CVE-2017-9445 Systemd DNS

Reminder that it's Lennart that we are talking about.

are you implying that only lennart writes buggy c code? or that it only matters when lennart does it?
please clarify what lennart has to do with it.

Oh look, another rust shill. Where is your init system? Where is your kernel?
Memory issues alone are not enough to convince developers to switch from C. There are other factors to consider.

0/10 damage control cfag.
what the fuck?
yeah that is the sad reality.

Given Lennart's history and the quality of code he puts out, he should be working with pajeets in Java where he doesn't have to worry about "difficult" things like memory management.

...

...

actually i agree that poetteringware is trash

HAHAHAHAHAHAHAHAHAHA
Fuck you pottering and redhat for sucking his dick.
As a linux admin who has to work with systemd daily FUCK THIS SHIT. Just because it works (mostly) and has a few neat features DOESN'T MAKE IT GOOD, see windows 10.

Not an issue.
Every fucking server admin use IP addresses instead of DNS on critical servers.

Every fucking server admin use Tor and onion addresses instead of DNS on critical servers.

So does this mean that Lennart Poittering is officially /ourguy/

lol do you enjoy getting v&? use i2p

bump

That's the real question. C was a fringe language until the late 1980s, when people started rewriting programs in C. The meme about C being important and that it was always popular is just a myth spread by C programmers to make it seem like you have no hope of replacing C, but that's not true at all.

systemd tantrums is why I switched to BSD
I might try plan 9 next

tantrums are why
sorry for spell mistake i live in delphi please help

poo in the loo pajeet

how would he be (((/ourguy/)))?

github.com/systemd/systemd/issues/6237

Is there a lot cases like that?

without-systemd.org/wiki/index.php/Arguments_against_systemd

Greed, aka redhat kikes.

and install java virtual-loo
nope

ALWAYS

systemd was shilled very heavily. There would be whole threads here and on halfchan dedicated to how fast systemd would boot your computer.

never happened. it was always shit on.

Well, Pottering is being a faggot as usual, and the bug might be a pretty big deal with templated unit files
sudo systemctl enable [email protected]/* */sudo systemctl start [email protected]/* */
I can't get any templated unit files to work like this, though. So, it's not that big of a deal. Services run as root by default when it's a system level unit, if you don't specify the user. The 'bug' here is in validation and not failing loudly, if anything. But falling back to a root user when it's unit files created and enabled by the root user is not a catastrophic thing. And if I can't get templated unit files to work like this, then it's not exactly a disaster, just lack of validation.

And retards like yourself would always be put in your place.

No, it wasn't, you're just mentally ill.

...

Modern unices don't follow the UNIX philosophy, UNIX died decades ago.
Daemontools can't manage processes that daemonize themselves (especially for proprietary shit that I do not have the option to disable this behavior on), systemd does it well, and gives you free resource limits using the same interface, not to mention the ability to express multiple failure modes in different situations for units whereas daemontools will loop forever.
Poettering is a faggot but that doesn't mean the software is bad, or whatever conspiracy theories autists want to vomit out.

These are bugs. They happen. I don't agree with Poettering being a faggot and saying that the username is invalid, thus not a bug, but it's not a catastrophic bug. It doesn't work on user units and as far as I can see it doesn't work on templated units. So, you have to be root, create a root owned file, and make the typo as root. systemd doesn't fail loudly and falls back to its default behavior as if a user wasn't specified at all. It'd be better if it failed loudly, but it's not that big of a deal.

Not to mention I imagine Poettering is going to get a stern talking to because RHEL7 is patched by default to allow these kinds of usernames (as mentioned in that thread) and you can imagine that there's some kind of proprietary shit out there or noisy customer that uses a name like this, so it'll be fixed at some point. But, again, not that big of a deal.

I'm GenZ and I know this (Upstart, Openrc or SysV btw). Your generation isn't an excuse, just ask the question without trying to find excuses for why you don't have this knowledge.

Do I need to link the amount of catastrophic bugs in the Upstart bug tracker? The cripple even ran into one when NT Technology forced him to run Ubanto on the servers. It's dogshit.
Wow, more dogshit RC scripts that can't even keep track of the real state of a daemon or child processes, complete with the racey behavior that's been hacked around for decades.
lol
I already went over daemontools above. systemd was the only option. You never see UNIX fags complain about Solaris SMF or old Unix greybeards complain about launchd on OS X. It's because they're not as autistic as the NEETs here, of course.

Oh, the only thing that Upstart was good at was a SysV shim, and systemd is better at it especially for obscure compatibility. Aside from that, Upstart jobs were catastrophically broken to the core when it came to anything besides babby's first upstart jobs, one of the developers admitting that it would require a lot of rewriting, so of course it won the fight.
Not to mention the useless cgroups functionality seemingly implemented the way it was because they wanted to be as different from systemd as possible. There's no other way to explain that.

as a sysadmin, i've had the opposite experience.

i ended up switching my personal machines to FreeBSD while my developer colleagues have been slobbering all over Lennart's knob.

No, it was more Spotify and anyone else who wasn't a fucking retard on the mailing list who chimed in.
Anything was better than Upstart, Red Hat even went through the systemd switch from Upstart because of the problems Upstart had, and no one wanted RC script hacks. Dion't pretend you know what you're talking about, because you don/t.

-Massive Shilling
-Backed by Red Hat

rustcuck leave. this is not your home.

That is a feature, not a bug.

It does it for good reasons, like everything else sane. The point is that systemd can do it no problem, and gives me easy to use cgroups administration of services. I have several bits of critical legacy proprietary cruft that double forks, and even if I didn't why should I have to modify every piece of shit that doesn't have an option to run as foreground?

...

I never liked systemd but at least I could rationalize it.

Then they took out the jit compiler(which is what gave them fast boot times, the moment they were above 50% they ditched it)
Then they added binary logs
Then they started replacing utils

I can believe that pottershit is capable of writing an init system but not that he can also do dns, journals, queuing, containers and whatever else he's added in there.

The fucking hubris to think that he can properly implement all of those services AND do them safely in C is laughable.

Being a systemd dev:
1. Break POSIX or create bug
2. Close bug reports by saying "expected behavior" or "you're doing it wrong"
3. Lock discussion after being called out by saying "this is a bug tracker, not a forum"
4. Call everyone trolls
5. Get banned from kernel development by Linus.

What a faggot