vdamewood has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
navi has joined #osdev
theyneversleep has quit [Remote host closed the connection]
<nikolapdp>
wonder how hard it would be to add journaling to bsd 2.11
node1 has quit [Ping timeout: 250 seconds]
<zid>
"Dear diary, today someone wrote a BIG file to me, it was 400 entire characters!"
<zid>
journalling in the bsd 2.11 era
edr has joined #osdev
kanzure has quit [Ping timeout: 268 seconds]
kanzure has joined #osdev
<nikolapdp>
how cute
node1 has joined #osdev
<Ermine>
it's so fucking big
<nikolapdp>
btw tty.c is 35992 bytes large
<nikolapdp>
now that's a massive file
<Ermine>
is it from 2.11 bsd?
<nikolapdp>
yes
<nikolapdp>
part of the kernel
<Ermine>
how many lines does it have? how many XXX are there?
<nikolapdp>
1726 tty.c
<nikolapdp>
and no XXX
<Ermine>
wow, that's a surprise
<nikolapdp>
heh i've checked the tty.c i've downloaded on my laptop
<nikolapdp>
must be a later version
<nikolapdp>
becase there are two XXX
<zid>
well it is tty code, you'd expect some ascii art porn
<zid>
two is about right
<nikolapdp>
kek
foudfou has quit [Remote host closed the connection]
foudfou has joined #osdev
joe9 has quit [Quit: leaving]
Arthuria has joined #osdev
MiningMarsh has quit [Ping timeout: 240 seconds]
foudfou has quit [Remote host closed the connection]
foudfou has joined #osdev
MiningMarsh has joined #osdev
apokryptein has joined #osdev
m3a has quit [Ping timeout: 264 seconds]
node1 has quit [Quit: Client closed]
npc has joined #osdev
Arthuria has quit [Ping timeout: 268 seconds]
npc is now known as obrien
obrien has quit [Remote host closed the connection]
npc has joined #osdev
dalme has joined #osdev
foudfou has quit [Remote host closed the connection]
foudfou has joined #osdev
foudfou has quit [Remote host closed the connection]
foudfou has joined #osdev
Arthuria has joined #osdev
Arthuria has quit [Killed (NickServ (GHOST command used by Guest684531))]
Arthuria has joined #osdev
<Ermine>
New posix was released
<nikolapdp>
has it
<nikolapdp>
*was it
<nikolapdp>
anything fun there
joe9 has joined #osdev
vdamewood has joined #osdev
foudfou has quit [Remote host closed the connection]
foudfou has joined #osdev
<kof673>
does it use UFS? "journalling" ....at some level it is "just" how much work to port a later UFS [*] [*] == *************************** == (potentially many other details)
<kof673>
three-star footnotes <runs>
<nikolapdp>
it is ufs yeah
<nikolapdp>
and not really
<nikolapdp>
quite different adding a simple journaling mechanism compared to actually porting newer bsds
<zid>
$676
<zid>
not bad
<nikolapdp>
not to mention soft updates that some newer UFSs are fond of
<nikolapdp>
that's that zid
josuedhg has joined #osdev
<zid>
what's what
<kof673>
"quite different" okay, that makes sense, ^^^^ [newer UFses] that is what i was getting at ^^^
<kof673>
bsdland i do not see the term "journalling" used much
<kof673>
"we have x instead"
<nikolapdp>
zid: what's $676
<zid>
posix 2024-1
<nikolar>
ah nice
<nikolapdp>
is there a draft or something
<zid>
you wish
<zid>
opengroup or someone will put a copy on their website soon I bet
<nikolapdp>
legally i am sure :P
<zid>
nah it's also the open group technical standard
<zid>
as well as the $676 ieee standard
<zid>
like iso C vs ansi C
<nikolapdp>
yeah yeah
<nikolapdp>
we'll see soon enough
<zid>
and opengroup hve 2008/2013/2016/2017 on their site
<nikolar>
Never really figured out the disklabel stuff
<nikolar>
Need to play around with bsds more
<geist>
so if yo uadd a new disk, set up the disklabel, put everything in 'b' or at least osmething
<geist>
you should be able to add it
<nikolar>
Yeah nice
<nikolar>
And just put it in fstab
<geist>
i generally move my home over to a second disk anyway just so that i can trash the original one and still have home left
dalme has quit [Read error: Connection reset by peer]
<nikolar>
Heh
<nikolar>
I should check that
<geist>
anyway you'll want swap. even with 4MB (max pdp11 would take) it'll start swapping pretty fast
<geist>
as soon as you start compiling things and whatnot
<nikolar>
I run it with 4mb
<geist>
swapping in a good way, swapping out stuff as memory pressure increases, etc
<geist>
swap was much more essential back in those days, it was just assumed you were overcommiting the machine a lot
<nikolar>
And randomly I started getting messages like coremap something something lost clicks
<nikolar>
So maybe it's trying to swap?
<geist>
since memory was expensive and everything being mega responsive was a luxury
<nikolar>
Yeah makes sense
<nikolar>
You've payed for the whole machine, you'll use the whole machine
<geist>
also 4MB was a maxed out machine. you could realisitically run with like 1MB
<geist>
actually my PDP11/53 has 1.5MB on the main board, i only upgraded it to 2.5 with an additional board
<nikolar>
Heh nice
<geist>
but i did most of my upgrading from 450-481 with 1.5 and though it wasn't mega swapping, it was definitely using it
<nikolar>
I don't think I can find a compact pdp11 near me unfortunately
<geist>
you could get ahold of one of the soviet clones, which would be amazing
<nikolar>
I could yeah
<geist>
i wonder if theres any sort of secondary market for those
<nikolar>
But no mmu :(
<geist>
ah, early era pdps
<geist>
they were only 64k total?
<nikolar>
I did find an elektronika from Ukraine for relatively cheap
<nikolar>
Yeah
<nikolar>
But unfortunately can't run bsd on it heh
<geist>
probably small industrial control versions
<geist>
wouldn't be surprised if the PDP11 that bootstrapped early VAXen was something trimmed down like that
<nikolar>
I think they are kind of commodores of the ussr
<nikolar>
in a way
<geist>
since it only really needs to run one piece of firmware
<geist>
yah pdp8s were kinda that way. later model versios running OS/8 is really pretty much a CP/M or DOS like experience
<nikolar>
Kek nice
<geist>
you can directly see how CP/M (and thus DOS) was derived from it. gary killdall was a DEC user before doing CP/M
foudfou has quit [Remote host closed the connection]
<geist>
a lot of the same concepts are there, even things like PIP.EXE (a CP/M thing that DOS generally replaced)
<geist>
drive labels, etc
<nikolar>
Yeah I am aware of some of those
foudfou has joined #osdev
<geist>
but only TSS/8 on PDP8 (very late addition) even attempted some sort of multi user/multitasking
<nikolar>
Heh maybe I can just buy a bunch of pdp11 boards off of eBay and recreate the whole machine from the pieces
<geist>
otherwise PDP8s were pretty much always single user at a console
<geist>
you can, that gets pretty pricey, but i see a lot of PDP11/53 or so motherboards for sale
<nikolar>
Heh yeah maybe over there
<nikolar>
I'm half a world away
<geist>
be careful, make sre they have the LSI-11 on them. some of the cheaper ones are missing the cpus if you're not paying attention
<geist>
but yeah you could probably build a barebones one from a QBUS backplate and then a single motherboard
<nikolar>
Heh yeah
<nikolar>
Maybe could be fun
<nikolar>
At some point
<geist>
the 11-23/53/75/83 were the last models made and pretty standalone. one main motherboard with everything + memory
<geist>
but no controllers for disks. but a console
<geist>
s/75/73
<nikolar>
Well ideally I'd like a disk emulator anyway so that's maybe fine
<geist>
yah i have it the expensive way: i have a scsi QBUS card with two channels, and i currently moved away from using the spinny disk and have a zuluscsi plugged in
<geist>
but the scsi card is fairly expensive
<nikolar>
Unfortunate
<nikolar>
I wonder if someone maybe a pdp11 recreation from modern chips like some of the other computer from that period
<nikolar>
Granted those were simper
<geist>
i thought i remember there being some mention of someone that recreated a LSI-11 on fpga
<geist>
but you start to get pretty close to why not just run an emulator, etc
<geist>
however i *do* recommend a pidp11
<geist>
basically the same as what you're doing with a front panel
<geist>
so yoyu get the experience
<geist>
very nice to have on the table next to your monitor, typing away
<nikolar>
Heh maybe
<nikolar>
The FPGA one needs a program on a PC for all io
<nikolar>
The only thing on FPGA is basically the memory and the CPU
<nikolar>
Not standalone
<nikolar>
But I feel it could be a fun project to build a pdp from chips if such a thing existed
<geist>
yah
<geist>
mind you it is a pretty complicated arch, so you're not goint to be bilding it from TTL or something without needing a few gigantic boards (ie, like real early PDP11s)
<geist>
would be interesting to get real LSI-11s and build up from that
<geist>
no idea what the bus interface looks like on those directly at the pinout of the chip
<bslsk05>
photos.app.goo.gl: Dynamic Link Not Found
bliminse has quit [Ping timeout: 256 seconds]
<nikolar>
Btw what are the differences between the pdp-11/5x, 7x, 9x variants
Arthuria has quit [Ping timeout: 240 seconds]
<Ermine>
geist: cool, thank you for sharing!
bliminse has joined #osdev
<geist>
yeah it's a pretty neat little computer history museum there on the edge of cambridge
voidah has quit [Ping timeout: 240 seconds]
<geist>
i haven't been since the pandemic, but i used to visit every year in the late 2010s when i would usually come over for the yearly ARM partner meeting they have in the summer
<geist>
i think it's still fine and doing well
<GeDaMo>
I'm always amused when I'm reminded that Lyons was a computing pioneer
<dostoyevsky2>
nikolar: who actually thought that 64K
<dostoyevsky2>
per process would be a good idea
<zid>
sounds like a great idea to me
<nikolar>
Kek I think it was luxury then
<heat>
does 2.11BSD swap when switching processes or was it way past that?
gbowne1 has quit [Remote host closed the connection]
<nikolar>
heat: it could
<nikolar>
I don't have swap though so doesn't swap for me
<heat>
oh it's a mode you can turn on?
<zid>
That's like, half a super nintendo
gbowne1 has joined #osdev
<nikolar>
You can just no have a swap
<zid>
per user
<nikolar>
Though ps dies without swap lol
<dostoyevsky2>
nikolar: 64k swap file should be fast
<heat>
yeah but i mean the actual swapping when swapping processes
<nikolar>
No swap swap, just swap devices
<heat>
like UNIX v1
<heat>
swap swap
<nikolar>
I'm not sure what you're referring to
<kof673>
<with apologies to pages the sizes of registers jokes...inserts entire RAM the size of a register joke>
<dostoyevsky2>
nikolar: weren't the ps programs back then reading the structs in ?kernel memory
<heat>
yes
<nikolar>
Yup
<nikolar>
It's reading /dev/kmem and /dev/swap
<nikolar>
But I don't have /dev/swap so it just doesn't work
<dostoyevsky2>
I wonder why it runs out of memory... maybe needs to copy all the kernel memor into userland?
<nikolar>
What do you mean "runs out of memory"
<nikolar>
What does
<kof673>
"why it requires swap"
<dostoyevsky2>
nikolar: "Though ps dies without swap lol "
<heat>
it might require swap just cause
<dostoyevsky2>
oh
<nikolar>
ps requires swap to read it
<zid>
is there not a version
<zid>
or parameter you can pass
<nikolar>
It's not running out of memory
<heat>
e.g ps will need to look into swap to get certain data structures
<zid>
to tell it to not try to print swap stats
<nikolar>
zid nope
<zid>
time to #if 0 some stuff
<nikolar>
Tired to patch it but I just get bogus processes
<nikolar>
*tried
<heat>
don't forget they could swap stacks and struct user back then
<dostoyevsky2>
heat: Maybe one should open a couple of issues in the netbsd2.11 github repo
<zid>
what's it looking in the swap for, swapped out processes?
<nikolar>
I wonder why it's getting bogus processes when I literally can't seap
<dostoyevsky2>
-net
<nikolar>
Since there's no swap device
<heat>
programming error?
<nikolar>
zid: eh haven't looked that deep into it
<dostoyevsky2>
nikolar: Doesn't Linux also show odd process in ps, some of which are running inside the kernel?
<nikolar>
I'm not referring to that
<nikolar>
I'm seeing bogus pids and commands etc
<nikolar>
It's just reading random data
<dostoyevsky2>
hmmm
MrCryo has quit [Quit: MrCryo]
<heat>
those are not "odd processes"
<nikolar>
Also no in kernel processes or threads
<nikolar>
Kek
<nikolar>
Who said odd
<heat>
dostoyevsky2 said it
<dostoyevsky2>
heat: write a phd thesis about it
<nikolar>
Ah right
<heat>
about what
<dostoyevsky2>
heat: The effects of short term memory loss
<dostoyevsky2>
nikolar: I guess ps could be compiled with the wrong kernel struct version to list the processes... I think that might be part of the reason they stopped doing it, besides security
<nikolar>
I don't think so but maybe
<zid>
ah fuck, I sneezed, now I'm in a lot of pain
<nikolar>
Sad times zid
foudfou has quit [Remote host closed the connection]
foudfou has joined #osdev
<dostoyevsky2>
nikolar: what kind of init does bsd2.11 have? Just a shell script?
<heat>
systemd
<nikolar>
init is a c program that runs a shell script
<nikolar>
/etc/rc is where all the action happens
<nikolar>
I think init just reaps the children
<heat>
no i'm pretty sure it's systemd
<dostoyevsky2>
Sounds familar
<heat>
nikolar do systemctl enable networkmanager
<heat>
make sure tls is set up right in systemd-resolved too
<dostoyevsky2>
heat: which version of libxz is the bsd2.11's systemd using?
<dostoyevsky2>
nikolar: and you have to use telnet to "ssh" into the system? Or via serial?
<nikolapdp>
heat: systemctl: Command not found.
<heat>
sadly jia forgot to compromise 2.11BSD's packaging system, the tarball
<nikolapdp>
dostoyevsky2: since it's a vm, it exposes the serial lines over telnet
<nikolapdp>
so you can open multiple ttys
<heat>
hmm... what was the first unix with a virtual terminal?
<dostoyevsky2>
heat: Does bsd2.11 even support ifuncs?
<nikolapdp>
doubt it
<heat>
not yet, but GNU/k2.11BSD glibc is on the way
<dostoyevsky2>
I asked in OpenBSD they don't even know what ifuncs are today
<dostoyevsky2>
#openbsd even
<nikolapdp>
kek the init.c is a tiny bit more fancy than just reaping childrem
bauen1 has quit [Ping timeout: 264 seconds]
<nikolapdp>
btw geist, vmstat claims i have 2980kb free
<nikolapdp>
doesn't sound like i am anywhere near swapping
<heat>
a whole 3MB, insane!
<nikolapdp>
indeed
<heat>
what's the max ram it can swallow before panicking?
lukflug has joined #osdev
<nikolapdp>
no clue
<nikolapdp>
kind of annoying to test since each process cannot possibly do more than 128k
<nikolapdp>
also i might have a corrupted directory or something in root
<nikolapdp>
it shows up as just ?
<nikolapdp>
(:
<heat>
run zfs
<nikolapdp>
i am unable to can
Turn_Left has joined #osdev
<nikolapdp>
i do wonder how hard it would be to add journaling to ufs here
<nikolapdp>
there's a ufs_bio.c which is probably where most of the code would need to go
<dostoyevsky2>
hmmm... reading through bsd2.11's init.c ... there is no syslogd?
<nikolapdp>
there should be
<nikolapdp>
i meant it prints it at boot
<dostoyevsky2>
oh, there is...
<zid>
good news, I didn't die from sneezing, barely
<nikolapdp>
oh you're still here
<lukflug>
Update on the wiki fork: the domain https://osdev.wiki/ now seems to work properly.
<bslsk05>
osdev.wiki: Expanded Main Page - Operating System Development Wiki
<dostoyevsky2>
I mean if you assume that /etc/rc has some kind of start stop functionality, then you don't need that in init... but it might be nice to restart services automatically I guess...
<zid>
There was a lot of blood involved but I appear to have survived
<nikolapdp>
kek
<nikolapdp>
lucky you
<nikolapdp>
lol there's a kern_xxx.c
<heat>
kernel porn!!
<nikolapdp>
mostly networking stuff, and for some readon, reboot
<dostoyevsky2>
it contains reboot()
<zid>
kernal porn.c?
<zid>
or xxx as like, a weird way to write misc
<nikolapdp>
i am leaning toward the second option
<dostoyevsky2>
would you call a function `execit' for when you want to exec something?
<heat>
no
<nikolapdp>
i don't know, maybe
<nikolapdp>
i am too lazy
<heat>
it's like calling a variable the_variable
<dostoyevsky2>
Seemed to be the thing with the bsd2.11 calling it execit
<dostoyevsky2>
+devs
<nikolapdp>
you'll find a lot of great names throughout
<heat>
musl coding style
<dostoyevsky2>
> sleep(10); /* prevent failures from eating machine */
<nikolapdp>
where's that from
<nikolapdp>
lol estabur is the name of a function
nikolapdp has quit [Remote host closed the connection]
<dostoyevsky2>
nikolar: init.c's execit
antranigv has joined #osdev
nikolapdp has joined #osdev
<nikolapdp>
oops
bauen1 has joined #osdev
<nikolapdp>
wrong terminal
<dostoyevsky2>
nikolar: vtxxx?
<nikolapdp>
emulating vt100s but not what i meant
<nikolapdp>
^D is for scrolling in vi(m)
<nikolapdp>
but it also closes stdin
<dostoyevsky2>
nikolar: ^F doesn't work?
<kof673>
that "estabur()" name got around, wherever it started
<dostoyevsky2>
> typedef signed char int8_t; typedef signed int int16_t; typedef signed long int32_t;
lukflug has quit [Quit: Leaving]
<nikolapdp>
maybe i should add proper 64 bit time before the y2038
<zid>
can you upstream the patches
<zid>
to.. oracle?
<nikolapdp>
why oracle
<zid>
who the fuck owns unix now
<zid>
and bsd
<nikolapdp>
the same guy has been maintianing the 2.11 since it came out
<zid>
nobody knows
<nikolapdp>
^ zid
<zid>
what
<CompanionCube>
definitely not oracle
<mjg>
nikolapdp: old unix is all avoidable linear searches
<nikolapdp>
the same guy that backported the crapload of 4.x has been maintaining the 2.11 for the past like 30 years
<mjg>
pretty f-ing f-eing
<mjg>
where
<zid>
is he mr. oracle
<nikolapdp>
no
<nikolapdp>
that i know lol
Left_Turn has quit [Ping timeout: 268 seconds]
<CompanionCube>
i think the current owner of unix (except the trademarks) is opentext, who acquired micro focus?
<mjg>
i don't know any of these names
<nikolapdp>
zid: Since the original release of 2.11BSD, Steven Schultz has been maintaining the system
<zid>
nikolapdp: Okay so set a calendar entry for december 2037 to switch 2.11BSD to full 64bit time
<nikolapdp>
lol i don't think i can manually set the date to a year after 2000
xenos1984 has quit [Read error: Connection reset by peer]
<nikolapdp>
date command only accepts two digits for the year
<nikolapdp>
lol
<zid>
can you set via unix timestamp?
<nikolapdp>
erm, nope
<nikolapdp>
it listens to ntp though
<zid>
or just write your own `date` I guess
<nikolapdp>
too much work
<zid>
sys_settime(rand(), rand());
<zid>
just run that until the date is correct
<nikolapdp>
yeah brute force the time
<zid>
bogodate
<nikolapdp>
basically kek
foudfou has quit [Quit: Bye]
foudfou has joined #osdev
<Ermine>
hey heat
m3a has joined #osdev
<heat>
hi
xenos1984 has joined #osdev
<heat>
mjg: linear searches are OPTIMAL when you have like 3 processes mon
<nikolar>
up to 20 per user
<nikolapdp>
and 4 users lol
<heat>
yeah but how many do you have?
<nikolapdp>
oh i have a few actually
<nikolapdp>
a few shells, irc client, vmstat, iostat, netstat, vi
<nikolapdp>
i might be getting pretty close to the limit heh
<mjg>
heat: some of these searches were PESSIMAL tho
<mjg>
they do them *everywhere*
<mjg>
not when you just have 3 fuckin' items
<heat>
i know
<heat>
but if there's a system you can excuse it's 16-bit unix
<heat>
having them on the VAX unixes is way worse
<heat>
or fucking linux
<nikolapdp>
you can excuse a lot of things on a 16-bit unix
<nikolapdp>
because you're working with bloody 64k of address space
<heat>
totes
<heat>
i don't think i'd be able to write 16-bit code
<mjg>
no c++ compilers for 16-bit?
<nikolapdp>
kek
<nikolapdp>
gcc does support pdp-11 as a target
<nikolapdp>
even for c++
<nikolapdp>
hello
Turn_Left has quit [Read error: Connection reset by peer]
<mjg>
:O
<mjg>
heat you would be able to write 16-bit code
<nikolapdp>
and you could presumably use it for rust once that's working in gcc properly
<mjg>
?
<mjg>
rust is llvm-based
<nikolapdp>
there's gccrs
<nikolapdp>
get with the times
<FreeFull>
There's actually two separate efforts to make rust work with gcc
<mjg>
at the ame time?
<mjg>
is that fearless concurrency?
<nikolapdp>
kek
<nikolapdp>
there's one proper upstream gcc frontend
<nikolapdp>
and the other that's using libgccjit within the rustc
<mjg>
i guess that's gcc fighting for continued relevance?
<FreeFull>
One is for using rustc with gcc, and the other is writing a whole new rust compiler
Matt|home has quit [Quit: Leaving]
<nikolapdp>
the libgccjit is completely unafiliated with gcc
<nikolapdp>
i think the guy just decided to do it
<mjg>
not that c/c++ is gonna disappear overnight, but if people are saddled with rust (+llvm) they mas well switch the toolchain completely
<heat>
oh god why are we talking abour rust now
<mjg>
it was few hours since last time
<nikolapdp>
we aren't even shitting on rust (yet)
<nikolapdp>
what are you complaining about
<mjg>
this is now a rust channel
<FreeFull>
I wonder what mrustc uses for codegen
<heat>
Mutabah would know
* Mutabah
is away (Sleep)
dgz has quit [Remote host closed the connection]
<mjg>
is Mutabah rate-limiting
<heat>
apparently
osdev199 has joined #osdev
<heat>
say what you will, rust and C++ actually give you decent data structures
<heat>
instead of C which (in POSIX XSI only!!) gives you bad binary tree and bad global hash table
<nikolar>
Ok and?
<heat>
it's good
<mjg>
c is a shite language
<mjg>
who tf is denying it
<heat>
and generally avoids silly linear searches
<dostoyevsky2>
rust probably already has a kernal crate, so you can just write your own OS in 4 lines of code
<mjg>
ye, it's basically download a linux image and boot in qemu
<mjg>
all wrapped with unsafe { }
<nikolar>
Fearlessly unsafe
<gog>
me too
<heat>
gog c#pilled
<gog>
i have suspended that endeavor
<heat>
oh dang
<gog>
until i learn more about the compiler
<heat>
consider not learning more about the compiler
<heat>
learn more about our lord and savior the nix package manager first
<gog>
i can't do a fake RAII pattern with using blocks because it wants IDisposable and claims that any object that i implement IDisposable, the zerolib mode doesn't build it
<gog>
says it's not implicitly convertible
<gog>
works fine when i compile against clr
<gog>
so i think the real fact is that everything will be unsafe and i would have to be manually managing memory anyway
<nikolar>
heat when can we expect nix to be reimplemented in rust
<mjg>
8(
<gog>
so why would i use c# and manually manage memory
<gog>
the whole idea is that i don't want to manage memory
<heat>
rust community don't make up a shitty awful project name for a project challenge [IMPOSSIBLE]
* gog
bite
<mjg>
gog: that's my schtick!
<gog>
:3
<heat>
colon three
* mjg
brings a neighbour's pitbull real quick
<nikolar>
Lol just read the comments
<gog>
i don't want to manage packages anymore either, that's why--no, too obvious
<mjg>
package management is what you do on linux
<mjg>
on a chad os like windows programs are just there
<mjg>
bill: 1; linus: 0
<heat>
flatpaks are based
<mjg>
is that so-hell?
<heat>
i haven't heard of flatpak hell
<gog>
appimage
<mjg>
just before you were born a common point raised against windows was that all programs come with their own .dll versions bundled in
<heat>
.so-hell is when you update one of your glorious system packages and you break some random package that depended on the .so
<heat>
they still do, it works fine
<mjg>
referred to as dll-hell
<mjg>
years later when linux gained real traction people learned the hard way that everything in the same version does not work
<mjg>
so you got things like snaps on ubuntu
<heat>
did they?
<mjg>
or the aforementioned flats
<gog>
sharing code is bad anyway
<gog>
communist bullshit
<mjg>
so i think it is amusing
<mjg>
ye commie .sos
<heat>
the linux package manager enjoyers would say all software should be FOSS why are you not running FOSS you weirdo
<Ermine>
snaps bad
<mjg>
vs capitalist static binaries
<heat>
FOSS *AND* a system package
<mjg>
heat: having all your .sos bundled does not mean you are not open
<mjg>
mofo
<mjg>
just that you don't use what's shipped with the system
<heat>
software broken by a .so bump would not be open, probably
<Ermine>
1) these are squashfsen that get mounted at boot time. Which makes boot times absymal
josuedhg has quit [Quit: Client closed]
<mjg>
Ermine: ye the implementation is.... peculiar
<mjg>
but the fundamntal concept of separting that prog from the rest of the system is pretyty standard
<mjg>
i only found out about the squash thing ater trying to install neovim 8S
<heat>
OH so you're a DLL hell fan
<heat>
okay
<heat>
fwiw, i'm a DLL hell fan too
<Ermine>
2) Awful apparmor policies that prevent apps from accessing anything but $HOME
<FreeFull>
Package management should be done using Prolog
<heat>
steam with the steam runtime (a full debian runtime) never broke for me
<heat>
and i use fucking arch linux
<heat>
btw
<nikolar>
bte
<nikolar>
btw
<mjg>
now i am somewhat curiosu what LINUX DEVELOPERS use
<mjg>
and not posers
<Ermine>
openSUSe suggested me uninstalling steam due to some conflict with glibc
<mjg>
:d
<heat>
linus uses fedora
<FreeFull>
It didn't suggest uninstalling glibc?
<heat>
that's all i know
<mjg>
you serious?
<heat>
yes
<mjg>
that would be pretty nasty
<mjg>
how old is that info
<heat>
not very
<Ermine>
Lennartie uses Fedora, I'm pretty sure
<mjg>
i don't expect people to stallman their way
<FreeFull>
Linus has also been trying Asahi
<mjg>
lenny is at microsoft now
<mjg>
so probably windows ;d
<Ermine>
idk, but then wsl
<heat>
it is well known that when he finds compiler breakage he says he has a recent fedora gcc toolchain
<mjg>
wsl is why
<Ermine>
or that cbl-mariner thing
<mjg>
would be my gueszz
<dostoyevsky2>
it's just a question of time until systemd includes <windows.h>
<heat>
probably ubuntu honestly
<heat>
i don't think they'd force wsl on people
<heat>
otoh it's microsoft
<mjg>
ms gonna ms
<heat>
it could be very entrenched
<Ermine>
btw, who even uses cbl-mariner
<dostoyevsky2>
systemd with recall default opt in confirmed
<mjg>
you have to connect to the ldap and other shit they have there somehow
<mjg>
sharepoint and fuck
<heat>
Ermine: servers
<mjg>
i doubt they support linux workstations for this stuff
<Ermine>
thats understandable
<Ermine>
but i wonder if anyone here uses it
<heat>
uses what
<Ermine>
cbl-mariner
<mjg>
tf is that name anyway
<heat>
i'm 99% sure it's not really usable
<heat>
like, it's now renamed to azure linux
<mjg>
maybe it is on their cloud?
<heat>
that tellsyou something
<heat>
yes
<Ermine>
oh i didn't know
<dostoyevsky2>
microsoft systemd would be nice name
<heat>
it'd be like daily driving RHEL
<Ermine>
well ok, seems like they made a distro for their own needs
<heat>
but with way worse testing
<heat>
or daily driving, i dunno, ubuntu server
<mjg>
my entire usersapce gets away with the following: rustc, cargo
<heat>
>usersapce
<mjg>
kernal
<heat>
typo lol seen dont care didnt ask
<Ermine>
usersapce/kernal
<mjg>
that is the split, yes, Ermine
<heat>
onyx
<Ermine>
I want this to be a topic
<dostoyevsky2>
> According to early Commodore myth, and reported by writer/programmer Jim Butterfield among others, the "word" KERNAL is an acronym (or, more likely, a backronym) standing for Keyboard Entry Read, Network, And Link, which in fact makes good sense considering its role.
<nikolar>
Lol what network
<nikolar>
It's a commodore
<heat>
mjg: btw gothamchess almost got a GM norm
<heat>
was half a point away
<dostoyevsky2>
> Surprisingly, the KERNAL implemented a device-independent I/O API not entirely dissimilar from that of Unix or Plan-9, which nobody actually exploited, as far as is publicly known.
<nikolar>
Yes I know what that means heat
<heat>
what
<nikolar>
What
<heat>
what do you know the meaning of
<nikolar>
The words you said
<heat>
good
<heat>
words sometimes have meaning
<heat>
other times you forge groups of words to give them even more meaning
<kof673>
+points for double meaning forge pun
<mjg>
heat: screw that guy
<heat>
haha found the botez fan
<nikolar>
heat: and other times you get a random mumbo jumbo from an llm
<mjg>
screw them too
<mjg>
(they have a beef or something?)
<heat>
no
<nikolar>
But in a different way I assume
<mjg>
than what's that about
<heat>
are you a kramnik fan mjg
<mjg>
kramnik is a unix geezer in chess mon
<heat>
he thinks everyone's cheating, it's hilarious
<Ermine>
> which nobody actually exploited --- Do you want to hear a joke about the Elusive Joe?
<mjg>
it is plausible ude is genuinely mentally ill though
<mjg>
i mean take a look at bobs fischer
<mjg>
i like ben finegold fwiw
<heat>
expected
<mjg>
OH
<heat>
mr. taking the piss
<mjg>
terrible
<mjg>
you may be misusing mr. though
<heat>
why
<mjg>
the only legit use is sarcastic, mr linux dev
<heat>
just because i write bad linux it does not mean i'm a linux dev