@aris

Proof of concept.

This is FreeBSD's #dd, lightly modified to use getopt_long() with a fairly straightforward set of long options mirroring the JCL-style ones.

It's a fairly simple modification, and a very similar one would apply to NetBSD.

This is not the hard part. (-:

@cstross @ryanc
#Unix #FreeBSD #NetBSD

Alt...

A shell session in (mostly) black on white. The dd command is run several times. 1: With the --help option, which yields some output describing both JCL-style and getopt-style options. 2: With getopt-style options writing 7 8KiB blocks from /dev/zero to /dev/null, which it prints a summary of. 3 and 4: With a mixture of getopt-style and JCL-style options resulting in usage error messages. 5: With JCL-style options doing the same as in run number 2.

Can any NetBSD knowers help me? I just installed the latest NetBSD (11.0, the i386 port) on an x86 laptop to try it out again. However, I can't find a recent Firefox (newer than 52ESR) in pkgin:

$ pkgin search firefox
basilisk-2025.10.10nb2  UXP-based web browser inspired by pre-Servo Firefox
fira-4.202           Mozilla's font for Firefox OS
firefox52-52.9.0nb80  Web browser with support for extensions (version 52ESR)
firefox52-l10n-52.9.0  Language packs for www/firefox (version 52)
firefox_decrypt-0.7.0nb3  Extract passwords from Mozilla Firefox profiles
lz4json-0.0.20191229  Unpack lz4json files as generated by Firefox's bookmark backups
py311-FxA-0.8.1      Firefox Accounts client library for Python
py312-FxA-0.8.1      Firefox Accounts client library for Python
py313-FxA-0.8.1      Firefox Accounts client library for Python
py314-FxA-0.8.1      Firefox Accounts client library for Python
syncstorage-rs-0.21.1nb4  Firefox Sync Storage server in Rust
=: package is installed and up-to-date
<: package is installed but newer version is available
>: installed package has a greater version than available package
$ pkgin search librewolf
No results found for librewolf

Please help?

In theory, I could attempt to compile it from pkgsrc, but compiling Firefox is already extremely unpleasant on a fast x86_64 computer with huge amounts of RAM, or (that one was SeaMonkey, presumably less heavy than Firefox itself) on a PowerPC computer with 1GB of RAM, I don't even want to try it this laptop with its 256MB of RAM...

#NetBSD

#NetBSD Pros & Cons
#Debian #runbsd

This is fascinating. I'd taken for granted that NetBSD had a no-slop policy, but it isn't quite that straightforward.

From their commit guidelines, "Code generated by a large language model or similar technology, such as GitHub/Microsoft's Copilot, OpenAI's ChatGPT, or Facebook/Meta's Code Llama, is presumed to be tainted code, and must not be committed without prior written approval by core."

So, it's not a closed door.

https://www.netbsd.org/developers/commit-guidelines.html

Compare with Gentoo's policy:

https://wiki.gentoo.org/wiki/Project:Council/AI_policy

#netbsd #noai #noslop

This also adds prebuilt #NetBSD binaries, logs when challenges are accepted, cleans up and consolidates the worker code, fixes an edge case in Content-Security-Policy parsing, and more.

@Jonas nothing to worry about. I would suggest first playing around with virtualbox or qemu to understand #NetBSD ask community here for help or mailing list or any community forum. Once you get comfortable with terminology and how things work you can try it on bare metal machine and see how it goes.

What are you switching to now that #Debian and even #Linux is getting slopified? I'm looking at #NetBSD, but I've got zero experience with #BSD so pls help

@cstross

Well it wouldn't be at random. getopt(3) is fairly well known and established at this point, and using it a fairly mainstream notion since the days of AT&T System 5 Release 3, although there are parts of the world that look askance at long options (invented in 1989 if memory serves) as if they are heresy.

And it wouldn't change the JCL-like syntax one whit.

I wonder whether anyone has actually attempted this. A quick look around doesn't reveal anyone even questioning this religious tenet, which is amazing given how people have reinvented stuff over the years. I would still be surprised, and I haven't looked everywhere. But it might actually be, @aris , that you are the first to suggest this in 40 years.

If it's taken that long for someone to even float the idea, my prediction of Sol going red giant is still looking spot on. (-:

@ryanc
#Unix #dd #NetBSD

Some napkin math for #BSD CVEs discovered with AI assistance:
#FreeBSD 27
#OpenBSD 1
#NetBSD 0
#DragonFlyBSD 0

@cstross

One wouldn't have to discard the old syntax, thanks to what #NetBSD and #Illumos have done long since. One could make it fall back to the JCL-like syntax in the cases where no modern getopt() options were used.

All existing shell scripts that run on NetBSD and Illumos or that use GNU coreutils or BusyBox, at least, won't be using any modern options at the moment, because that's a fatal error, so there's no existing scripting to trip up.

The GNU and BusyBox people have snuck --help into their dd(1)s, but I don't see that being conflicted with. (-:

@aris @ryanc #dd #Unix

Upgrade du serveur de #Debian 12 -> Debian 13 ✅

Juste un énorme plantage durant l'install - probablement à cause d'un dépôt mal configuré (et qui semblait encore référencé Bullseye...). Je suis reparti du dernier snapshot #hetzner , puis upgrade classique, reboot et on est à jour.

... même si je serais sérieusement tenté par une VM #NetBSD si je devais repartir d'une install.

@cstross

@aris made a good point. BSD fixed the ps(1) command to have getopt(3) syntax in 1990. Maybe it's finally time to do the same to dd(1).

#NetBSD put the beginnings in place in its dd(1) 26 years ago, and is currently ahead of all of the other BSDs in this respect.

https://cvsweb.netbsd.org/bsdweb.cgi/src/bin/dd/dd.c.diff?r1=1.17;r2=1.18

#Illumos also introduced the beginnings of getopt() to dd, at some point before going open source.

https://code.illumos.org/plugins/gitiles/illumos-gate/+blame/7c478bd95313f5f23a4c958a745db2134aa03244/usr/src/cmd/dd/dd.c#495

At this rate, we should get modern argument syntax as an optional extra in dd just before Sol goes red giant. (-:

@ryanc
#dd #Unix

#missionreport #netbsd #bsd
Installation - Dell latitude E5470 (i3/nvme/ddr4 16Go)
#grub bootloader from #mxlinux / #xfce4 /systemd (🤷)
Menuentry > add - ok.
Multiboot > add - ok
Kernel NetBSD11.0 - ok
Wifi - ok
Pkgsrc / pkgin - ok.

Done
Shutdown -h now.

if netBSD ever solves the issue of certain capture devices not working i'd consider it #netbsd

@nina_kali_nina #NetBSD 🙌

#omarchy #NetBSD 🚩

FIND yourself a a linux distribution or BSD that does reject AI; use it and donate to it.

If you can't move your daily driver move a secondary machine. Tinker with it, fit it into your life where it makes sense.

YEAH the linux kernel has slop. Most programming languages have slop. Most desktop environments have slop, you will have dependencies that are still slop. This is a big problem; there is no magic off switch to fixing this overnight.

We are in the early days of dealing with a contaminate that polluters are actively dumping into our digital environment.

It is going to take a long time before anyone can truly be slop free.

But support, and invest in the folks doing that hard work which will build a slop free future.

Starlight Network maintains a list of slop rejecting operating systems and distros here: https://noai.starlightnet.work/list.html #floss #AntiAI #linux #bsd #netbsd #elementaryos #gentoo

@argv_minus_one I think #NetBSD's stance is better it not just avoid AI LLM code but also any means of copy paste from other places or any means which can taint Base operating system.

When I see someone say that moving away from an OS that has questionable choices would require "living a hairshirt existence" I laugh out loud. The first 1/3 of my "Linux lifetime" was just that.

I was there when Linux was new. The chances of a UNIX app not being provided by your distribution was quite high. You had to manually calculate monitor and video card clock timings to get a GUI. You probably had to dual boot if you wanted to do anything like gaming.

1

Early adopters and contributors refused to accept this situation and stuck with it. It went from hairshirt to mainstream in part due to people sticking to their principles. And, above that, the principles of open source.

So yes, switching is hard. It will be difficult for a time, but enough people standing by their principles can lead to critical mass. It's unbelievable that the very people whose careers are built on open source are forgetting how it works.

1. This list is NOT exhaustive. There was also a binary format change, non-modular kernel and more.

#linux #noai #netbsd

Sending good vibes to everyone in the Debian community navigating the recent transitions. If you're using this moment to do a little distro-hopping, consider jumping over to the BSD side. NetBSD offers a completely independent codebase built on a classic, predictable philosophy. You might just find your new permanent home! 🚩💻
#NetBSD #Debian #FOSS #noai #TechCommunity #antiai

@oblomov hi, #NetBSD doesn't allow AI LLM code you can ask for the core on the mailing list :) @netbsd

The tragedy of #Debian accepting AI 'contributions' is that, most of the time, you are accepting stuff that:

a) is someone else's code, with all the potential legal issues (hence #NetBSD stance on it)

b) is of a very dubious quality, since it went through an AI blender

Just remember: an 'AI' does not exist and has no concept of code, quality and correctness -- I feel like we will soon see the first CVE with a CVSS 9.99 score in AI supplied slop vile coding due to this.

Funnily enough, I was never really into #Debian ... I appreciate the quality of the distro, but... I don't know, too many moving parts, I guess.

I will continue supporting #Slackware I'll look into #Guix and, of course, #OpenBSD and #NetBSD as well as #HardenedBSD .

The line is drawn. It is time to choose your side.

#netBSD is the one that's super-anti AI. it even has a macbook install page - for twenty year old models. this should be good.

oh, wow. the iso that was dd'd as were all the other ones isn't even detected as bootable. they got a special, USB image .gz available for download. how bad do I wanna dick with that at 9 MB/s write speeds?

Fuck.

Well, #TempleOS it is then

https://come-from.mad-scientist.club/@algernon/statuses/01M199QQV3ZYMW315SMD26VD7X

#BSD #FreeBSD #OpenBSD #NetBSD #slopware #AI #LLM

We are counting down to the European *BSD event of the year! 😈⛳🐡

Big thank you to our bronze sponsor: Prodottoinrete
https://prodottoinrete.it/

If you haven't secured your spot yet, now's the time!
https://tickets.eurobsdcon.org/

📅 You can check out the program at https://events.eurobsdcon.org/2026/schedule/

EuroBSDCon 2026 in Brussels, Belgium 🇧🇪
September 09-13, 2026

#EuroBSDCon #EuroBSDCon2026 #BSD #RUNBSD #FreeBSD #NetBSD #OpenBSD

Alt...

Logo Prodottoinrete

#Debian was structured as a democracy of developers (maybe technically a platonic aristocracy, but it does not matter here).

Thanks to this, it got trust that turned it into a fundamental building block for #Linux users (most used distro are Debian derivatives).

Unfortunately, such trust also attracted commercial interest.

Now most of Debian Developers (just like most of #Linux kernel's devs) are corporate employees that would not be assigned to Debian development if they would oppose corporare interests.

So ultimately Debian cannot be trusted anymore: the vote on #LLM is just an evidence of this fact.

To people who use it since decades (I started with Potato because of its focus on #freedom), it's a pain.

Right now I'm slowly moving to #NetBSD.

@skjeggtroll@mastodon.online @freshstart@hachyderm.io @neil@mastodon.neilzone.co.uk

It seems that #pkgsrc/www/firefox-155 under #NetBSD/amd64-current has no major problems...

As of today, I believe NetBSD is one of the most important operating systems to consider for your deployments.

With hardware costs rising dramatically, portability and efficiency have become more important than ever. If prices do not come down, we will probably have to adapt and create new kinds of hardware, at lower cost. And NetBSD, by its very nature, will be very easy to port to that hardware.

Does it run on your modern PC? Maybe. Maybe not.

But it runs on your server. On your VPS. On your SBC. On your old computer, which may still have a lot to give, especially if you want to keep ownership of your own data.

And I know something about old hardware still having a lot to give:
https://it-notes.dragas.net/2023/08/27/that-old-netbsd-server-running-since-2010/

No jails or containerisation? smolBSD has already shown that a reduced NetBSD kernel can be so small that it can be started in milliseconds, while also adding the security benefits of virtualisation. A different concept, certainly. But one that already exists and can already be used.

Linux itself has opened up to AI. That is a choice I will not comment on.

But from now on, for those who do not want LLM-generated code in their operating system, there are essentially two options: fork Linux (!!!), or choose something different.

And NetBSD is that something different.

Already in 2024:

“...code generated by a large language model or similar technology (e.g. ChatGPT, GitHub Copilot) is presumed to be tainted (i.e. of unclear copyright, not fitting NetBSD's licensing goals) and cannot be committed to NetBSD.”

#NetBSD #RunBSD #smolBSD #IT #SysAdmin #OwnYourData #ModernTech

@linuxuserspace
For the kernel:
#Linux❗
#Hurd❗
#FreeBSD ☑️
#NetBSD ☑️
#OpenBSD❔
#DragonflyBSD❔
#KolibriOS❗
#RedoxOS ✅
#Haiku ☑️
❗LLMs used
❔Unknown or unclear position about LLMs but I didn't found any LLM generated commits
☑️ LLM use forbidden or limited for the code
✅ LLM use forbidden

...let's lower the barriers even further!
Now this instance is running on a Raspberry Pi A+ (1 core @ 700 MHz, 256MB RAM), so even less than the Raspberry Pi Zero W it was running on before.

On NetBSD, of course!

#littleFedi #NetBSD #RunBSD #OwnYourData #Decentralization #OwnYourVoice

Alt...

The Raspberry PI A+ that is currently powering this instance.

@cffiegel @hermod @phaeton @pOtiron @bastamedia
#Linux❗
#Hurd❗
#FreeBSD ☑️
#NetBSD ☑️
#OpenBSD❔
#DragonflyBSD❔
#KolibriOS❗
#RedoxOS ✅
#Haiku ☑️

❗LLMs utilisés
❔Position inconnue ou peu claire mais pas de commits par LLM (en tout cas à ma connaissance)
☑️ Usage de LLMs interdit ou limité pour le code
✅ Usage de LLMs interdit

Certaines distributions linux comme Gentoo ont également interdit les contributions par IA. Malheureusement, il me semble qu'aucun système grand public n'a banni les LLMs.

...let's lower the barriers even further!
Now this instance is running on a Raspberry Pi A+ (1 core @ 700 MHz, 256MB RAM), so even less than the Raspberry Pi Zero W it was running on before.

On NetBSD, of course!

#littleFedi #NetBSD #RunBSD #OwnYourData #Decentralization #OwnYourVoice

ternyata oh ternyata
gw reboot box, request IP baru, dan coba dari awal dengan cara mirip di #OpenBSD akhirnya bisa juga. #NetBSD ane bisa terhubung ke IPv6!

Right. I've done the same. I've been more forthright than the #pkgsrc people.

The #xmlto in #slashpackage here removes 'type' completely, gets rid of all of the other #Bashisms that I can find, uses sh as the interpreter, and in a modicum of testing (re-building the nosh manual set) works with dash on #Debian and the #FreeBSD and #NetBSD Almquist shells.

It also defaults the wwwbrowser configuration setting to www-browser (as used in Debian's alternatives system).

https://github.com/jdebp/slashpackage/tree/trunk/textutils/xmlto-0.0.29

hopefully #NetBSD gets some fresh development in the days ahead. it would be nice to have even a basic Nvidia driver that supports my card (RTX 3070) because I don't think the available nouveau driver does. that puts me on #FreeBSD solidly for now, LLM sludge or not.

Upstream wasn't experienced/use NetBSD enough to want to merge it (fair), so pushed my fork of susmb (a FUSE implementation of SMB v3) for NetBSD support to Codeberg here: https://codeberg.org/gourd753/susmb-netbsd

If I can hammer out one remaining issue (the PERFUSE_BUFSIZE environment variable needing to be an arbitrary size for large file transfers to work) I'll see about trying to get a netbsd package for it but atm it's too finicky for me to feel comfortable with that.

#NetBSD

Most of this setup has been progressively pulled out of ewaste piles over a period of several years and I think it's finally time for it to run NetBSD and fulfil its true destiny.

Is a computer from 2012 actually "retro"? Well ... it will be when I'm done with it.

#retrocomputing #thinkpad #NetBSD

Alt...

A ThinkPad T530 at a gnome desktop attached to a dock with a key lock on the side, and a Lenovo keyboard and mouse set in front

@jperkin

The #xmlto people did actually try to remove the #Bashisms themselves, over a decade ago. But they left it still auto-configuring to pick bash instead of changing that as well. Which is what it does if one builds from vanilla source today.

https://pagure.io/xmlto/issue/3/

Then, 10 years later, the change happened that put the bashism back in that creates the #NetBSD noise, perhaps because they did not go the whole way and make it a /bin/sh script, which would have highlighted that.

Found it. pkgsrc is explicitly setting the XMLTO_BASH_PATH environment variable (amusingly, entirely ignoring the strong hint in the variable name) before running configure . I was looking for a patch; but this isn't being done with a patch.

Since I've patched out all of the type -p and type uses (sometimes being used in wildly overcomplex ways) and replaced them with command -v, I wonder whether I can do the same.

#NetBSD #Bashisms #xmlto #slashpackage

I added #xmlto to my #slashpackage tree because I wanted to patch out the 'type -p' thing that causes a complaint on #NetBSD every time the tool is run.

Oddly, building from vanilla source makes it a bash script, as (from reading the Fedora issues list) the authors actually intend. It isn't supposed to be a POSIX sh script. bash has type -p .

Yet the installed-from-package xmlto on NetBSD has /bin/sh as the interpreter, hence the complaint about the #Bashisms.

I wonder what's going on.

#NetBSD / #ctwm folks,

Any idea how to create a keybind for super+. (period)?

This doesn't seem to work at all:

"."      = mod4     : all : !"~/bin/emoji-picker"

Pressing super+. just gives me a period, and there's nothing in ~/.xsession-errors to indicate that it triggered anything at all.

Nevermind! Found the solution, derp:

"period" = mod4     : all : !"~/bin/emoji-picker"

#NetBSD is famous for being able to run on any hardware imaginable, yet somehow I've not been able to install it successfully on even the most bog-standard machines.

@chadmccullough This saddens me as well, I recently started testing NetBSD for this reason, but I was holding out hope on this Debian vote, as it has been my preferred OS distribution for decades.

Oh, well. Time to put more energy into the testing of/transition to NetBSD. I also just updated my software donations rotation, replacing Debian with the NetBSD foundation.

#NetBSD #debian #noai

Wrote a patch to get SUSMB, a FUSE implementation of SMB working on #NetBSD, and running into an issue where files larger than a certain size hang the driver. This appears to be related to the "PERFUSE_BUFSIZE" environment variable, and some forum posts I've seen indicate this is an issue across PERFUSE given it came up with NTFS-3G.

Does anyone here with more NetBSD experience than my full uh, week lightly toying with this have any experience with this?