Apparently I know a shit more about Debian that lkcl does. He claims those are packages that depend on libsystemd0, which largely exists to make sure packages don't depend on systemd. And he gets it wrong!
Just take the first example on that list, "0ad":
# apt-cache show 0ad | grep systemd [ crickets ]
If you want to know what depends on libsystemd (but why care?)
Which is precisely what people are complaining about - more and more packages (like gdm3, gnome, etc) *needing* (being written to require) systemd... or some "shim" to get around it, which starts making packages less and less able to be ported to other OS's (BSDs, Solaris...).
gdm3 needs some functions that used to be provided by consolekit. Until very recently nobody could be bothered to maintain consolekit.
(Of course/sbin/init also gets the exit status -- it's just that/sbin/init has no concept of services, the whole sysvinit system just pretends to manage services but hardly uses/sbin/init at all).
Actualy that's not true for systemd -- if a process double forks and exits systemd can still capture tbe exit code -- that's (part of) what cgroups is for.
I can find no evidence of the mysterious 'dropped syslog' claims, nobody ever seems to have reported it as a bug and it doesn't happen on any system I have access to.
It'll also need an IOMMU and a driver which prohibits the NIC from stepping out of line, or a NIC with open firmware. Otherwise, someone could (theoretically) own your NIC and then browse your memory from it.
On the (as yet inexistant) Neo900 the wireless module is a USB device. It doesn't get to access the memory if the CPU doesn't want it to.
Slashdot Mirror
Pointing out that someone is a troll is flamebait now?
GTK+ is rife with serious problems. The first is that it's affiliated with the GNOME crew.
Troll.
Another reason to use an encrypted root.
Of course there are ways to get round that, too, but the size of the hard drive firrnware is limited.
Yup, sounds exactly like the way we're "dealing" with global warming.
You know, your google-foo is lacking. http://devzing.com/blog/index.php/how-to-delete-bugs-from-bugzilla/.
So you work for an organism that employs sociopaths. Nice for you.
Luckily some of us are not in that position.
Ah, I love the smell of conspiracy theories in the morning.
Ever seen the bug list for consolekit?
No.
"somebody" (i.e. Lennart Poettering) went to some trouble to make sure that gdm3 still worked with consolekit.
"somebody" (i.e. you) didn't bother to fix the bugs in consolekit, so nobody wants to use it.
Stop whining. Work.
No, naming every release after the same animal is "professional".
Then you don't know Debian either. 4583 packages depend on "systemd" - see http://lkcl.net/reports/removi...
Apparently I know a shit more about Debian that lkcl does. He claims those are packages that depend on libsystemd0, which largely exists to make sure packages don't depend on systemd. And he gets it wrong!
Just take the first example on that list, "0ad":
If you want to know what depends on libsystemd (but why care?)
Alarming? Not.
No it isn't.
Why bother posting such nonsense?
Yeah, trying to work out the backronym for DIE MUTHERFUKER was a pain.
Should be embarrasingly easy.
Why haven't you done it if you want it?
The "big" "error" was to not write text logs and a binary index/metadata file.
However it wouldn't have stopped the trolls, they are clearly insane.
(For me "journalctl" isn't the win, it's "systemctl status").
Which is precisely what people are complaining about - more and more packages (like gdm3, gnome, etc) *needing* (being written to require) systemd... or some "shim" to get around it, which starts making packages less and less able to be ported to other OS's (BSDs, Solaris...).
gdm3 needs some functions that used to be provided by consolekit. Until very recently nobody could be bothered to maintain consolekit.
systemd provides those functions.
And that's systemd's fault?
Stop whining. Work.
That's exactly what it does.
No, it isn't.
Or, hell... just plug it into Google to save some time, since all the heavy lifting has been done for you time and time again, you astroturfing troll.
Plug what in? I've tried every way of posing the question I can think of and it all comes back to here or that one incorrect reddit post.
People are tired of submitting bug reports to a project that doesn't address them
Name one.
(Fucking slashdot went down again while I was posting this reply. Did they install systemd or something?)
If you want text logs install syslogd.
Is that hard?
Would you rather have stderr scrolling up the (probably not attached) screen in some remote data centre or have the message written to syslog?
(Of course /sbin/init also gets the exit status -- it's just that /sbin/init has no concept of services, the whole sysvinit system just pretends to manage services but hardly uses /sbin/init at all).
Actualy that's not true for systemd -- if a process double forks and exits systemd can still capture tbe exit code -- that's (part of) what cgroups is for.
I can find no evidence of the mysterious 'dropped syslog' claims, nobody ever seems to have reported it as a bug and it doesn't happen on any system I have access to.
Just talking about Debian, because that's tbe only one I know.
How many packages depend on systemd?
Answer -- one, gummiboot.
Some vandalism.
It'll also need an IOMMU and a driver which prohibits the NIC from stepping out of line, or a NIC with open firmware. Otherwise, someone could (theoretically) own your NIC and then browse your memory from it.
On the (as yet inexistant) Neo900 the wireless module is a USB device. It doesn't get to access the memory if the CPU doesn't want it to.
UKIP could those problems.
That's a feature, not a bug.
So you'd prefer the job of a full time spreader of FUD?
(setq sarcasm 'on) Well, that will certainly help me convince the boss to upgrade our infrastructure. (setq sarcasm nil)
What crappy programming style. setq? Really? The proof that a real Fortran programmer can wriet Fortran in any language.