rwp: no

rwp: conscious "management" decision to roll a new quarterly regardless of build failures/missing dependencies and packages available are the result of that quarterly.. even when notified a patch/upgraded package has caused an issue with dependent packages.

Does pkg "latest" improve that situation? I have seen just as many people complaining here running latest too.

rwp: the only safe and sure way to ensure everything you depend on is patched and available is run your own repo... even if you're a home user... *very* bad IMHO.. however i was shunned for calling it out (as were others) so that'll be the last i say of it for another few years.

point of order: "dummy spit"?

It's an idiom. en.wiktionary.org/wiki/dummy_spit and usingenglish.com/reference/idioms/spit+the+dummy.html

We don't call them dummies on this side of the pond. It's not a familiar US idiom.

'pacifier spit' doesn't have the same ring to it :P

Someone having a cow would also be a hard idiom to translate too.

it sure doesn't

hehehhe actually this is a better description of what happened/what i did.. urbandictionary.com/define.php?term=Spit%20the%20Dummy

... num num spit?

(and yes I'm Aussie)

ah ic :)

We are a great people separated by a common language. :-)

lol

my dummy spit was that bad i forked freebsd and ran my own version (incompatible with freebsd binaries in the end) on my own systems (across SORBS) .. used it until the 30 June this year when SORBS was shutdown. only thing i never got ported was haskell ...

(was a really dedicated dummy spit)

Michelle: it seems you're not the only one to dummy spit regards to pkg and FreeBSD - i get the impression it (pkgbase in this case) was part of the reason ixsystems started to go linux

in my case, it's like...well...I'm puttng together a userland, but I can't write a kernel x.x

and I'm like, well, what's closest?

entrop, not surprised in the least.. someone wanted it to be more linux package management like.. which is fair enough if done right... but i personally got the impression it was about someone making their mark at the cost of everyone else.

i don't feel like switching to linux is innovative

^

and I thought of how Debian had at one point been experimenting with BSD kernels

(I actually ran "Debian GNU/kFreeBSD" on a friend's server for a while as a compromise with him)

i mean, it might be a great way to focus on the problems you really want to solve, because there's a horde of other people working on everything else

Licca, can't you still get debian with the freebsd kernel?

Michelle: lists.debian.org/debian-devel/2023/07/msg00176.html

(except for openssl and log4j)

not anymore

they stopped after 2 versions

"kFreeBSD" is basically my second choice as a potential kernel for the project I called "Unitas"

(my first is "kNetBSD")

woop woop

mine was BSDSUniX (aka BSDSUX).. however as it was really only me, and i have a new AI project I'm currently working on ( linkedin.com/feed/update/urn:li:activity:7216240768015851520 ) i don't know if i'll keep it going, especially as i need to stand up new (public) servers as was using SORBS machines as distribution nodes for the base OS and packages.

I use github for now

i have a soft place in my heart for netbsd

I feel like NetBSD is the most "BSD" of the BSDs, with FreeBSD a very close second

Me too.

grrr, pkg upgrade still wants to remove my media player :( let's see if pkg lock will keep it. /yay, it does :)

so I wanted to make an actual OS around my "Strix" userland project, but I was having trouble writing some elements

the second biggest issue is that although it's not possible for the whole system to have one license (my Strix code uses the FreeBSD/NetBSD license; I use various similar licenses for different projects) I do want to avoid a bunch of different licenses, and even more, I want to avoid "herpetic" licenses

I'd have to prepare a custom build system

the usual issue was that the idea of creating standalone build trees for certain projects and methods of rolling them just led to negative reactions

like "why would you want to build your own clang?"

or "why would you want to have a kNetBSD build tree without the rest of NetBSD?"

but what I was trying to create was distinct from what any BSD is

I just wanted to use a BSD kernel because they're generally close enough and it would save me a lot of work lol

but trying to create even just enough of an environment to cross-compile a basic system from my native OS (lol, Debian) and then start hacking on it in a gradually more native environment, and get it to the point where it starts to resemble a "proper" OS...

blame RMS for gcc cross compile woes

I don't even want to use gcc, lol

I'd prefer to use clang

but I can use gcc if push comes to shove

easy as pie on clang, I haven't checked lately but it's easy as knowing the target triple

about, imo, the only thing RMS did right was convince Keith Bostic to Theseus BSD

anyone knows how I can get pkg to print the .pkg file in the repo path

so my ultimate goal would' be to have a tree where I can use GNU or BSD make to "bootstrap" the base system from outside, or to roll it from inside; which is based somewhat on a BSD but feels more like SVR4; and ultimately runs the CDE

... and have it be something you made, yes?

as far as I can

like if such a thing magically already existed, *pouf*

100% isn't possible. and I'd be using FreeBSD and/or NetBSD to try to fill holes

well, illumos is prolly the closest thing that does

but the CDDL makes me twitch

that's what i was sort of wondering about

yeah? i have only tldrlegal'd it lately

well, I really wanted something more BSD-licensed

so I started doing it myself and created Strix

that is cool.

the license on my own code is the FreeBSD/NetBSD (afaict it's the same) 2-clause license

so I'm like...well...I'm gonna have to roll a few things myself. I'll prolly replace some of my existing stuff with BSD code.

one might ask, what would I use for /bin/sh... ash - why would I use anything else?

works fine for you, works fine for the NetBSD guys, even works fine for Debian and Ubuntu

no rc or es? :)

es might be something worth making a secondary shell, on the same level as csh

9front is MIT-licensed...

lol certainly doesn't feel like SysV though

I don't want "Plan 9 from Bell Labs", though, I want SVR4 lol

yeah so SVR4... all i'm coming up with when i think of it is numbered init scripts and sysv shared memory. i don't actually know what it feels like

NetBSD seems to be going in that direction

but they're also getting a bit GNUy-chewy too

I also decided that just as original SVR4 put a more or less 4.3BSD "universe" in /usr/ucb, I'd do similar - but with something more modern, i.e., pieces of FreeBSD and NetBSD

yeah k see that's what i remember distantly of solaris - what's this whole other set of binaries doing over here??

and why, *why* is ping only there and not on my path??

SunOS 4 was a BSD that had SysV elements, and then AT&T forced SVR4 on them

what issues might I be running into when upgrading from 13.3-RELEASE to 14.1-RELEASE ? My setup is simple, out of the box, with a single jail setup via bastillebsd, let's encrypt certs setup using dehydrated.

asking pre-upgrade, or having a problem?

you could easily get 90% people saying "worked for me", and 10% saying it caused them trouble. It all depends on your setup. Just make backups and go for it

bdrewery: pre-upgrade. I recall that when 14.0-RELEASE came out, there were people talking about certain things not working, etc.

I've been having issues with sslh after upgrading to p5 for 13.3-RELEASE, in which I'm seeing sslh process has died for some reason and the service no longer running. I was considering maybe I bit the bullet and move to 14.1-RELEASE

the only thing that comes to mind is that the step to upgrade the boot code was mentioned in a release note in 13-release somewhere, and i skipped that step, so i needed to do it when upgrading to 14-release..

the other thing that comes to mind is that the root shell changed, and it needed to be paid attention to in the file merges

I don't recall what I did for upgrading from 12.x to 13.0. I don't think I did anything with boot code. Will have to verify and check.

Thanks scoobybejesus

skered: no, i installed this system this year

maybe in january

managed to pull a 10.3 tree out of Debian's site lol

So strange now.

I tried to connect to my XMPP server with dino

But it doesn't work, though profanity client works

It throws out some "Authentication error"

I'm pretty much sure FreeBSD-related. Because on postmarketOS dino works very well

And pidgin also doesn't authenticate. Only profanity.

Probably something with SSL? or with something else common external.

Should I expect SSL working fine on 13.2?

Or is that too old?

For SSL to work

there's also the potential issue in my concept: it would have to be a full-on fork, I think, even if it's just part of the operating system. just like Debian has to say "kFreeBSD", "kNetBSD", not "FreeBSD" or "NetBSD"

in this case too, uname(3) will have to say something else so that FreeBSD doesn't get blamed for my tweakery ;p

should i expect the computer to reboot normally after freebsd-update?

freebsd-update fetch

and

freebsd-update install

will this override the kernel?

i both did pkg update, pkg upgrade

and the one that i told previosly

What's wrong with my installation?

I upgraded with pkg upgrade

And did freebsd-update, so no there are no upgrades available

But now I can't start sway.

ld-elf.so.1: /usr/local/lib/libepoll-shim.so.0: Undefined symbol "kqueue1@FBSD_1.7"

pkg-static upgrade -f

for what reason this command is? ^

can or can't the system be upgraded without this?

through freebsd-update

pkg-static upgrade -f will upgrade or reinstall all the packages using statically build pkg(6)

(8)

freebsd-update can update only OS

why would i need this? do I?

that was minor upgrade

but now my zpool is not working

or something. i had to set old kernel

otherwise it didn't even boot

V-T60: did you read the release notes and upgrade the bootloader as described?

I'm reading Chapter 26, but don't understand much...

next big changes you predict to freebsd after pkg base?

does anyone know how to use this on FreeBSD? github.com/thesofproject/sof-bin

without that, my speaker wont active

freebsd.org - having trouble accessing it via IPv6

I did freebsd-update rollback and now my computer is not booting at all

full data loss?

writing something like 'module kernel required but not loaded'

rwp: the repo is switched only after the new one is build, but if you use old versions for failed packages you risk to run into inconsitencies with versions (well, iirc for libs pkg has a little bit of protection against that) (what if some dependency changed ABI and only that dependency is updated and the package you want to use is not because it failed to build and you put the old version in the

repo, similar to that problem with icu and prosody V-T60 had recently?)

now my issue is worse...

looks like i totally lost everything from the computer

after freebsd-update i couldn't run sway

but after freebsd-update rollback i can't even run bash

it crashes at the very beginning

writing something like 'module kernel required but not loaded'

i could clarify, if it had any sense

to me it looks like full data loss

i tried single user mod3

i tried single user mode. not booting*

now i have even three kernels...

hm, freebsd-update should have made a boot environment before updating, can you boot that?

boot environment?

before updating?

how is that?

where is that?

with root on zfs, the datasets are created in a way that you can have multiple versions (boot environments), so that you can switch back to an older version in case of problems when updating for example

i saw something about boot environments

but it shows 14.0

what on earth

i have never even installed such version, isn't that so?

what version did you install? what updates did you do?

i am 13.2 user

in could only do minor upgrade

because the program (that is freebsd-update)

only shown me that

it should have been 13.2 something i guess?

but sway didn't work

(hm, for freebsd-update upgrade you have to specify the release with -r, so you would have noticed)

i didn't specify.

but boot environments never appeared for me before

in the menu on boot up

now i see some default that tried booting

and 14.0. but how? i doubt any of them chosen will be working

after sway stopped working i tried to revert

and rebooted. now it is absolutely broken

Oh. I changed to 14.0

And now even gui is working

wait what...

% freebsd-version -k

14.0-CURRENT

what was i actually updating from and to

how and when did you install -CURRENT?

09:11 < V-T60> Should I expect SSL working fine on 13.2?

i don't understand...

during revert somehow?

I was updating 13.2 and reverted back... to 14?

insane. that is nonsense. i don't understand

or wait.

what is that? Fetching metadata signature for 13.2-RELEASE from update2.freebsd.org... done.

Probably I misunderstood.

but how? 14 was -CURRENT until november 2023

Look.

The following files will be added as part of updating to

13.2-RELEASE-p12:

try freebsd-version -kru

How can i upgrade 13.2 to 14?

% freebsd-version -kru

14.0-CURRENT

13.2-RELEASE

13.2-RELEASE

Triple output?

what does it mean?

you have a 14.0-CURRENT kernel, running a 13.2-RELEASE userland with a 13.2-RELEASE userland installed

Before it refused booted

Before it refused to boot*

I don't how that was ended up like this

I don't know how that was ended up like this*

Shouldn't I reinstall everything from scratch to get clean environment?

Or is that easy to arrange logically?

did you do anything with make inside /usr/src?

Oh, could be. I'm using this system so long time ago.

I had some troubles

With kernel modules not working

or something with kernel. that it prevented me from doing what i wanted

I don't remember what exactly. Probably I did something to modules. Probably kernel. It is even possible that nothing worked good for me and I didn't do anything

But I tried something for sure.

ok, first you should do a backup of everything you want to keep from that system, I guess a normal upgrade to 14.1 via freebsd-update should work, you already had problems there

I can move

to HDDs

I already had problems there?

Where?

On 14.1?

I thought that 13.2-something was tried?

then you have to decide if you want to try the freebsd-update route (-> latest -p for 13.2 -> 13.3 -> 14.0 -> 14.1) or just a reinstall to have a clean system

with using freebsd-update

I believe that with freebsd-update it asked me to proceed with going to 13.2-something

but i doubt that it is working

probably i didn't try whether reinstalling packages helps

with that trouble

(sway not booting)

but that was minor upgrade, wasn't that?

so i shouldn't have...

well, 13.2 is eol, so the pkg repo isn't up anymore

oh,okay

probably that's why i couldn't use?

the freebsd-update utility

no, freebsd-update is seperate and afaik it should always be possible to use, no matter how long the release is eol

but probably the packages are incompatible?

that i took with pkg upgrade

you only need base utilities for freebsd-update, ignore pkgs until you are on the release you want (login as root)

login as root?

So should I have upgrade again after reboot?

i could omit this one...

not sure now.

well, a reinstall is probably faster and it gives you a clean system, with freebsd-update you should switch to the latest -p release first and than go thru the releases

The process will require 956 MiB more space.

2 GiB to be downloaded.

Strange.

Didn't I upgrade everything before my manipulations?

how did i revert them...

probably i did freebsd-update before pkg upgrade?

i'll go with reinstall probably

you booted some boot environment, try bectl list what does it say for created for the active one?

that was 14.0 one

that i said non-working

i than tried to change the kernel too

not only environment

(but with previous environment none of the three kernels was able to but the system, though not clear why 3)

% bectl list

BE Active Mountpoint Space Created

14.0-CURRENT_2024-08-17_081749 N / 9.01G 2024-08-17 08:17

default R - 38.7G 2023-06-11 20:47

What is default and why is that so massive?

ok, so it seems that one is the one freebsd-update did before the update

14.0 current?

is my backup?

no, it seems like your system was pretty messed up before you tried the update, a 14.0-CURRENT kernel with a 13.2-RELEASE userland

so moving asap?

otherwise something might go even more wrong?

the space calculation there is a bit strange, as that other boot environment is a clone of default at some point in time its size count to bothand then you had some changes after that

as I already said, I expect that updating to 14.1 via freebsd-update carefully doing the full chain should work and fix your system, but reinstalling is probably faster and definitly has less risk

It seems that V-T60 had installed a 14.0-CURRENT kernel but was still running the 13.2-RELEASE kernel. Hadn't rebooted to it. A "uname -r" would confirm or deny this.

nimaje, A couple of issues about pkg repositories. One is that the build fails and skips packages and then is switched into place. That's when people can no longer install those important packages. But if there are failures and skips then that new repository switch looks more like a failure than an update.

Two is that yes an older program either pkg or self compiled might be broken by the base removal of a shared library or by an ABI breaking change. FreeBSD should but does not protect against ABI changes. I think it should. Other operating systems do protect against breaking ABI changes successfully. I think FreeBSD would be better if it did.

easy as pie on clang, I haven't checked lately but it's easy as knowing the target triple]

my friend's server has a triple of x86_64-portbld-freebsd13.1

well, for required libs, these are recorded in the package metadata somewhere and most of them are named .so.<so-version> where the so-version should get bumped on ABI changes, so I would hope that pkg had at least some protection against updating a lib, but not its reverse dependencies when that name changes

It's not always bumped though. I don't think there is any plan to ensure that it gets bumped.

I'm not even sure what decides what so-version even is

Maybe this is a variation on it. The problem often occurs with kernel drivers in ports. pkg is compiled against the oldest supported base and so kernel drivers in pkg are compiled against the oldest base kernel. But in point releases the kernel has changed ABI which breaks the previous kernel driver. And then it is months between breakage and pkg ports becoming "fixed".

I experienced that first hand with the radeonkms driver from ports in 13.0, 13.1, 13.2 working perfectly then in 13.3 this was broken. It only became fixed in pkg when the 13.2 base became EOL and the pkg build system moved to the 13.3 base for its build.

that does seem to be one advantage of the Linux philosophy, IMO

where they try to avoid breaking changes to the ABI

Agreed. It's a definite advantage on Debian/Ubuntu for example.

I mean, if you're using *x properly it shouldn't matter, as everything should always be rolled from source, but still, ain't nobody got time for that

But one might argue that the FreeBSD philosophy is to build it from source, as many long timers do, and that also avoids the problem.

you can prolly tell I'm not a FreeBSD "native" lol

(my main OS is actually Debian, to which I'm pretty much platform-agnostic but have mostly used it as a Linux distro)

I am a newcomer here myself! I have been using FreeBSD only since 12.

I've used it on and off. a friend uses it more than I do, and I've got access to all his servers

in fact, I'm currently hijacking one of his FreeBSD servers (with his permission) to do some encoding because my own rig is a bit unstable, especially in the summer

I have been trying to skill myself up with FreeBSD. I can do the basic things no problem but learning curve means it takes some time to get into all of the nooks and crannies of it.

when I was younger, one thing I wanted was a "real" (read: SVR4) Unix box. illumos is prolly as close as a mortal can get to that and it basically _is_ that. but I also became, like, I wanted to see if it could become "more free", though that necessitated the same kind of Theseus job that happened to BSD

so although I had a couple false starts I started writing something I called Strix, figuring I'd do what I could myself, and then use NetBSD and/or FreeBSD to fill in holes

as such, when I started coding, that's also the exact licensing terms I used

so that Strix code and BSD code could more readily be "blended"

now, I'm not that good a programmer.

I know a lot of evil C tricks though

I'm not saying I prefer SysV better than BSD (apart from, say, init(8))

though no doubt, BSD's init(8) could prolly be kitbashed into a SysV-style one

You were talking about Strix yesterday. Building an entire operating system is a large project. Something similar to base can be managed but when trying to do EVERYTHING it becomes a very large task.

you're not wrong

there's a reason I said the main focus is base, and a lot of the rest would need to be pinched

...mostly from NetBSD and FreeBSD ;)

I did decide a few things for outside

like, I'd have a full BSD-style userland _available_. (it would be in /usr/ucb, like SVR4, but it would be *modern* BSD, not 4.3)

and I'd prolly offer "traditional" Unix desktops by default - the CDE and OpenLook - but no reason I couldn't offer others (I main MATE)

an example: I have my own implementation of cal(1), but I could either use ncal(1), or a derivative of Reno's cal(1) (placing ncal in /usr/ucb)

the purpose of /usr/ucb was for bug for bug compatibility with ancient scripts on SunOS, not as an alternative userland to Solaris

I see it kinda like /usr/gnu on some OSes

isn't /usr/gnu an exclusively Solaris thing as well?

dunno, I would think it's a perfectly reasonable thing to do

basically, "Unitas" is intended to be a spiritual successor to SVR4

but aiming to have the freedom and freeness of BSD

does that mean I have to compromise one side for the other...yeah

SVR4 make was properly horrendous

I think Unitas would frankly need its make system to be bmake

I have no clue what this unitas thing is

I've been mentioning that I've been trying to create a "SVR4-like" system, and filling in holes in the part I *can* write ("Strix") with the BSDs

hence trying to figure out a way to build "kFreeBSD" or "kNetBSD" independently

I came up with the name "Unitas" because I wanted to suggest Unix without running afoul of their trademark.

(it's Latin for "unity")

I tend to see Solaris as the "model" SVR4, prolly because it's the most readily available

so ultimately what I'm trying to achieve

is a "cleaner" SVR4 based on some sort of "kBSD"

and then bring it all up to date

using FreeBSD and NetBSD as a model

hm, the FreeBSD vm I have is 13.1

is there a rule of thumb for how long it takes for a port (available via ports) to be available as a binary via pkg?

b_h, Welcome to the recent topic of the week! We have all been pondering that very question for various packages. Which one are you waiting for? Firefox? VLC? Another one waiting for the x265 dependency?

I still think h.265 is overrated

I still use h.264, but a lot of my encoding is still of SD material

I've become kinda infamous for my work on Dragon Ball Z a few years back

can anyone here reach the freebsd.org site via ipv6 ?

where I prepared some decent SD encodes but the Japanese audio I included with them was better than almost anyone had ever heard

while I can connect, I have no way to tell if it's via 6 or 4

thanks... Ishut offipv6 in FF and it works ok. I have seen this beforel Wireshark shows a lot of repeating packets

other sites besides freebsd.org have had this problem, something about MTU back then. debian.org still works fine. This is a somewhat recent problem with FBSD, maybe a cloud server doing it.

not that many well known ipv6 services that I can reliably test with

same

I only know... ip6only.me ?

I'll try that, thanks

yep that works fine. the problem may be MTU related to the SSL handshake again for a cloud server. Last time it was someone else I think (not freebsd.org )

rwp: curl :)

curl... well... "fetch" is in base and always available. (shrug)

ccurl is not necessarily for getting files, but it gives you LOTS of capabilities - I use it for lots of stuff where I want to attach files like "PUT" and naciguessing redirects, etc.

and iirc, it's BSD-style license too?

hm. it's an X-style license with a trademark clause (similar to the one in the 3 and 4-clause BSD license)

MIT license

unless latest changed

that's as of 2020, at least

freshports.org looks like ipv6

the source tree I've got

freshports.org/ftp/curl but I totally get keeping a working ports tree as-is for as long as you can

yeah fresjports works via ipv6, but not freebsd.org

it's not that I'm allergic to the GPL, per se

but its fanboys put me off lol

The same can be said of the reverse too. It's a shame that the two largest free software communities are at opposition to each other. It leads to a lot of wasted energy.

:P

"you can't do that!" "b—, I'm the G.D. copyright holder, I can do whatever I want with it!"

the funny thing is I don't think we'd HAVE FreeBSD without RMS

rwp: unfortunately, it's used by transmission daemon, so I can't use fetch as a stand-in

right now the daemon crashes with the current curl 8.9.1 package after a torrent is started

rwp it's a philosophy thing. One wants to allow closed source use, the other does not. I try to dual license anything I contribute (whether useful or nit) to cover all bases

it's fixed in ports but I don't want to bulk out my jail to build from ports

you might beable to revert to an earlier package version. not sure where that repo is stored online

you can 'make pakage from ports on the host, too. Make sure /usr/ports/packages exists first

then use pkg to install iit within the jail

er 'make package' I mean

anway YMMV with other hoops needed to make that work properly

I tried reverting to an older package but it was too old to work and I can't find an 8.9.0 arm64 pkg anywhere

er, amd64

forums.freebsd.org is working via ipv6 - just not the handbook docs and main pages I guess. ~2 weeks I think since it started

well need to do work now

random: just polished off a minipizza

C, ASM or get out lmao

that's just my personal opinion though

I mostly code in C when not doing server or client side web schtuff

but the C util I wrote that works like a shell command gets used a lot too

;p

a lot of what I do is stuff where people would be like "why would you ever want to do that?"

...because I do, got a problem with that? (not you)

Hello shbrngdo

heh yeah. old school coders

still running FreeBSD ?

👵🏻

C ftw

Rust ftL

:)

I actually started a few years ago trying to rewrite DOS ... and then Microsoft and IBM MIT'd MS-DOS/PC DOS 4

mzar - yeah - building ports for somewhat crammed at me update to 14.1 and latest port. Stupid web sites not support old firefox.

obviating most of my work

I am using 2+ year old ports tree

related to that, though, I'd thought of trying to rewrite diff(1) without reliance on either Ancient Unix or GNU...

I might actually be able to, because FC does basically the same thing

working well in VBox VM except mate panel crashes when I start X - not the panel, just everything on it. need it working before I "deploy"

shbrngdo: that's good to hear that you are still active in software world, have a successful build

building packages from ports, to install on 2 systems

ack thankx

last I checked, mwm didn't properly handle dual head which is a pity - I suppose dtwm (i.e., CDE) prolly doesn't either :/

well

CDE is a bit long in the tooth

but I like it

honestly, I think it'll need a lot of reworking of the code to modernize the guts

Licca: back in the day (circa 2007) someone was trying to recreate SGI's Indigo Magic desktop env on linux and others. Indigo Magic used Motif but was vastly superior to CDE in usability

I don't know it this effort ended up in packages somewhere

(the old nekochan (SGI enthusiast) forums (now defunct) had posts by the dev)

MaXX?

slash, 5dwm

I do have questions about its legitimacy but it would be a potential option. whatever source he released is apparently BSD

yeah this

inb4 fsn: "it's a Unix system, I know this"

lol

people thought fsn was Hollywood film magic lol

...and it was the real deal XD

yep, people stuck in dos or win *.* had no idea, nor did the mac or amiga people

hell, people in Solaris or AIX were prolly comparably blown away

probably yeah, xdm couldn't touch that

NeXT people must have been jealus :P

I used to have an O2 during my uni years. Wrote my thesis on it

s/jealus/jealous

21" SGI CRT

nice

I've got a Indigo2 R10K with SGI 21" CRT too

only Impact Solid gfx though

I am in physics, but this was from our Bio dept whom I helped with their molecular dynamics simulations (because all science is ultimately quantum field theory :)) and they donated that machine to me as thanks

lol

they used to have huge octanes

these days NeXTstep is called "macOS"

unfortunately I no longer have it

Octanes were such a cool arch

I have never seen such a striking computer

looking at an Irix source tree it looks like a System V/BSD hybrid, but not "SVR4"

xfs on even the o2 was jaw droppingly fast

I'm taking early 2000s

yeah XFS has always been bulletproof and fast

so long as the underlying drive held together it worked superbly

Licca: well the kernel had so many performance improvements (multi-cpu, large ram management, 3D and video, etc, NUMA management...)

hybrid SysV4/BSD wasn't unusual in the early 90s

AMIX from Commodore was that for ex

nearly pure SysV4 port but some BSD enhancements

SVR4 itself was a hybrid

brought in lots of issh from 4.3BSD

yep

which kinda informed my roadmap for Unitas: "SVR4" (a BSD-licensed clone) base, but then add in a lot of "BSD" elements, by which I mean FreeBSD and NetBSD instead of 4.3 ;p

but...haven't been able to figure out how to get off the ground yet.

Licca: why go with a 33+ year old unix though? Isn't there a more uptodate version of System V that is open source now?

old SCO/Caldera stuff?

well, yes, there's illumos.

ah yes

my idea was to create an up-to-date "spiritual successor" which is BSD-licensed

well, as far as possible

quite the endeavour

yeah

I started work on a userland, which I called Strix

based roughly on SVR4, and on Posix

but new code.

the idea being, I'd add either kNetBSD or kFreeBSD as its kernel, and the corresponding libc, and clang as the system compiler

and use BSD to fill in some gaps in my skills

for example, /bin/sh would be ash

mate-panel definitelyat the core of a major bug in Mate - ubuntu-mate.community/t/mate-panel-…e-errors-and-crashes-in-24-10/27438

I checked freshports, been there since March

So hpefully this will be "the bug" this time around, and no others

mode #freebsd +p

b_h, curl upgraded: 8.9.0 -> 8.9.1 just recently around Aug 5. Do you still have /var/cache/pkg/curl-8.9.0.pkg available? You would return to it until the problem is fixed.