Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
BlankSystemDaemon
Mar 13, 2009



Sapozhnik posted:

system76 is kind of a lovely actor in the linux space

https://blogs.gnome.org/christopherdavis/2021/11/10/system76-how-not-to-collaborate/

not as bad as some of the stuff ubuntu does but they're still quite parasitic and spread lies about other projects on top of that.
i wonder how many times we have to go through this before we learn that corporations are not the friends of opensource

Adbot
ADBOT LOVES YOU

BlankSystemDaemon
Mar 13, 2009



Sapozhnik posted:

"open source" itself is a corporate neologism designed to domesticate the free software movement and counter some of its more anti-capitalist tendencies
and the free software movement leads to open sores, so it's full circle

BlankSystemDaemon
Mar 13, 2009



Sapozhnik posted:

"open source" itself is a corporate neologism designed to domesticate the free software movement and counter some of its more anti-capitalist tendencies
it's kinda late to be adding it now, but it's pretty safe to assume that if i'm talking about opensource, i mean the kind of source code that's so open a line containing the ntty struct from 1980 (before the free software movement existed) can survive to this day completely unaltered or slightly changed.

SYSV Fanfic posted:

Looking at system 76's website I'm left with one question: Will hardwood panels replace tempered glass in hi tech fashion? My HAF 912 is still chugging along, but the black quasi-militaristic aesthetic makes me feel pretty dated, kinda like an avacado green fridge.

Should I hold out for one glass and one hardwood panel on either side? Are they refinishable so I can make it match my dad's 1976 pioneer stereo reciever?
are you lgr?

that being said, i do think there's room for a lot more interesting designs than exist in the current market

BlankSystemDaemon
Mar 13, 2009



finally you're talking sense!

BlankSystemDaemon
Mar 13, 2009



lol cabal

BlankSystemDaemon
Mar 13, 2009



I'm insomniac and bored out of my mind, so have a lot of :words:

I think there's a fundamental bit of philosophy that a lot of copyleft proponents miss (intentional or not, though I think it's mostly the latter), which is that to a lot of people working on copyfree licenses (not just BSD or MIT, but all of them), the important part isn't that someone else contributes back, that they get kick-backs, or any of a number of other reasons that copyleft proponents often talk about.

Instead, it's that there should be at least one good implementation of something that should be available for everyone to use, instead of everyone having to implement their own thing - in the same vein that everyone isn't expected to be able to correctly implement cryptography.

In the specific cases of Microsoft and Apple, there's some history there that I'm sure there are lot of people that don't know about, so here's the short of it:

Microsoft didn't "steal" the FreeBSD stack. Spider Systems (based, at the time, in Edinburgh and which has since been swallowed up by Intel) was a company that Microsoft contracted to develop a TCP as well as a STREAMS (the native UNIX System 5 networking) stack(s).
Since it's proprietary, not a whole lot is known, but from what I remember at the time, a pre-release version of one of the early products at one point got released/leaked, which didn't have symbols stripped - and from that it was learned that the TCP stack had FreeBSD headers, and a few of the userland utilities (such as tracert) had a FreeBSD license embedded in them).
This, combined with Microsofts use of FreeBSD for Hotmail (because the company that owned Hotmail before Microsoft bought it used FreeBSD, until Microsoft made ISS capable of handling the load for the servers, some 5 years later) as well as the Windows Update servers, led to rumours of Microsoft using a lot more code than there's ever been any actual evidence for.

As for macOS, that's got an even longer history. It all, predictably enough, starts with NeXT. For reasons forgotten to history, it integrated part of BSD out of CSRG with a Mach kernel. Then, NeXT did its NeXT thing and Steve Jobs ended up at Apple with a lot of the people who'd worked on NeXT.
This led to what would eventually end up becoming Mac OS X, which was itself a remarkable shift away from Classic Mac OS, which would, partly because of previous experiments with Linux in the form of MkLinux, also not just have POSIX compatibility, but full on SUS compliance (because that was still somewhat of a selling point at that particular time).
So since the developers responsible were already familiar with the Mach kernel and BSD subsystems, they yoinked code from the biggest descendant of BSD (which was and is FreeBSD), hired some of FreeBSDs developers (Jordan Hubbard being the most prominent, but others too, as well as several others later on and even as recent as just a few years ago).
This leaves the question of what FreeBSD code is in macOS - and perhaps more importantly, how much of it is used. Firstly, there's a bunch of command-line userland tools (although many have been modified by Apple subsequent to being imported, and others either have been replaced or are being replaced with copyfree-alternatives). I honestly don't think these count for much, since only powerusers tend to use these, and while Apple is popular among powerusers, there are far more regular users who simply don't use the terminal.
Then there's the parts of the kernel, where there are three "big" uses: the netstack (as mentioned elsewhere, because it really has been used widely), the process model, and the VFS. Obviously, the process model and VFS get used all the time, and the netstack is hardly not going to see use considering almost nobody uses any system today without networking. However, even all-told, those three parts still only account for a tiny part of XNU, and the userland utilities are only a tiny part of Darwin.
There's also the elephant in the room, which is all the parts that make up macOS, which aren't part of Darwin: The entire UX in the form of Quartz (including compositing), things like Metal, Core(Video|Audio|Image|Animation|OpenGl|Graphics), the entire driver interface as well as every single driver, and too many other things to mention (if you're really curious, there's a CCC talk by "Lucy" from 2007 called Inside the Mac OS X kernel which provides more detail).

Now, what has Microsoft given back, monetarily? Who knows. Probably not much. As for code, around 1300 commits - although depending on how cynical you are, you can either view this as a good thing or Microsoft having customers on their Hyper-V cloud who want to use FreeBSD. To me, it can go either way, honestly - I'll take Microsoft commits just as readily as I'll take those of anyone else, but I don't think they're doing it purely out of the goodness of their hearts (they're a big corporation, and while they have individuals who're good, it's a difficult thing to mesh).
So what about Apple? Well, they haven't really sponsored much code (though that's perhaps because nobody's interested in picking through the huge changesets that result from them offering only date-based snapshots of Darwin with no VCS information). But they have given back, in the form of sponsoring a lot of conferences over the years (as in, sometimes being the primary sponsor), plus they've been instrumental in things like TrustedBSD (which has given us OpenBSM for auditing, MAC for RBAC et cetera, OpenPAM plus a bunch more things that I'm not sure Apple has been directly involved with). Oh, and they're also the primary reason why FreeBSD is almost GPL free because of LLVM - without that, FreeBSD and a lot of other software developers would still be dependent on the monoculture of GCC - and even if GCC has been enormously helpful, there's plenty of evidence to suggest that the best thing to happen to GCC was that it got some competition.
(I think over 8000 bytes is enough words)

BlankSystemDaemon
Mar 13, 2009



Did you read what I posted, or just picked out a few phrases here and there? Because you're the one who asked.

EDIT: Or are you just being intentionally obtuse and refusing to acknowledge that even if somehow, magically, there was nothing but copyleft software as an alternative to proprietary software, there's still legal compliance which can and does affect whether or not any copyleft software can be used?

BlankSystemDaemon fucked around with this message at 04:46 on Nov 26, 2021

BlankSystemDaemon
Mar 13, 2009



SYSV Fanfic posted:

I obviously read every word. Personally I think it's really lovely of apple they can't even toss $5 or $10k to the freebsd foundation annually. Microsoft still does.
Apple is listed under corporate donors and has been for many years.

BobHoward posted:

fyi the process model is mach tasks, not bsd

they hide a lot of it underneath more abstract libraries, including ones which provide traditional unix interfaces, but mach IPC primitives and the mach process (task) model are the base layer for a lot of stuff in macOS
Huh, interesting - but there is some of the BSD process model, because top(1) on macOS has a column for BSD syscalls (even if there are never any processes that use any; I've tested the entire base system).

Ocean of Milk posted:

What specififcally do you mean by that? Examples?
From trading war stories over drinks with people who've worked in banks, I know that copyleft software is ill thought of in the banking industry, and that at least one major bank has a mail sorting system where FreeBSD is used in some capacity.
Other industries with legal compliance are telecommunications, the medicinal industry and the industry around delivering software to the armed forces - all of which I don't know specifics, but I know that people who work on the BSDs have been hired by subcontractors of those, because they worked on the BSDs.
One thing they all have in common is that they're required by law or contract to keep things secret.

Needless to say, none of it is verifiable, but that's the nature of these things.

BlankSystemDaemon
Mar 13, 2009



DoomTrainPhD posted:

Real talk: In the several decades I have been using Linux on my desktop, sound has never been an issue. Especially after kernel 2.4 (when OSS was replaced with ALSA and you could play sound from multiple sources AT THE SAME TIME!!!!) :corsair:
Are you saying that OSS on Linux had ASIO-like characteristics?

BobHoward posted:

linux style package management is unironically the #1 reason why linux can't have good software
It's certainly the reason Adobe gave for never releasing any of their products on Linux.

infernal machines posted:

windows has the microsoft store.

no one uses it, but it's there.

a lot of those old paid mac utilities are holdovers from some earlier version of macos before that feature existed in the os itself. for whatever reason they keep puttering along long after apple has integrated it into the os.
For the longest time, getting onto the Micorsoft Store required you to provide UWP, which nobody wanted to for obvious reasons, partly because it also involved compatibility for Windows RT.

Microsoft have, for the longest time, had the technology ready to do proper third-party software repositories in the form of NuGet (which gets used by Chocolatey), but it's only until recently they've started doing sensible things with Microsoft Store.
So I'm sure it'll only be a matter of time before they stop doing that.

BlankSystemDaemon
Mar 13, 2009



mystes posted:

The microsoft store isn't even really a package manager anymore because now, in addition to packaged non-sandboxed win32 applications, it can install arbitrary programs that can't even be updated through the store, which I think is a dumb idea, but they decided that that was the trick to get more people to use it, I guess.
Welp, it didn't take them long to ruin it, did it.

BlankSystemDaemon
Mar 13, 2009



Mr. Crow posted:

What if I'm installing random crap from the Debian repo and "getting raw dogged"?

In short: computers are a land of contrasts
that's very near peak opensource.nzb

BlankSystemDaemon
Mar 13, 2009



Shaggar posted:

if your package system contains scripts its a piece of poo poo
isn't debians packaging system entirely based on a multitude of shell scripts, each with their own set of duplicate variables for everything under the sun?

BlankSystemDaemon
Mar 13, 2009



DoomTrainPhD posted:

cgroups v2 now!
still no isolation though

BlankSystemDaemon
Mar 13, 2009



Phobeste posted:

no, it's not. no matter what you're building for the love of god do not have a workflow where artifacts get built on a random loving developer laptop and deployed directly. come on
but op, that's how all firmware is built
if it's good enough for them, why is it not good enough for you?

BlankSystemDaemon
Mar 13, 2009



DoomTrainPhD posted:

What the gently caress are you talking about? Deployment happens in a pipeline, no random loving developer laptop get's to deploy anything
ah, you're a stranger to systemd development, i see

Poopernickel posted:

just use SSH and tunnel everything

boom, problem solved you're welcome
ssh -D 1080 owns, though

mawarannahr posted:

has anyone used mosh? is it worth it? would IT kill me?
the only advantage to mosh that i've been able to find is if you're on a link with a high latency and high delta between packets

so if you're on a train while using WWAN, basically

BlankSystemDaemon
Mar 13, 2009



nudgenudgetilt posted:

does that even matter any more? I was holding open an ssh tcp connection through the sf subway, transbay tube, and all that poo poo a decade ago with umts/hsdpa.

mosh was that thing everyone wanted to keep their eye on for a few years when it came out, because it solved a problem we had at the time, but we didn't consider mosh safe/stable enough to see widespread use. then connectivity got better and mosh was kinda pointless for the majority of the users who thought it was neat.
connectivity can still be spotty in areas; there's a place near where i live where if you walk down a bike path between some houses, reception just drops for no adequately explainable reason
but connectivity has gotten a lot better for the most part

the only other use-case for mosh i know of is the way i heard about it originally, with someone being trapped in an elevator

BlankSystemDaemon
Mar 13, 2009



imagine that i posted a flowchart with the question "are you a hyperscaler" leading to a boolean where yes is "use kubernetes" and no is "don't use kubernetes".

BlankSystemDaemon
Mar 13, 2009



Poopernickel posted:

The pendulum will swing away from containers within the next 4 years, guaranteed.
some of us have had containers since 1999

BlankSystemDaemon
Mar 13, 2009



oldie but goodie:
https://www.youtube.com/watch?v=CPRvc2UMeMI

BlankSystemDaemon
Mar 13, 2009



SYSV Fanfic posted:

And the cloud was just called "time sharing".
a time shared butt

BlankSystemDaemon
Mar 13, 2009



NihilCredo posted:

tfw you have a list of codes in a text file and you want to find all files that contain any of them, so you do:

Bash code:
cat barcodes.txt | tr '\n' '|' | head -c -1 | xargs -i grep -Elr '{}' .
and you feel all greybeardy and wizardy then realized you could have just done

Bash code:
grep -lr -f barcodes.txt .
it's kind of a GNUism because GNU grep evolved a bunch of options that replicate other functionality, but it doesn't mean they have to be used - i still would've done the first way, simply because it's easier to think in terms of pipelines and small utilities for me
same thing happened to GNU ls - and both of them ended up being ported to BSD grep and BSD ls because they were so commonly used that people kept asking about them

later on, these options also started being added to speed things up - for example `find <expr> -exec grep <expr2> {} +` being used as a weird incantation of magic characters, instead of piping to xargs because it used to be faster than using piping, instead of you know making pipes faster
same weird optimization philosophy also got applied to GNU yes, so that it's incredibly fast, for no easily ascertainable reason

long options are also a GNUism and i seem to recall that they were originally supposed to be used only for scripting - the idea being that it's easier to read the script if you know what's being done, but then what's the point of manual pagesGNU info?

BlankSystemDaemon fucked around with this message at 19:23 on Dec 2, 2021

BlankSystemDaemon
Mar 13, 2009



mawarannahr posted:

-f is posix, op
small utilities and pipelines aren't a posix thing, it's a unix principle
-r doesn't exist and i'm pretty sure the behaviour of -l is subtly different

BlankSystemDaemon
Mar 13, 2009



Rufus Ping posted:

Wow. Hostile gatekeeping and toxic reaction much. This does NOT reflect well on the Linux community
We're gonna have this copypasta for as long as we've had the BSD is dead copypasta. :allears:

BlankSystemDaemon
Mar 13, 2009



Kamrat posted:

BSD is just for people that thinks Linux is too mainstream
what does that make Illumos?

BlankSystemDaemon
Mar 13, 2009



freeasinbeer posted:

if you really hate K8s that much; the real hipster alternative to K8s is Nomad, way simpler and solves a bunch of the stuff K8s did with better tooling like consul and vault.
A neat thing is that nomad is also fairly platform agnostic

Poopernickel posted:

it's weird how "unstable" has such different meanings for Linux and Windows

for Windows, unstable means "might crash randomly for no reason"

for Linux, unstable means "updates will break your poo poo constantly"
So both mean that you end up losing data? That sounds fairly consistent.

BlankSystemDaemon
Mar 13, 2009



mycophobia posted:

macos is cool. or at least was, i havent used it since like 10.5
that was the last good version, op

BlankSystemDaemon
Mar 13, 2009



carry on then posted:

this is how i know you've never used it

snow leopard motherfucker
10.5 is the last good version because it was the final one to support the powerbook g4, where the brushed aluminium look inspired the look of cocoa

had 10.6 supported the powerbook g4, it would've been the last good version (although if you went to the trouble of acquiring a development version, it could work but it meant no updates)

SYSV Fanfic posted:

The main thing to know is that if you are most concerned about browser security, snaps/flatpaks add extra security over unzipping to opt. The browser will run in a sandbox, you'll get the latest updates, and there's a better chance dependencies won't have vulnerabilities. IDK much about appimage these days, but it's the same idea as old appimage except it uses kernel features to better protect your system from vulnerable or poorly packaged software.

I only compile software I want or need to have full access to my system anymore. If I need the latest version of something I use snap.
application-enforced sandboxes are only as good as the application itself - proper sandboxing is enforced from an elevated privilege

BlankSystemDaemon
Mar 13, 2009



carry on then posted:

the first gen macbook pro was identical to the powerbook while being 5x faster. leopard is obsolete.
it wasn't a powerbook though :mad:

BlankSystemDaemon
Mar 13, 2009



how about, i dunno, fixing the OOM killer? oh wait, people have been trying for over a decade

what it invariably comes down to is either people talking about performance as if that's the most important thing over system stability when the system is in production use, or it's some vague "you don't need that" post, as in the first response to the lwn.net article above

BlankSystemDaemon fucked around with this message at 06:30 on Dec 11, 2021

BlankSystemDaemon
Mar 13, 2009



SYSV Fanfic posted:

There's nothing to fix. Exhausting all available physical and virtual memory is an error condition caused by bugs or bad design on the part of the user. "System stability" is kind of a moot point when the other option is to automatically flush all dirty buffers/cache to disk and halt the system or wait for a human oracle rather than a heuristic.

Giving people (more) control over the heuristic just makes it more likely the OOM killer won't be able to bring the system into a safe state to even respond to sysrq keys. Truly important work gets check pointed/has a commit log. Those mechanisms will fail if the kernel can't recover to the point that pending file writes can complete.

SYSV Fanfic posted:

Any time my computer starts swapping I get paranoid that it's dying/freezing. Can't be the only one.
There's something deeply wrong with the entire loving virtual memory management when the response to fixing a tiny piece of software is to completely abandon (demand) paging and insist on going back to physical memorymapping, because there's apparently a guarantee that if any swapping happens, the system will crash.

BlankSystemDaemon
Mar 13, 2009



SYSV Fanfic posted:

Talking about being anxious my hardware is dying when swapping happens. Like... did this thing just freeze for five seconds ten times in a row b/c I need to repaste something, am I going to have to do an RMA/buy a new board? Thermal throttling? Oh no, just firefox using 10,000% CPU and 150% available ram. Nope just working as intended, swapping out thousands of 4k chunks of memory like it's 1994 and I'm running emacs or some poo poo.

On a desktop, OOM killer always gets its man b/c it's always the browser.
Yeah, something is deeply loving broken with the VM subsystem if that's what you're experiencing.

BlankSystemDaemon
Mar 13, 2009



hifi posted:

anyways bsd probably sucks just as much poo poo if not more when some docker crapware is spawning itself 300 times a second
I've seen people fail to forkbomb a FreeBSD based honeypot that I used to manage, without success.

BlankSystemDaemon fucked around with this message at 19:57 on Dec 11, 2021

BlankSystemDaemon
Mar 13, 2009



Tankakern posted:

so they succeded?
Ha, holy poo poo I did a double negative unintentionally. :v:

BlankSystemDaemon
Mar 13, 2009



The_Franz posted:

freebsd oom killer uses the "new guy in the prison yard" approach: find the biggest process and take it out. linux attempts to use some heuristics when trying to figure out which process to whack
it might also be worth mentioning that protect(1) is meant to be used to protect processes from it, and that the _oomprotect suffix for rc.conf can set it for services

BlankSystemDaemon
Mar 13, 2009



Nomnom Cookie posted:

my earlier point, which you seem to have either missed or ignored, is that in the scenarios kubernetes is designed for and deployed into, crashing the system is better than slowing down. a crash is a fail-stop, fail-stops are well understood and distributed systems are usually pretty good at dealing with them. fail-slows are a lot less well understood, are more complicated to handle, and some of the mechanisms used to detect fail-stops will aggravate the effects of a fail-slow. that is why kubernetes was built to operate without swap and why people who know what they’re doing won’t use kube with swap unless they really know what they’re doing and really need it
what definition of distributed system are you using?
kubernetes is scale-out fail-over with load-balancing, not a ssi with single root/ipc/IO/process space, process migration and all the other fun stuff

Lysidas posted:

there is a similar mechanism to give weights to processes for the linux oom killer, i have never looked in to details of how these are set but ive seen a bunch of oom killer ouptut in dmesg, and most systems are configured out of the box so e.g. sshd will never be oom-killed
protect(1) and _oomprotect isn't really about weighing anything for how important it is; it's just a way to ensure that certain things don't get killed - i'm not sure how much value there is in having a weighted process if the heuristic for killing processes doesn't kill the right process to begin with

BlankSystemDaemon
Mar 13, 2009



SYSV Fanfic posted:

You don't have to wait, here's an blanksystemdaemon flowchart you can just respond to.


:v:

BlankSystemDaemon
Mar 13, 2009



Antigravitas posted:

They can't host a central repo. They would have to secure redistribution rights with every vendor, and good luck with that. It's still just a bunch of wrapper scripts around third-party garbage binary installers. The only party that could conceivably pull that off is Microsoft, and they only built a dpkg-equivalent in MSI, then lost interest when it came to writing the part that actually manages packages…

At least you are strongly encouraged to put a hash of the downloaded executable in the chocolatey install script…
Microsoft has had the perfect solution for years in the form of NuGet, which ironically is used by Chocolatey - and it's what underpins the Install-Package cmdlet in PowerShell.

BlankSystemDaemon
Mar 13, 2009



Basically, everything is terrible.

BlankSystemDaemon
Mar 13, 2009



what is linux desktop, i post as i browse sa on a used-but-new-to-me t480s with freebsd, wayland and sway

Adbot
ADBOT LOVES YOU

BlankSystemDaemon
Mar 13, 2009



Kazinsal posted:

is sndio standard on freebsd these days or is it still a port? apparently some folks ported it to linux because porting openbsd's sound layer to linux took less time than getting sound to universally work on alsa and friends
it's still a port, as freebsd has its own pcm(4) stack that does a bunch of things which sndio doesn't do like OSSv4 compatibility, bitperfect audio, low latency mode, high-quality re-sampling (known as project z when it was being developed), virtual channels (so each app can get its own set of channels, but it still gets mixed together in the kernel when playing at the same time; this is not necessary for things that support OSS multiplexing though - only for things that insist on having control like pulseaudio and alsa if you need either or both)

sndio is great for getting the same interface across multiple OS' though and for a while it was also the only way to do webrtc in firefox on freebsd (although this has since been fixed, and oss in libcubeb now works again after the bitrot got fixed by one of the freebsd developers)

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply