Systemd Gets A Stable Release Repository, Backports

[gens]

Here is part of the quote you left out (my emphasis):

"Yeah, I see a few things coming down the line that may cause a shakeup to our usual way of doing things, and could force Slackware to become, well, perhaps less UNIX-like. I guess the two big ones that are on the horizon are Wayland and systemd. Whether we end up using them or not remains to be seen. It's quite possible that we won't end up having a choice in the matter depending on how development that's out of our hands goes. It's hard to say whether moving to these technologies would be a good thing for Slackware overall. "

http://www.linuxquestions.org/questions/interviews-28/interview-with-patrick-volkerding-of-slackware-949029/

Slackware will move in the same direction as every other Linux distro: Sysvinit and X are on minimum life support right now, with systemd and Wayland being the future of Linux. Patrick is perhaps sceptical, but Slackware is such a small distro that it won't have the manpower to even maintain status quo. There simply doesn't seem to be any developer impetus to maintain a full featured Sysvinit Linux distro. Maybe some Debian Sysvinit derivative will be made, but it will mostly be a server version, since DE support for non-systemd systems have been bit-rotting for years now.

here's another part

"To the typical end user, if this results in a faster boot then mission accomplished. With udev being phased out in favor of systemd performing those tasks we'll have to make the decision at some point between whether we want to try to maintain udev ourselves, have systemd replace just udev's functions, or if we want the whole kit and caboodle. Wayland, by comparison, seems fairly innocuous, assuming that they'll be able to implement network transparency either directly or through some kind of add-on compatibility layer. Again, another thing that most desktop users don't have a lot of use for but many users can't do without. I like X11, and would probably stick with it if moving to Wayland meant losing that feature, even if Wayland's rendering method carried with it some benefits like reduced rendering artifacts or increased video performance. I guess we'll just have to see what the overall benefit is when it's far enough along to make such comparisons."

so udev
at the time of the interview (06.07.2012) there were not as many alternatives to upstream udev

later logind might be problematic, but there will probably be a good enough alternative (also systemd will release a stable API version, maybe)


slackware is not small and is kept quite up to date (newest alsa, gcc, firefox etc and last stable kernel, xorg etc.)
i suggest doing some research before assuming
more then that i suggest not putting things into other peoples mouths

[gens]

Wayland is against Unix too now?

no, on the contrary
it does one thing and does it well

also interview was two years ago, wayland was fairly new then

[interested]

here's another part

"To the typical end user, if this results in a faster boot then mission accomplished. With udev being phased out in favor of systemd performing those tasks we'll have to make the decision at some point between whether we want to try to maintain udev ourselves, have systemd replace just udev's functions, or if we want the whole kit and caboodle. Wayland, by comparison, seems fairly innocuous, assuming that they'll be able to implement network transparency either directly or through some kind of add-on compatibility layer. Again, another thing that most desktop users don't have a lot of use for but many users can't do without. I like X11, and would probably stick with it if moving to Wayland meant losing that feature, even if Wayland's rendering method carried with it some benefits like reduced rendering artifacts or increased video performance. I guess we'll just have to see what the overall benefit is when it's far enough along to make such comparisons."

so udev
at the time of the interview (06.07.2012) there were not as many alternatives to upstream udev

later logind might be problematic, but there will probably be a good enough alternative (also systemd will release a stable API version, maybe)


slackware is not small and is kept quite up to date (newest alsa, gcc, firefox etc and last stable kernel, xorg etc.)
i suggest doing some research before assuming
more then that i suggest not putting things into other peoples mouths

Slackware _is_ a small distro and I am not just talking user base and number of packages, but also in terms of developers. Nothing wrong in that, just a fact. It has been a niche distro for more than a decade (and yes, I have installed Slackware from floppies before that).

Regarding the fork of "udev" and "logind" then it is highly symptomatic for the lack of non-systemd development, that the only vaguely viable alternatives to systemd, are in fact forked systemd code made by Lennart Poettering and other systemd developers.

The "logind" fork is just a temporary band aid; it will stand still while upstream DE's start to develop features that aren't compatible with the old forked 204 version of systemd.
Then there is cgroups and kdbus etc. together with Wayland. It will become harder and harder not to support systemd on any full featured Linux distro. It isn't impossible, but there just doesn't seem to be that much developer interest in doing so.

[gens]

Slackware _is_ a small distro and I am not just talking user base and number of packages, but also in terms of developers. Nothing wrong in that, just a fact. It has been a niche distro for more than a decade (and yes, I have installed Slackware from floppies before that).

Regarding the fork of "udev" and "logind" then it is highly symptomatic for the lack of non-systemd development, that the only vaguely viable alternatives to systemd, are in fact forked systemd code made by Lennart Poettering and other systemd developers.

The "logind" fork is just a temporary band aid; it will stand still while upstream DE's start to develop features that aren't compatible with the old forked 204 version of systemd.
Then there is cgroups and kdbus etc. together with Wayland. It will become harder and harder not to support systemd on any full featured Linux distro. It isn't impossible, but there just doesn't seem to be that much developer interest in doing so.

so from your point of view there are only 3 big distros (ubuntu, debian, fedora and what comes of it)
since slackware is bigger then all the others (over 6%)

fork of udev shows that people think making a key part of linux tied into a specific boot system is dumb
and udev is a key part
and there was nothing wrong with it
Kay Sievers took maintainership of udev
Linus himself said it went to shit then
and no, nothing is changed in udev itself, it is there for systemd's benefit only

and yes, desktop people have a loooong history of not cooperating
that is their problem
i think it is 'cuz they all artistic and want to do their thing

it does not mean we should all praise GNOME for being dependent on systemd
funny enough the official reasoning is multiseat support, that you can't do without a network in systemd


cgroups are, as i become bored of repeating on phoronix, simple and easy
and in my opinion not that useful on the desktop, but that is my opinion

point is it can all be done with simple daemons
and yet systemd is being forced for no clear reason

[Teho]

so from your point of view there are only 3 big distros (ubuntu, debian, fedora and what comes of it)
since slackware is bigger then all the others (over 6%)

Source? That's most definetly not true. Try openSUSE/SLE, RHEL/CentOS or Gentoo.

fork of udev shows that people think making a key part of linux tied into a specific boot system is dumb

No it doesn't. It only means that some people think so and the fact that no one, not even Gentoo has adopted eudev should tell you something about the importance of the project.

Kay Sievers took maintainership of udev
Linus himself said it went to shit then
and no, nothing is changed in udev itself, it is there for systemd's benefit only

Seriously what are you talking about? Greg K-H transfered the maintainership of udev to Kay Sievers in 2005

Kay has been instrumental in getting udev as
feature complete and stable and actually working well. Without his
help, it wouldn't be the program it is today. He also was the one who
implemented the persistant naming policy for disks, which is now in all
of the major Linux distributions (and a few minor ones.)

I'll still be around and doing minor work on udev, but Kay is the one
who is now in charge, and I know it is in good hands.

Also from the systemd/udev merge announcement:

We are about to merge the udev sources into the systemd source tree. After that, the next version of systemd will continue with udev?s version numbering, i.e. jump immediately from 45 to 184.

Great job on this, it makes a lot of sense.

So the creator, original maintainer, second most active udev developer and the current maintainer and the most active developer, responsible for virtually all udev developement, both agreed that merging was the right choise. It has allowed a lot of code sharing between systemd which then again makes the maintanance of udev easier.

funny enough the official reasoning is multiseat support, that you can't do without a network in systemd

What....

and yet systemd is being forced for no clear reason

No one is forcing you to use systemd.

[DeepDayze]

And that is a good thing!

Slackware does well because it does not always get involved in petty politics. I have cut my Linux teeth on Slackware and gravitated towards Debian because how conservative it is as well as its robust package system.

[gens]

Source? That's most definetly not true. Try openSUSE/SLE, RHEL/CentOS or Gentoo.


No it doesn't. It only means that some people think so and the fact that no one, not even Gentoo has adopted eudev should tell you something about the importance of the project.

Seriously what are you talking about? Greg K-H transfered the maintainership of udev to Kay Sievers in 2005



Also from the systemd/udev merge announcement:





So the creator, original maintainer, second most active udev developer and the current maintainer and the most active developer, responsible for virtually all udev developement, both agreed that merging was the right choise. It has allowed a lot of code sharing between systemd which then again makes the maintanance of udev easier.

What....


No one is forcing you to use systemd.

linuxcounter.net, currently down so no direct link

troll

troll

out of context or link troll

same as above

it is forced, troll

[Gusar]

cgroups are, as i become bored of repeating on phoronix, simple and easy

You have to keep repeating it because it's just words with nothing to back them up. Where's your cgroup manager? And with which service system is it integrated, to provide the seamless service control that systemd does?

See, cgroups have been around for some time. But until systemd, they weren't mainstream. Those who wanted to use them came up with their own homegrown solution that probably only works for their specific use-case.

point is it can all be done with simple daemons

Again, where are they then? Why has no one but the systemd devs provided a solution for the things DE devs want?

and yet systemd is being forced for no clear reason

There is a very clear reason - you haven't given us alternatives! Why you? Because you're the one with the big words on how simple everything is.

troll

troll

out of context or link troll

same as above

it is forced, troll

So this is your MO now? When you get confronted with facts that contradict you, you yell "troll"?

[Teho]

linuxcounter.net, currently down so no direct link

You do understand that it's not a realible metric, right?

It was started as a "for fun" project to find out how many Linux users there are worldwide. The basic idea is for people to register themselves as being a Linux user. Of course, this way you won't get all Linux users counted as not every Linux user will register himself at the Linux Counter site. Thus, the only way to "know" the number of Linux users worldwide, is to make a guess, preferably a not-too-wild guess of the number of Linux users. Not making wild guesses there is only one way to go: statistics. And so, there we are.

quoteout of context or link troll

It's not. Here's the source on the later one.

it is forced, troll

Yeah... how about you explain how that works.

[gens]

You have to keep repeating it because it's just words with nothing to back them up. Where's your cgroup manager? And with which service system is it integrated, to provide the seamless service control that systemd does?

See, cgroups have been around for some time. But until systemd, they weren't mainstream. Those who wanted to use them came up with their own homegrown solution that probably only works for their specific use-case.

Again, where are they then? Why has no one but the systemd devs provided a solution for the things DE devs want?

There is a very clear reason - you haven't given us alternatives! Why you? Because you're the one with the big words on how simple everything is.



So this is your MO now? When you get confronted with facts that contradict you, you yell "troll"?

quoting things in separations kills context, thus killing the meaning of the things quoted

if you read this whole conversation that i was not having with you
you would see that it started by one person misinterpreting what a well respected BDFL of a well respected distribution stated a long time ago
it was not for troll bait, and yet there were trolls

there are many opinions
you can have your own one
if you plan on sharing it with others, good
don't share it with me unless you want to hear mine
simple as that

only thing systemd brings to the desktop is logind that does some basics

if i personally wanted cgroups on my desktop, i would put things into cgroups using the shell and the cgroups filesystem
if it were up to systemd way http://www.linux.com/news/featured-b...roups-redesign
i would not be able to do it so simply since the filesystem would be removed
i would have to write a C program to do something that usually takes about 20 sec to do
also with the filesystem being there it can be easily done for background services in scripts, be they shell, python, java, ruby, whatever

but "raw nobs to userspace"... filesystems have permissions you know
and if a hacker gets root privileges you are done anyway, no matter how you complicate things

just try it out, you might find cgroups fun
think about the usefulness in some special circumstances