SiFuh changed the topic of #crux-social to: Offtopic Talks | Project https://crux.nu/ | Logs: https://libera.irclog.whitequark.org/crux-social/
<remiliascarlet> Chat: "Why can't those arrays be sorted?" Jonathan Blow: "Because fuck you."
serpente has quit [Remote host closed the connection]
<SiFuh> remiliascarlet: Jeff Berwick's dog Lucy died a few days ago
<remiliascarlet> Condolences.
<remiliascarlet> SiFuh: Also, I got st working correctly on NetBSD now. Turned out it was me fucking it up, and all I had to do was remove the TERMINFO env variables from my .zshrc file. Because while Linux, OpenBSD, and FreeBSD didn't mind it, NetBSD really hated that.
<remiliascarlet> OpenBSD is my main OS, FreeBSD is my NAS OS, Linux is my gaming OS, and NetBSD is my secondary OS in places where I need extra performance I can't get out of OpenBSD.
<SiFuh> Sucks to be you :-P
<SiFuh> Laptop, OpenBSD. Main driver, OpenBSD. NFS, OpenBSD. TV Box, OpenBSD. Repo, CRUX.
<remiliascarlet> I actually like running multiple OS's side by side.
<remiliascarlet> OpenBSD and NetBSD seem to be the best options among the OS's that aren't dead. And I support FreeBSD and Linux because most people use those 2 among the non-proprietary ones (talking about software development).
<remiliascarlet> The only part that sucks is the lack of a good Linux distro with the musl library.
<SiFuh> I might format the CRUX repo and turn it into a monero miner
<remiliascarlet> If I had the patience, I would take the CRUX base, remove sudo and PAM from core, replace glibc with musl, and recompile.
<remiliascarlet> But if I were to make my own Linux distro, I would probably just hard fork the Linux kernel, delete all the bloat a la OpenBSD, and replace all the GNU utils with more POSIX-like alternatives.
<remiliascarlet> Simply because I'm not content with the current state of Linux and GNU.
lavaball has joined #crux-social
<SiFuh> I beat the wall and won
farkuhar has joined #crux-social
<SiFuh> farkuhar: This is completely broken https://crux.nu/Wiki/BuildingISO
<farkuhar> SiFuh: not completely broken, just in need of some updates. The closing advice "I highly recommend reading through the Makefile" is still sound.
<SiFuh> Heh
<SiFuh> What Makefile when the page is borken
<SiFuh> Hahaha
<farkuhar> SiFuh: I'm intrigued by your OS monoculture across so many devices. That TV box running OpenBSD -- did you flash the firmware of an off-the-shelf model, or assemble a PC from generic components?
<SiFuh> It's a mini PC like an Intel NUOC
<farkuhar> SiFuh: the BuildingISO page loads just fine for me. Granted, the git clone commands need revised arguments in light of the latest updates on crux.nu, but there's still a good overview of the ISO building process.
<SiFuh> Loading is fine. The entire page is broken
<SiFuh> It needs a re-write and the links replaced
<farkuhar> Anyone who was paying attention during the recent changes to the crux.nu server will know the correct links to use. And the Makefile still has all the targets mentioned in the 2016 document. Re-writing is justified, but don't oversell it.
<SiFuh> Welcome to CRUX, broken web pages, incorrect links, and you should have been around awhile ago so that you'd know.
<farkuhar> SiFuh: remember how the recent update of meson-python led to a build failure of python3-numpy? And those were two ports maintained by the same person! You can hardly expect jbrooks to come back and update a 2016 Wiki page as soon as the server gets an overhaul. Instead, a long-time user (like yourself) points out the discrepancies, and eventually they get fixed.
<SiFuh> The point is, if you are advertising something that you want people to use, then that should have been a big priority
<SiFuh> One of the cool things I liked about PHP is I can use a config.php to specify variables. I could have changed every URL on the site by editing one line
<farkuhar> jbrooks clearly expressed the limited scope of the BuildingISO wiki page: just a collection of "notes to self" that others might find useful. Nobody is advertising or promising anything. And here's the great thing about a Wiki: if you find something wrong with a page, just log in and fix it!
<SiFuh> jaeger is suppose to keep that up-to-date since he is the ISO king
<farkuhar> jaeger is responsible for the system/iso git repository. If he wants to fix an old Wiki page, he's free to do so, but he doesn't inherit that job automatically. Otherwise there's no upper bound to jaeger's workload (people could just write thousands of Wiki pages mentioning the ISO and you would expect jaeger to keep them all updated?)
<SiFuh> Yes
<remiliascarlet> SiFuh: By the way, any idea why I can't VM's in QEMU more than 6 GiB RAM on OpenBSD, even though I have 32 GiB of RAM total, of which only 2 GiB is being used?
<SiFuh> No idea. I don't use QEMU. I think it sucks
<remiliascarlet> Of course the solution is on REEEEEEEEEEEEEEddit...
<remiliascarlet> I love how OpenBSD also has manpages for config files.
<SiFuh> And manpages for the manpages
<remiliascarlet> The only thing missing is a comfy way to browse the manpages without having to know beforehand what pages you have (unless you ls).
<SiFuh> remiliascarlet: Didn't I tell you to use apropos?
<SiFuh> apropos, whatis – search manual page databases
<remiliascarlet> You did, but the name is so hard to remember.
<SiFuh> whatis is easier
<SiFuh> If you can't remember you can do man man and scroll to the bottom and it will be mentioned there
<remiliascarlet> `whatis love && echo "baby don't hurt me"`
<SiFuh> whatis: nothing appropriate
<remiliascarlet> Yeah, didn't work.
<remiliascarlet> But the `whatis` command kind of reminds me to that song.
<SiFuh> remiliascarlet: I also can never remember the command apropos :-P
<SiFuh> search_man would be more apropos/appropriate*
<SiFuh> I saw this, this morning
<SiFuh> What a nutcase
lavaball has quit [Remote host closed the connection]
lavaball has joined #crux-social
<SiFuh> remiliascarlet: hmm
<SiFuh> remiliascarlet: Should be root:_sndiop
<SiFuh> On my 7.5 box it is root:_sndiop
<remiliascarlet> On my T400 it is root:_sndiop.
<remiliascarlet> But I'm curious, let me test it on my main desktop, where this problem exists.
<SiFuh> My TV Box is also root:_sndiop
<remiliascarlet> Have been testing audio with this video for 6 minutes so far, but no crashing yet: https://invidious.jing.rocks/watch?v=iwAIXeljqe8
<farkuhar> The 2024-04-08 whistlez message mentioned running OpenBSD in a virtualbox VM, but I can't see why that would have anything to do with the incorrect groups of /dev/audio*
<SiFuh> farkuhar: I seem to remember a permission issue with CRUX in a VM a few years back. As for OpenBSD, it was never designed to be run in a VM anyway.
<remiliascarlet> FreeBSD devs use Macbooks and run their OS exclusively in a VM under macOS. OpenBSD devs use ThinkPads running their OS on it natively. And this is instantly noticeable when you install either OS on real hardware.
<remiliascarlet> Well, seems like sndiod is not willing to crash today.
<remiliascarlet> OK, it finally crashed!
<SiFuh> Congratulations!
<SiFuh> remiliascarlet: Mine doesn't crash anymore. I actually found out what is wrong.
<remiliascarlet> And as it turns out, whistlez is wrong.
<remiliascarlet> So what is wrong then?
<SiFuh> remiliascarlet: 7.5 ;-)
<remiliascarlet> I mean apart from that.
<SiFuh> The update
<remiliascarlet> Because once again, this never happens on my ThinkPad T400, only on my desktop.
<SiFuh> So far the TV Box has no issue either.
<SiFuh> But I also had network issues in 7.5 as well on the main driver. And the laptop has GPU issues on 7.5. So I downgraded my main driver and the laptop to 7.4
<remiliascarlet> What GPU's are these 3 using?
<remiliascarlet> Because I remember that FreeBSD had a fuckton of issues of all sorts of different things on an AMD GPU, but it works relatively stable on Intel HD Graphics.
<remiliascarlet> Which makes sense, because MacBooks use Intel GPU's too usually.
<SiFuh> Can't remember. I think Green Sardine or something is what the laptop has
<SiFuh> But it is the only one with the GPU issue on 7.5
<remiliascarlet> One thing I noticed is that if I do `doas audioctl` before the crash, the name is "azalia0", but after the crash it changed to "azalia1".
<remiliascarlet> Then I did `doas chown -R root:wheel /dev/audi* && doas chown -R root:_sndiop /dev/audi*` now sndio no longer reports errors and plays audio normally, except it's muted because it's listening to the wrong device.
<SiFuh> You should have two azalia
SiFuh_ has quit [Remote host closed the connection]
<SiFuh> One for HDMI and one for your speakers
SiFuh_ has joined #crux-social
<remiliascarlet> Sounds about right.
<SiFuh> Back in the day, you could tweak the kernel to get HDMI to work
<SiFuh> I think it was from 7.3 onwards, it no longer works
lavaball has quit [Remote host closed the connection]
lavaball has joined #crux-social
lavaball has quit [Remote host closed the connection]
DesRoin has quit [Ping timeout: 256 seconds]
DesRoin has joined #crux-social