Will be ever be free from pic related?

It's still Bell Labs garbage anyway.
You aren't fooling anyone, Rob.

Better OSes already existed before Unix was even created in the first place.

More like pozix

>Better OSes
In what way, exactly?

Redox OS is another example of a POSIX-compliant microkernel.

POPULAR == BAD
SMALL == GOOD

Attached: 015.png (434x327, 55K)

Whats wrong with posix?
It makes writing cross-platform code super easy

Attached: 1562540656873.jpg (512x422, 29K)

Nothing wrong with it. A bunch of NEETs can't handle not feeling special by running some obscure kernel design.

seriously is POSIX actually used beside mutexe objects? it was a dream that never came true

>Never came true
The fact that you can compile something on Linux, then simply recompile it to work on FreeBSD is solid proof that it did its job well.
(Of course some software might require minor changes)

my systemd install isn't posix compliant so I'm already free

How so? systemd doesn't have anything to do with POSIX compliance.
Sure, it doesn't follow the UNIX philosophy of "do one thing and do it well", but nothing more beyond that.

no I mean that AFAIK POSIX was meant to provide a unified OS API. So you could use the same call for linux, windows, mac or whatever unix or anything else. Wasn't it? This was a perfect vision but of course this is strictly against the interests of big companies like Microsoft and Apple. Since this would make any application just works on any OS.

So? Syscalls still work the same on POSIX compliant operating systems.
(Unless you're using platform-specific syscalls, like seccomp(2) on Linux or pledge(2) on OpenBSD, etc)

But OS X is certified POSIX compliant and Windows has had POSIX compatibility from NT's initial release through 7, and it's back now that WSL is a thing

>Plan 9 has POSIX compatibility layer
ftfy

>posix refers to kernel design

fork() needs to fucking die already, it's such a shit design

Fuchsia is getting rid of that apparently. I remember seeing based Lunduke mention that about the OS in a video he seems to have deleted.

Use TempleOS

Please provide a detailed, well thought out explanation of why you are posting this and not jsut because you are a contrarian shitposter.

The burden is on people justifying and imposing the omnipresence of POSIX, not those refusing it.

POSIX is a well-designed standard that has proven to be effective and it has been in use for decades, so your arguments are completely bullshit unless you (or any other contrarian shitposter) can explain exactly why the standard is bad and what is a better alternative to it.

>POSIX
>well-designed standard

Being able to compile most software on this planet. Even win32 supports a castrated posix subset, like read write calls and sockets.

You people really are terrible at arguing, eh?

Not an argument. Go have sex

This is why no one takes you seriously.

>Not an argument
>Go have sex

Beats win32.

t. poettering

>How so? systemd doesn't have anything to do with POSIX compliance.
exactly

I think you have any idea what POSIX is

Serious question: do you know what POSIX is? Or are you simply making assumptions that POSIX means a minimal system?

POSIX literally just defines some C standard library functions, some shell utilities, and some other pretty minor things.
It's not some massively monolithic thing that specifies how the kernel is designed or how your fucking init system works.

I'm not using a POSIX compliant system so I'm free from it, read OP's question.

Oh so you're trolling then, nice.
Seriously try harder, this is an obvious one.

Use Haiku, unironically. Haiku is beautiful and you can do most of your daily stuff in it. And it isn't posix

Emacs isn't POSIX. Use Emacs.

Sorry to break your bubble, but Haiku is POSIX compliant and it has bash and shell utilities as well:
en.wikipedia.org/wiki/POSIX

>fucktard still doesn't understand what POSIX is

Ah sorry, I must have been misinformed/I'm not the brightest.

The fact that this thread exists and that there are people who don't even know what POSIX is yet argue about how "bad" it is clearly reflects on the current state of this board and how bad it has gotten of the years.
I blame the moderators for keeping low quality shit like this on this site.
This doesn't even mention the current stupid shitstorm of AMD/Intel that has completely shat over this board for the past few months, yet the moderators are ***still*** not doing anything about it.
Fuck this board and fuck every single one of its moderators.

It is only partially posix compliant actually

Even FreeBSD (a direct UNIX descendent) is considered partially compliant.
This is not because it's not POSIX compliant, but because POSIX certification is now handled by a scam corporation called "The Open Group" that will sue anyone (including Haiku) who claims to be POSIX compliant without paying them money.

>AMD/Intel that has completely shat over this board for the past few months
Months? That shit's been going for for 15 years.

It was mild back then, these months it has completely taken over the board.
The *only* way to solve this is to ban consumer technology (and maybe create a separate consumer technology board).

Fuck off Rob, it's you Bell Labs worshippers who keep screaming about how "well designed" it is, ignoring how even 70s shit such as Multics was better, which is ironic since Unix was supposedly based off it.
And the usual argument, apart from plain ignoring it, is bringing up stuff Windows does worse all the time and shout "see? Windows is shit, therefore POSIX good!".
You want these threads off the board because they scare you, they shatter your dream world.

Thankfully no.

Show a concrete example of a better standard that could be used instead of POSIX based for this task: create a file, write 10 bytes to it, close the file. You can just state the name of the standard, and the API functions involved.

An answer for POSIX would look like: POSIX, open, write, close.

I'll be waiting.

He won't show you anything, he's just a moron troll who couldn't make a proper bait thread.

Genera, years before the invention of POSIX, directly provided a with-open-file construct, which doesn't even require you to close the file yourself, avoiding potentially flooding the system with open files.
>inb4 b-but it's not an api!!!1
It's directly provided by the OS and directly usable by an application programmer, I'd say it qualifies.

So write a few line wrapper function? How is this an argument against the posix standard?

The POSIX API isn't perfect. Most APIs that exist from standardising common practice usually aren't.
But the fundamentals are solid, and is still the best API we have for this kind of shit.

Attached: 1375773299400.jpg (425x419, 19K)

that's what APE is for you dipshit

>inb4 b-but it's not an api!!!1
It's not one that can exist on the lowest layers like POSIX. What did Genera do if you needed to keep the file open for the duration of the program?

Microsoft/NT

Much worse than POSIX.
>sockets

It is an argument against POSIX because it challenges the usual fundamental notion of how you approach the problem you yourself presented: writing to a file.
The way POSIX is designed, opening, writing and closing are separate concepts, and so they are to the application programmer, unless some higher-level implementation, separate from the OS, provides wrappers on top of it.
Now, considering how file writing is used 99% of the time, it makes sense to have a primitive concept of "use this file in this portion of code" and let the OS, natively, take care of opening and closing before and after said portion.
Encumbering the *application* programmer with such details is unnecessary and counterproductive, which is the opposite of what an *application* programming interface should do.

You could also just open it of course, but it's much less frequent usage. See

your points just make me laugh at Jow Forums

See The API being low-level encumbers nobody since programmers can just use fitting libraries built on top of it if they need a simplified (but less capable) API.

If this is the level of arguments against posix you have, you aren't convincing anybody.

How would you implement databases, dipshit?
POSIX isn't a high level framework. It's an OS interface. You can implement high level frameworks on top of it. Many programming languages implement the abstractions you're thinking of on top of POSIX.

Fucking retard. Did you make the thread? Jesus Christ.

>The API being low-level encumbers nobody since programmers can just use fitting libraries built on top of it if they need a simplified (but less capable) API.
It encumbers low-level programmers.
The fact that you are pretty much admitting "lmao just build stuff on top of it" is itself an argument against POSIX, and ironically encourages the so-called "bloat" many people hate.
>If this is the level of arguments against posix you have, you aren't convincing anybody.
Sorry, next time I'll avoid merely answering questions you yourself presented.

Not only you could implement databases in the same way, but it would arguably be much simpler with a standardized structured reader and writer, instead of, ince again, bothering the application programmer with making their own parsers.
The fact that "many programming languages implement such abstractions" supposedly on top of POSIX is again, not an argument in favor of POSIX: in fact, such an inadequate base forces wheel reinvention every time, which I, quite frankly, find absurd.
>Fucking retard. Did you make the thread? Jesus Christ.
No, but I welcome and encourage everyone who challenges groupthink and questions how things are now, so that we can hope for changes for the better.
Many people here instead just accept the way things are and refuse to accept that we can, and must, do better.

>t's not one that can exist on the lowest layers like POSIX.
It literally can, though, and it did. Why couldn't it?

>The fact that you are pretty much admitting "lmao just build stuff on top of it" is itself an argument against POSIX
This is called abstraction. It's what enables us to write highly abstract programs on machines basically built of nand gates. And here you're calling this use of layered abstraction "bloat".

Repeatedly opening and closing files is inefficient, and also leads to a bunch of annoyances and even race conditions.
Trying to cram several operations into the same instruction severely reduces flexibility for what you can do. You'd only do that if it provides some kind of atomicity advantage, which open-then-write doesn't.

Attached: 1541532591349.png (821x3109, 559K)

Posix is based as fuck.
Just like XDG baseD* directory specification.

Most of those microkernels are still POSIX compliant, i dont think you even know what you're talking about.

You have no idea what you're even talking about