Systemd 244 Released With New Init System Features For Black Friday

[starshipeleven]

I won't bother wasting more time on you. You already proved yourself troll in this with your "all GNU software is xxxx except GCC". Better things to do.

copying here the relevant part about it from the 3 posts above

More ignorant bullshit on your part.

One thing is what third party project they accepted and host on their infrastructure, one thing is what they actually make themselves.

And yes one of my main gripes is with coreutils, which IS a GNU project. I've yet to find something I can't do with busybox, that is like 1/10th of the size of their crap. Seriously, just look at the "true" tool. That's 35 kbytes to just return "true" when called.

Also their build system (Autotools) is ass.

Make is OK I guess.

[aht0]

I'd like to see some numbers on that. Not that I don't believe you but I think we are both right. BSD is gaining users, but it's not really anywhere near enough to matter.

Kinda hard to provide numbers, since there's no user tracking whatsoever. I can just say I keep seeing new users who term themselves "systemd refugee" in various forums and BSD-groups in Facebook.

You troll about Linux and I troll you about Slowlaris. That's poetic justice.

I troll? Dude, whatever I write has generally some reasoning behind it, which causes me to write whatever I write. I don't go just writing "xxx is shit, you suck, you moron". In fact I try to avoid going personal..

You have your own smaller heavy weights like iXsystems (FreeNAS and its proprietary counterpart) and Netgate (Pfsense), and Netflix.

Not one of them has effectively imprisoned BSD family of OS'es and dictate direction like RH/IBM does with Linux distros through systemd

You talked about how Linux is raped by megacorps and shoves stuff down your throat.

So, lets look at it.
Check the commits to Linux kernel. What is the percentage authored by developers hired by some or other enterprise/company/conglomerate? 50+% last time I checked. Which mathemetically means "MAJORITY". So, in effect, enterprises largely have control over what gets added to the kernel, how it's implemented etc. Because if they would not do it, most likely nobody would. Torvalds & his minions only stop commits which are outrageously stupid, disruptive or bad quality - not even then always (code churn could exceed their ability and we do see often enough how big chunks of code are pulled back out). What concludes: it's mostly enterprises that steer the direction of Linux -> going vulgar: Linux is their butt-boy like I said. That's from the kernel side of things.

Now, systemd.. Big piece of software, bent on taking over the Linux ecosystem. Designed in a way to achieve maximum usage among Linux distros - which also coincidentally translates to maximum control by Red Hat.
Control [yes, control : It's devs are hired by RH/IBM, unpaid freelance devs don't have any say in this matter] by Red Hat, multi-billion profit subsidiary of IBM megacorporation. NOT by non-profit steering commitee or any single individual.
It's development is controlled by profit-seeking business entity. You may indulge in wishful thinking about it being benevolent or whatever: in the end, all it cares is using and shaping Linux so it makes it more money. It (in the form of it's shareholders/owners) does not give flying fuck about Linux's itself outside using it for making more.. money.
Are you still arguing my saying "Linux is buttboy for corporations" or you actually comprehend my reasoning? I wasn't just trolling, there was solid reasoning behind. You can agree with the reasoning or not, that's your choice.

You talked about how Google's own puppy OS is going to be so great and have a permissive license.
I pointed out that Google's own puppy OS is going to be orders of magnitude more "raped by megacorps" and "shoves stuff down your throat" than Linux will ever be as while in Linux there isn't a single master and they do develop competing features and frameworks (i.e. AppArmor vs Selinux for example), on Fuchsia you will have only one true master, Google.

Actually I didn't say anything you claim I said. I literally only mentioned it's in the works and it could cause havoc to Linux. What you assume I said through your haze of angry emotions is frankly your problem. Read it over after cooling down.
Maybe you cannot cool down tho, so I'll explain it for you: Google has been developing Fuchsia some years now, maybe in order to ditch Android in it's embedded products, maybe it want's to use it for something like laptops etc as well. Or servers, hell knows - WHY does this thing have x86_64 arch support?
Since Google has global influence, more money than most nation states and technological infrastructure required - anything they do start to push seriously is also pretty fucking likely to gain traction too. How much, how widely, depends on actual implementations, so there's no point in further speculation. But it's probability.

[aht0]

copying here the relevant part about it from the 3 posts above

More ignorant bullshit on your part.

One thing is what third party project they accepted and host on their infrastructure, one thing is what they actually make themselves.

And yes one of my main gripes is with coreutils, which IS a GNU project. I've yet to find something I can't do with busybox, that is like 1/10th of the size of their crap. Seriously, just look at the "true" tool. That's 35 kbytes to just return "true" when called.

Also their build system (Autotools) is ass.

Make is OK I guess.

I find it HILARIOUS that you whine over GNU Core Utils being "too big" but behave like rabid attack dog if anyone dares even subtly criticize systemd.

IMHO, GNU has lots of useful software. Contrary to your beliefs I use GNU stuff quite a lot, some of what is always mandatory in my installs, even on BSD.

[aht0]

There is a limit. Once you get past it, having too many options is just as bad as too little.

That'd be fine theory if systemd domination did not mean elimination of all other options. limit=0 can not be defined as 'too many options' but as 'zero options'

For core components, you don't want too many options you can choose as this means a lot of duplication in all the distro.

That used to be the strength of Linux. Options and choices. Monoculture is never good.

This is not true and you need to fuck off with this bs.
Before systemd everyone was using more or less the same script-based init and more or less the same userspace core applications anyway too.

Actually, if you actually bothered really checking facts, you'd emerging picture contrary your claim. runit for example was first released ~15 years a go, upstart ~13 years a go, OpenRC ~12 years a go. They ALL predate systemd (it's been around slightly less than a decade).

Daemontools and it's various iterations for service management have existed nearly 20 years (since 2001).

I don't see why this is a problem. Doesn't BSD license (used in such applications) encourage you to take it and never contribute back?

Functionally, no difference and you know it, despite parroting old idiocy.

If someone does not want to contribute back, he is not going to, despite the license. If they do want to contribute back - they are going to. Desire to reduce your future maintenance load could be equally potent motivator - it's easier to submit your patches and let OS developers fix the problems pre-release than work around problems cropping up and try to patch these in-house "post-release".

You want to lock your device in order to prevent modification but it has GPL software in it? Just lock the boot loader and it works same. Or dump the sources in unusable form (follow the letter of the law but violate the spirit). So, in effect if some user does not want to contribute back then license makes literally zero difference.

There isn't much more to take, and new stuff coming to BSD is not really a thing.

Funny claim, since you aren't even using any of the BSD's and just parrot your belief based on prejudice.

Can I remind you that BSD does indeed ignore POSIX when they see fit too? Neither Linux nor BSD are fully POSIX-compliant.

There is no willing, willing ignorance backed by contempt against POSIX.
https://www.freebsd.org/doc/en_US.IS...standards.html
https://wiki.freebsd.org/AvoidingLinuxisms

Where FreeBSD breaks it's POSIX-compliance in base OS, is it's signal.h where it does not define SIGRTMIN and SIGRTMAX.

For comparison, let me remind you general attitude from Linux. "We know better, we are better and we are not going to follow it, we go beyond POSIX, it's outdated" etc. Very similar attitude to MS Windows.

But the problem isn't POSIX, the problem is using kernel features. Linux kernel features are not the same as BSD kernel features (actually these features are different even from a BSD to another)

Problem is by now two-fold.

• Emulation of non-standard (read:non-POSIX) system calls found in newer Linux. It's the reason why Linux ABI in FreeBSD is stuck on Linux 2.6.3x (it's emulating either CentOS6 or optionally CentOS7, 32/64-bit variants both). It's better in NetBSD (somewhere around Linux 3.2x), but it may have had negative consequences security-wise for NetBSD (happened to read some audit once) because binary emulation ain't easy to implement, complicated code is far more prone to errors and OpenBSD gave binary emulation of Linux up completely because they did not like resulting security implications and code complexity. At all.
• systemd. You want to emulate it properly, depending on particular component you may have to emulate most of the suite. Interlocked, interdependent tightly coupled interfaces do that. Plus, it's changing. Yeah, OpenBSD has some stubs. elogind exist, it's not enough.

There is also WolfSSL and MbedTLS that are both real and serious projects (can be found and used in OpenWrt among other things)

That's actually good. Crypto is one area tho where I'd first check the background of developers. Do they have background in math, cryptograhy AND security. If yes, then even better.

and dnsmasq and something from systemd project and a whole lot of others https://en.wikipedia.org/wiki/Compar...erver_software

Well, dnsmasq is not per se DNS server, it is "DNS forwarder" which is not nearly the same animal. What it CAN serve, is DHCP. Same thing with systemd-resolved. It's not DNS server, it's at best haphazardly written placeholder stub and at least year a go it still exposed rather weird resolving issues and behaviour (I don't sit on BSD all the time you know).

It's not like outside of it there are no other applications.

After we exclude dnsmasq and systemd-resolved due the fact they are NOT in fact DNS servers but at best forwarders, then what would be your reasonably good-featured alternatives to Bind and Unbound?
DNS servers were also just a random example I came up with. Life has shown that unexpected critical flaws may hit from literally any direction. HeartBleed, CRACK and other such security flaws illustrate my case. Not to mention whole debacle around Spectre/Meltdown. You just never know.

Y didn't break anything in any way, shape or form.
Y offers a better way to do something and someone else contributed a way using Y to the upstream software that was accepted as it was indeed considered useful and merged.
If X can't be arsed to contribute or maintain a way to run the same software on X systems, then it's their own problem.
If the upstream software does not want to merge a way to run on X systems, then it is their own problem, not Y problem.

[/quote]
Y offers a better way to do something and someone else contributed a way using Y to the upstream software that was accepted as it was indeed considered useful and merged while Y was fully aware his creation was in the future going to hurt alternative components and due the "vendor-lock/viral" nature of Y's offering - other OS'es who also made use of said alternative components. Because Y has strong personal opinion that non-Linux FOSS OS' should just die off, he don't give flying fuck about the consequences his creation has.

[arokh]

You realized that you don't actually have any arguments other than your BSD friends not liking systemd for reasons you don't even understand. I'm sure you have better things to do than complaining about an awesome piece of software like systemd indeed. It's loved by admins and you managed to claim it takes control away from them trollol, you obviously never worked professionaly maintaining *nix systems.

[aht0]

You realized that (1)you don't actually have any arguments (2)other than your BSD friends not liking systemd (3)for reasons you don't even understand. I'm sure you have better things to do than complaining about an (4)awesome piece of software like systemd indeed. (5)It's loved by adminsand (6)you managed to claim it takes control away from them trollol, (7)you obviously never worked professionaly maintaining *nix systems.

(1) I've posted a ton. If you cannot read, it's your disability.
(2) Bs, when I switched from Linux to BSD, I did it especially because of systemd. As typical lennartware-product (like PulseAudio) initial versions were too buggy and made simply using Linux too frustrating.
(3) No, It appears to me that you do not have basic ability to read and understand written text. Or refuse to understnad
Otherwise that sentence in itself would not have been so empty of any meaning.

(4) It's "awesome" for enterprise users. For anyone else it's from "meh" to that "fucking piece of buggy lennartware-shit".
(5) Admins are not the entirety of Linux user base.
(6) I never wrote that. Somebody else might have tho. Again exemplifies your basic inability to read and understand what's written.
(7) Linux ain't Unix, so no. But my first jobs included maintaining Linux systems. In fact I was first hired to be sysadmin by the very school I learned IT in. After compulsory draft into military I chose that occupation because I liked that life much more.

[arokh]

You've posted a lot of complete non-sense, that's all. As far as I can tell, it's the usual conspiracy theory fear of "Redhat is taking over the world and it's a huuuuge problem". This whining about "imprisoning" BSD is the stupidest thing I ever heard, you don't seem to comprehend that it's a different OS with absolutely no connection to Linux. I really wonder why a complete noob like you is so opinionated on the Linux init system, you obviously have never contributed a line of code yourself. BSD devs are welcome to contribute the code they need.

You don't seem to understand the word "parroting". It's what you are doing when you are expressing the opinions of "systemd refugees" on facebook or some BSD forum users. You've completely failed to provide a relevant reason as to why systemd is bad. Get off the internet and stop complaining about free software OK, like you said you have better things to do. Like start coding shims for BSD for example.

[aht0]

You've posted a lot of complete non-sense, that's all.(1) As far as I can tell, it's the usual conspiracy theory fear of "Redhat is taking over the world and it's a huuuuge problem". This whining about "imprisoning" BSD is the stupidest thing I ever heard, you don't seem to comprehend that it's a different OS with absolutely no connection to Linux. (2)I really wonder why a complete noob like you is so opinionated on the Linux init system, (3)you obviously have never contributed a line of code yourself. (4)BSD devs are welcome to contribute the code they need.

You don't seem to understand the word "parroting". It's what you are doing when you are expressing the opinions of "systemd refugees" on facebook or some BSD forum users. You've completely failed to provide a relevant reason as to why systemd is bad(5). Get off the internet and stop complaining about free software OK, like you said you have better things to do. Like start coding shims for BSD for example.

(1) Let's see about conspiracy theories or associations.

• U.S Army Major General Nicholas Justice has publicly declared that U.S military-industrial complex is 'the largest client of RHEL'. (Fact One).
• RHEL-hired/associated devs (exclusively!) are developing systemd. (Fact Two).
• Most Linux distros, at this point, are using systemd. (Fact Three).
• systemd devs are a group of privileged people who have direct control/say-so over the direction of developing systemd. When they don't want code contribution from "outside", they will reject and there's not a thing anyone else really can do about it. Forking ain't really a practical option because no single developer can compete with group of hired devs. RHEL thus (because RHEL is employer of said devs), has total control over development of systemd. (Fact Four).
• We have by now determined that RHEL has total control over systemd (see Fact Four), thus we can conclude that RHEL has subtle control over majority (see Fact Three) of Linux distros - subtle, in the form of "by shaping systemd development as it wishes, RHEL can steer the direction Linux distros using their systemd, develop". In fact, the scope of it's subtle control is increasing because systemd keeps replacing stand-alone components in GNU/Linux distros (Fact Five).
• RHEL is a business entity, not some non-profit software steering commitee, so it's primary goal is to make money. Thus RHEL can be greatly influenced by what it's larger customers want to see in it's products. RHEL can only comply if it does not want to lose it's customers to competitors (like Oracle Corporation, which has very similar distro among it's offerings AND capability for additional software engineering if need should arise). Conclusion: RHEL can be influenced/dictated by governmental organizations (Fact Six). Precise degree of incluence: unknown.

It absolutely looks like RHEL has no control over anything and all "conspiracy theories" are groundless, not.

(2) 'Noob' is a derogative form of 'newbie'. Most often used in gaming, in insulting manner. Considering the fact my first Unix/Linux association happened on Red Hat Linux 5.2 nearly two decades a go (back in the beginning of a 1999) your assumption is void.

(3) Don't worry, I have. Not for Linux, admittedly. But you were coming out with a general assumption I can negate.
(4) Are they indeed? Your claim would seem to contradict with L.Poettering's publicly declared opinion that BSD's should simply die. In that light I find your claim rather arguable.
(5) If I repeated all the "why's" again, you'd accuse me of "parroting", completely ignoring the fact that you seem to simply choose to be in flat denial about ANY argument I bring forth.

[AsuMagic]

Pretty amusing how essentially most arguments there can be used against linux as well.

[arokh]

Of course Redhat has control, what exactly is the problem with that? It's only natural. Other distributions are free to use whatever they want, of course they use systemd as it's the best solution. Do you have any technical issues with Pulseaudio? Which of your PR's have been rejected?

Obviously your first encounter with Linux does not decide whether you're a complete noob or not, the fact that you are on here complaining about nothing does. BSD can adapt or die, it's all up to their devs and really has nothing to do with systemd at all. In case you haven't noticed, Unix has been slowly dying since Linux showed up. Too bad nobody cares to provide code for your hobby OS anymore, go cry in the BSD section?