:18:20 Debian has posh 14:18:39 there's also yash (Watanabe shell) and mrsh 14:20:20 i do tend to prefer zsh these days, i feel like the completion is better implemented than in bash. also i like having PS1 with dynamic codes. 14:20:27 So we're already running out of letters. 14:20:44 nah. [a-z]+sh can fit as many as you like 14:21:09 Right, then now I'm going to make a jskdjsdkhdsfklghsfdghnsh. 14:21:42 Can't really make an ssh though, since that would conflict with Secure Shell. 14:21:52 and rsh already is overloaded 14:22:01 restricted shell or remote shell 14:22:20 Oh, I thought it would have meant "Rust Shell", which would fit with it being overloaded. 14:22:44 Memsafe shell 14:22:57 Because I tried some of the Rust rewrites before, and I always find them to be way too overengineered. 14:23:19 I feel like a *x should be written mostly in C and the rest, at the kernel level only, in ASM 14:23:30 remiliascarlet: you could make ssssssssssssssssssh the very secret secure shell 14:23:42 👵🏻 14:23:44 Like that Vim clone written in Rust of which I forgot its name for example, it's so bloated, it doesn't even feel like Vim anymore. 14:24:03 clearly written by someone who doesn't understand vi(1) 14:24:11 I think the bootloader should be written in ASM. 14:24:23 kinda has to be 14:24:30 That's the point. 14:24:47 bah, no worries. systemd will boot for you now 14:24:53 but I cut my teeth in the 8-bit world, 65SC02 14:25:10 you want primitive, you got it 14:26:42 Demosthenex: But systemd is almost a complete operating system. 14:26:51 All it has left to replace is the kernel. 14:27:01 I think they're working on that. 14:27:10 at that point, you're no longer running "Linux" 14:27:11 systemd-linuxd when? 14:27:55 You can already run GNU without Linux by using one of the forever experimental GNU Hurd distro's. 14:28:03 or BSD, or Solaris 14:28:20 remiliascarlet: its nearly windows. soon you'll need an antivirus 14:28:25 But those don't come with all the GNU coreutils. 14:28:35 systemd: converting linux to windoze, one system service at a time 14:28:35 that's just one package 14:28:53 Demosthenex: True, some time ago I even read they've been implementing a BSOD into systemd. 14:29:10 I feel like the canonical init system is System V-style (*gets shot*) but I suppose you could coax init(8) into doing that if you really wanted 14:29:12 this BSOD is actually helpful 14:29:28 it lets you page kernel panics IIRC 14:29:29 I want BSD, not BSOD! *gets shot* 14:29:53 remiliascarlet: wasn't the BSOD stuff inside the linux kernel? not systemd? 14:29:54 Licca: You can take a look at CRUX actually, that distro uses BSD-style init scripts. 14:30:05 remiliascarlet: also qr codes 14:30:09 I actually like the "runlevels" 14:30:15 it is funny to see solaris mentioned given the systemd takes a bunch of ideas verbatim from SMF 14:30:17 runlevels was fancy enough 14:30:21 but that's kinda antithetical to how y'all work lol 14:30:22 Grabunhold: https://news.itsfoss.com/bsod-linux/ 14:31:00 duncan: And apparently, the Windows source code is filled with code stolen from Linux, the BSD's, and Solaris. 14:31:39 But then again, we all know Microsoft doesn't give a single fuck about open source licenses. 14:32:01 I know that ftp.exe in Windows is derived from BSD 14:32:15 it is well-established that it includes BSD licensed code, but the notion it includes code from Linux proper is… unlikely given the GPL 14:32:44 same for CDDL unless they did some licensing gig 14:33:04 well, they could pinch from SVR4 without CDDL 14:33:29 hell, Microsoft legitimately owns some of the SVR4 code 14:33:37 Well, at least the BSD and ISC (in the case of OpenBSD) permits the re-use of code. GPL and CDDL don't. 14:33:54 At least, not in the way Microsoft did. 14:34:51 I have a DGAF attitude toward most of my code 14:35:59 I used to release code under GPL, then I switched to BSD, and now to ISC. 14:36:17 BSD and ISC are more or less the same, I just found ISC's wording much clearer. 14:37:50 unless I have a reason 14:37:55 I usually use UIUC 14:41:39 Strix, though, I went with the 2-clause BSD license because it just fits better 14:43:56 Hello, Grabunhold. I am ashamed to say I have not submitted that ticket yet. Give time till tomorrow, and then consider my initiative as failed :-) 14:45:41 ant-x: ah well, it's not like i've been very active either ;) summer hole i guess 14:46:33 at least i figured out the situation with my irc bouncer and nick registration to send you that memo :D 14:50:47 Good for you! So, you are running your own server with a bouncer on it? 14:51:30 ant-x: i'm using https://quassel-irc.org/ 14:54:28 I dunno why I want to chug some liquor when I see that name lol 14:54:38 maybe because it reminds me of "quaff" and "wassail" lol 14:55:03 and yes, the quassel core component has it's own server. I'm in the process of migrating that from linux to a (Ansible-managed) FreeBSD jail 14:55:41 Grabunhold, Quassel must have a server part, which you have to have running on some server machine, right? 14:56:11 Licca, I think of Lev Kassil. 14:56:56 Licca: "quassel" is a wordplay on a german verb, "quasseln". it's something like chitchat 15:04:49 ant-x: yeah, as I said, there is a server machine. my quassel core is using postgresql as a backend, but sqlite is supported, too. 15:21:44 I was not sure what you meant by "its own server." It might be a server maintained by some Quassel foundation... Thanks for the clarification. 16:28:05 quassel core as a service? :D not sure how many users quassel even has any more these days... 16:30:05 Grabunhold: you're on all the cool channels 16:40:34 Do you have a list of bluetooth devices that work on FreeBSD ? 16:57:18 that would be a forever incomplete list 17:00:29 GRRrrr - qt6 hauls in libwayland requiring wayland package and NO OPTION to SHUT IT OFF... wireshark needs it. [scream profanity] 17:01:21 * shbrngdo wanted a system untainted by wayland... 17:01:27 Wayland is the thing that should not be 17:01:34 a *x system should use X Window 17:02:12 it could be streamlined to leverage its client/server nature with multi-core 17:02:35 X already has that 17:03:04 but NOOooo we have to do something that no longer has DISPLAY= whixch I need for embeddeddev so I always will use X11 17:04:00 yeah some optimization but there are still some stack bottlenecks that could use a bit o' tweeking 17:04:39 at least in older versions, won't know until I run the new stuff if it has been tweeked 17:05:43 I still have issue with new mate-panel crashing all panel apps. apparently fixed in github but not in ports or freshports 17:06:44 I think it is dbus-related. lots of logs but it's known already, so maybe if I can patch no need to update all of mate 17:07:26 just submit patch after test, and then let the normal cycle get updates in 17:17:46 well as much as it irritates me I should spend time fixing mate-panel instead of eliminating every last vestige of wayland... 17:18:33 Demosthenex: and the uncool ones as well! :D 17:18:45 that and getting the rest of packages built so I can get workstation 1 up and running 17:22:21 but we need to toss out all the old stuff so we can watch youtube videos without screen tearing! it's SOOOOOO important 17:23:15 sounds like a driver problem 17:23:29 I don't have tearing on X Window, and I have CPU-level graphics 17:23:39 with dual 1920x1080 17:25:02 drivers? (mate-panel issue was fixed 2 months ago according to github if that is what you referred to) 17:25:20 I was referring specifically to "screen tearing" 17:25:28 oh nevermind missed the snark. hweh 17:25:30 since I *did* have an issue with that once and it *was* a driver issue 17:25:52 nothing requiring Wayland ;p 17:26:22 VESA has issues, but usually you can have SDL do frame buffers and switch during monitor blanking. yeah 17:27:33 MPlayer can use OpenGL or VDPAU 17:28:09 right. Idonot reember what VDPAU is but I think it's driver-specific... 17:28:28 I think it's nvidia-related? but somehow my Intel on-CPU graphics seems to work with it 17:28:44 I added SDL support for mplayer 17:29:09 helps to read the config options (usually) 17:29:11 I wish mplayer2 were still around 17:29:31 the same enhancements as mpv but didn't mess with my muscle memory 17:30:52 I use mplayer most of the time, and mencoder. I have mostly gotten used to the weirdness 17:31:48 the main annoyance 17:31:56 is that seeking turns off pause 17:32:19 oh yeah - press any key un-pauses 17:32:39 mplayer2 changed that and mpv keeps that. but mpv messes with 20 years of muscle memory 17:32:55 all the switches are different, all the keybindings are different 17:32:58 hard to get screen snapshots also - at least the 2021 version 17:33:29 I end up using vlc to do that 18:28:39 Does someone publish a log of this channek somewhere? I got some advice on July 21 or 22 I think, and I didn't record it properly and would like to refresh my memory. 18:28:50 channel 18:31:46 gh00p: I don't publish any logs, but I do log everything for this exact reason. I can provide my logs from July 21st to 22nd. PM me if you'd like and we can off-channel this. 19:32:10 does rofi pkg on 14.1 crash? 19:36:51 nmz: Doesn't crash for me 19:38:22 you use it normally? 19:39:25 I sometimes use the calculator 19:40:18 hmm 19:40:24 so its a me problem 19:40:26 thanks 21:46:13 does wayland works with intel gpu ?