View previous topic :: View next topic |
Author |
Message |
yngwin Retired Dev
Joined: 19 Dec 2002 Posts: 4572 Location: Suzhou, China
|
Posted: Tue Oct 02, 2012 8:12 am Post subject: |
|
|
Fitzcarraldo wrote: | My question related to official policy or goals by this distribution's developers (the Gentoo Council), given the statement in the article I referenced. My understanding from reading various posts in the past in these Forums is that the official init system in Gentoo is OpenRC (The Gentoo Handbook, updated September 11, 2012, for example, refers only to OpenRC and not systemd) and will stay that way. The aforementioned IT World article appears to indicate that could change in the near to midterm future. Hence my question. |
The article is wrong. Gentoo is most definitely not switching from OpenRC to systemd. OpenRC is, and for the foreseeable future will remain to be, our default init system. All that happened is that some people maintain the systemd package as an optional alternative for those who want it.
Any suggestion to change the default will be controversial, and suggesting systemd as a replacement extremely so. I would even say it would be a guaranteed recipe for a fork. _________________ "Those who deny freedom to others deserve it not for themselves." - Abraham Lincoln
Free Culture | Defective by Design | EFF |
|
Back to top |
|
|
ryao Retired Dev
Joined: 27 Feb 2012 Posts: 132
|
Posted: Fri Oct 05, 2012 4:51 am Post subject: Re: Gentoo 'lining up behind systemd'? |
|
|
Fitzcarraldo wrote: | http://www.itworld.com/data-centerservers/297977/better-booting
Quote: | It looks like the crowd lining up behind systemd is growing fast. With Fedora, OpenSUSE, Mandiva, Arch, Debian, Gentoo, Frugalware -- Did I forget anyone? And RedHat will probably be joining with the release of RedHat7. Ubuntu seems to be sticking with Upstart. |
Well, is Gentoo 'lining up behind systemd'? I thought the official policy in Gentoo is to stick with OpenRC? |
There are no restrictions on the packages that Gentoo developers can commit to the tree. The only real requirements are that the ebuilds follow Gentoo's QA standards, the package is not malware and someone promises to maintain it.
As far as I know, OpenRC is our official init system. However, there are a few Gentoo developers that are interested by systemd. The ability to use systemd in systemd largely depends on their ability to integrate it in a way that plays nicely with everything else in Gentoo. With that said, only those few developers see value in systemd. If Gentoo ever switches to systemd, it is quite likely that P = NP will have been proven long before the switch. |
|
Back to top |
|
|
Kollin Veteran
Joined: 25 Feb 2006 Posts: 1139 Location: Sofia/Bulgaria
|
Posted: Fri Oct 05, 2012 6:36 am Post subject: |
|
|
May be udev fork will happen after all:
http://article.gmane.org/gmane.linux.kernel/1369384
_________________ "Dear Enemy: may the Lord hate you and all your kind, may you be turned orange in hue, and may your head fall off at an awkward moment."
"Linux is like a wigwam - no windows, no gates, apache inside..." |
|
Back to top |
|
|
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
|
Back to top |
|
|
Kollin Veteran
Joined: 25 Feb 2006 Posts: 1139 Location: Sofia/Bulgaria
|
Posted: Fri Oct 05, 2012 7:40 am Post subject: |
|
|
OMG, i have to try it! _________________ "Dear Enemy: may the Lord hate you and all your kind, may you be turned orange in hue, and may your head fall off at an awkward moment."
"Linux is like a wigwam - no windows, no gates, apache inside..." |
|
Back to top |
|
|
Fitzcarraldo Advocate
Joined: 30 Aug 2008 Posts: 2034 Location: United Kingdom
|
Posted: Fri Oct 05, 2012 8:47 am Post subject: |
|
|
Crikey! He doesn't pull his punches, does he? _________________ Clevo W230SS: amd64, VIDEO_CARDS="intel modesetting nvidia".
Compal NBLB2: ~amd64, xf86-video-ati. Dual boot Win 7 Pro 64-bit.
OpenRC udev elogind & KDE on both.
Fitzcarraldo's blog |
|
Back to top |
|
|
Kollin Veteran
Joined: 25 Feb 2006 Posts: 1139 Location: Sofia/Bulgaria
|
Posted: Fri Oct 05, 2012 9:40 am Post subject: |
|
|
Fitzcarraldo wrote: |
Crikey! He doesn't pull his punches, does he? |
The velvet glove is definitely in the bin _________________ "Dear Enemy: may the Lord hate you and all your kind, may you be turned orange in hue, and may your head fall off at an awkward moment."
"Linux is like a wigwam - no windows, no gates, apache inside..." |
|
Back to top |
|
|
aCOSwt Bodhisattva
Joined: 19 Oct 2007 Posts: 2537 Location: Hilbert space
|
Posted: Fri Oct 05, 2012 10:43 am Post subject: Re: Gentoo 'lining up behind systemd'? |
|
|
ryao wrote: | If Gentoo ever switches to systemd, it is quite likely that P = NP will have been proven long before the switch. |
<trolling>
ryao, on some other place of the internet... wrote: | I have reason to believe that P = NP has been solved |
So... Richard, what's next ?
Riemann hypothesis ?
</trolling> _________________
|
|
Back to top |
|
|
ulenrich Veteran
Joined: 10 Oct 2010 Posts: 1480
|
Posted: Fri Oct 05, 2012 12:56 pm Post subject: Soon we will lining up behind systemd! |
|
|
ryao wrote: | As far as I know, OpenRC is our official init system. |
No more than some preconfigurations. Systemd is two USE flags away: +systemd -consolekit
Quote: | The ability to use systemd in systemd largely depends on their ability to integrate it in a way that plays nicely with everything else in Gentoo. |
Nearly nothing in a free forkable software world "plays nicely with everything else" though. In my humble opinion it is nicer integrated than openrc just yet. After decades of openrc there are discussions how network dependencies should be considered. But systemd comes to accepted conclusions after two years, finally.(*)
Quote: | With that said, only those few developers see value in systemd. |
The interested audience will grow:
1. Nobody having had good experience with systemd will retreat. (unidirectional flow of change)
2. Many developers will explore new areas not existing in openrc world. (nspawn)
3. Features needed for openrc don't have upstream maintainence any more (consolekit)
4. New software will be more and more dependend on systemd, because of possibilities not been available with openrc. For example Kde developers will explore the unreliable startup sequence with systemd user space
@ryao, I can't see any hard facts backing your sided feelings.
(*)PS: Because of the weak concepts openrc represents many controversial cases were swapt back and forth. A software implementing more distinctive concepts can settle down and will have been tested better in a timely manner.(**)
(**)PS: These decisive concepts do have the side effect of systemd developers being unable for compromise. Just seen with Kai Sievers rejecting a firmware bug. Against all traditions of the bazaar this was upsetting Linus to his edge. But isn't it real software engineering? When Con Kolivas compared Solaris with Linux, he also observed this effect of the opensource bazaar! |
|
Back to top |
|
|
khayyam Watchman
Joined: 07 Jun 2012 Posts: 6227 Location: Room 101
|
Posted: Fri Oct 05, 2012 2:35 pm Post subject: |
|
|
ulenrich ...
more of your nonsensical ramblings, "(decisive|distinctive) concepts", yadda, yadda. Your convinced that your statements are meaningful even though entirely lacking a propositon => conclusion ... a meaningful sentence is more than a jumble of words.
Also, when asking for "hard facts" you should be sure to provide them yourself, ie, in the case of "Kai [sic] Sievers rejecting a firmware bug.", which is completely fanciful.
best ... khay |
|
Back to top |
|
|
ulenrich Veteran
Joined: 10 Oct 2010 Posts: 1480
|
Posted: Fri Oct 05, 2012 2:39 pm Post subject: |
|
|
Surely I meant:
Sievers rejected an udev fix regarding firmware loading.
I am not an english speaker, hope I express somehow understandable ....
And I was not that much argueing but trying to show ryao having just feelings.
Which I see the other way round:
freedom, bazaar, week concepts
- versus -
strong principles, software engineering
which I thankfully appreciate - in this case .... |
|
Back to top |
|
|
khayyam Watchman
Joined: 07 Jun 2012 Posts: 6227 Location: Room 101
|
Posted: Fri Oct 05, 2012 7:05 pm Post subject: |
|
|
ulenrich wrote: | Surely I meant: Sievers rejected an udev fix regarding firmware loading. |
... whatever was ment, and whomever is on the wrong side in that discussion, Linus is absolutely correct when he writes "[i]t's more robust because we have maintainers that care, and because we know that regressions are not something we can play fast and loose with. If something breaks, and we don't know what the right fix for that breakage is, we *revert* the thing that broke."
ulenrich wrote: | I am not an english speaker, hope I express somehow understandable .... |
Its not so much your English, but your lack of coherent argument, stating that "[b]ecause of the weak concepts openrc represents many controversial cases were swapt back and forth" is not only poor English, but presents nothing other than equating openrc with "weak concepts", its in the realm of vague associative semantics. Similarly when systemd is mentioned its in terms of "distinctive" and "decisive" "concepts", etc, but given that these claims are not argued but simply presented, its nothing more than jingoism.
ulenrich wrote: | And I was not that much argueing but trying to show ryao having just feelings. |
... and there I was just waiting for you to move the goal posts, and ta-da! Can you point to one statement made by ryao that you think are illustrative of these "feelings"?
Now, a quick change of course, readers please don't blink you may miss the connecting train of thought ...
ulenrich wrote: | Which I see the other way round: freedom, bazaar, week concepts - versus - strong principles, software engineering |
Questions of the validity of these distinctions aside, can you please explain how this relates to the statement made by ryao's, specifically in what way was the case made that you are seeing "the other way round"?
best ... khay |
|
Back to top |
|
|
ulenrich Veteran
Joined: 10 Oct 2010 Posts: 1480
|
Posted: Fri Oct 05, 2012 7:28 pm Post subject: |
|
|
Quote: | Can you point to one statement made by ryao that you think are illustrative of these "feelings"? |
ryao answers the negative connotated question "lining up behind systemd" with reassuring jokes:
- ability to integrate (systemd is integrated since some time)
- plays nicely with everything else (not one little peace of software plays well with everything else)
- few developers (everything starts with few)
- P = NP (every one heaving read an informatic book knows about this)
I have done more than jokes. I don't want to give an akademic rationale why systemd has stronger software concepts implemented. But I gave a hint what sympton this shows. Your critics would fit better the other direction in this propaganda thread
Last edited by ulenrich on Fri Oct 05, 2012 8:06 pm; edited 1 time in total |
|
Back to top |
|
|
NeddySeagoon Administrator
Joined: 05 Jul 2003 Posts: 54099 Location: 56N 3W
|
Posted: Fri Oct 05, 2012 7:58 pm Post subject: |
|
|
ulenrich,
ulenrich wrote: | I don't want to give an akademic rationale why systemd has stronger software ... |
Please do. Google cannot find any such rationale and so far you hove only posted empty and apparently exaggerated assertions.
It would be good to read some facts about systemd and a contrast with the facts of one or more other init systems.
Of course, its possible you don't have the akademic rationale you refrain from posting.
Perhaps it does not exist anywhere. _________________ Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
Back to top |
|
|
khayyam Watchman
Joined: 07 Jun 2012 Posts: 6227 Location: Room 101
|
Posted: Fri Oct 05, 2012 8:29 pm Post subject: |
|
|
ulenrich wrote: | Quote: | Can you point to one statement made by ryao that you think are illustrative of these "feelings"? |
ryao answers the negative connotated question "lining up behind systemd" with reassuring jokes: |
ulenrich ... I'm having a hard time detecting these jokes, or the attempt at reassurance.
ulenrich wrote: | - ability to integrate (systemd is integrated since some time) |
That is not what was said ... "the ability to use systemd [...] largely depends on their ability to integrate it in a way that plays nicely with everything else in Gentoo".
ulenrich wrote: | - plays nicely with everything else (not one little peace of software plays well with everything else) |
Thats certainly true of systemd, but this all depends on what we mean by "plays nicely", anyhow, again, see the the above. From the perspective of "everything else in Gentoo" then requiring systemd is not "playing nicely with everything else".
ulenrich wrote: | - few developers (everything starts with few) |
Such a use of truisms doesn't change the fact that there are just "a few".
ulenrich wrote: | - P = NP (every one heaving read an informatic book knows about this) |
ahhh ... but, sorry, what? Its an unsolved problem, and thats how I read it in the context of ryao's statements. So, what does "everyone [...] know"?
ulenrich wrote: | I have done more than jokes. I don't want to give an akademic rationale why systemd has stronger software concepts implemented. But I gave a hint what sympton this shows. |
No, you have made no case at all, your simply presenting one of your incoherent ramblings.
best ... khay |
|
Back to top |
|
|
ulenrich Veteran
Joined: 10 Oct 2010 Posts: 1480
|
Posted: Fri Oct 05, 2012 8:38 pm Post subject: |
|
|
@NeddySeagoon, to your pleasure a few examples. These citations are from http://lwn.net/Articles/389149/ in 2010. This shows strong comittment to early on software concepts:
"Systemd uses cgroups to track daemon processes that it has created; they allow these processes to be monitored even if they use the familiar daemon tricks for detaching themselves from their parents."
"Cgroups would also make it easy for system administrators to set up specialized sandboxes for daemons to run in." (systemd-nspawn)
"just wait until somebody tries to connect to the service and start it on demand. ... The systemd approach to managing daemons allows a fair amount of boilerplate code to be removed" |
|
Back to top |
|
|
ulenrich Veteran
Joined: 10 Oct 2010 Posts: 1480
|
Posted: Fri Oct 05, 2012 8:47 pm Post subject: |
|
|
khayyam wrote: | ulenrich ... I'm having a hard time detecting these jokes, or the attempt at reassurance.
ulenrich wrote: | - P = NP (every one heaving read an informatic book knows about this) | ahhh ... but, sorry, what? Its an unsolved problem, and thats how I read it in the context of ryao's statements. So, what does "everyone [...] know"? |
You clearly have a big problem with humour. Ryaos original was:
Quote: | If Gentoo ever switches to systemd, it is quite likely that P = NP will have been proven long before the switch. |
|
|
Back to top |
|
|
NeddySeagoon Administrator
Joined: 05 Jul 2003 Posts: 54099 Location: 56N 3W
|
Posted: Fri Oct 05, 2012 8:48 pm Post subject: |
|
|
ulenrich,
There are very few facts there and no contrasts with other init systems except to say its different.
As I have already said, different != better. _________________ Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
Back to top |
|
|
ulenrich Veteran
Joined: 10 Oct 2010 Posts: 1480
|
Posted: Fri Oct 05, 2012 9:05 pm Post subject: |
|
|
NeddySeagoon wrote: | There are very few facts there and no contrasts with other init systems except to say its different.
As I have already said, different != better. |
Surely you will be able to implement following with openrc:
- catch daemons, which behave badly
- having this you will be able to provide better virtualisation features
- replace consolekit without its flaws and its hundreds of dead entries
- having this you can offer automatic multiseat features
- implement dbus better integrated
- implement inetd ondemand capabilities
- offer dependend service starting facilities for user space
I know of new openrc efforts to reach some these goals, really But the problem might be openrc has maximal 10 percent of the fast pace and developer resources of systemd. What is plant and actively worked on:
http://cgit.freedesktop.org/systemd/systemd/tree/TODO
Last edited by ulenrich on Fri Oct 05, 2012 9:35 pm; edited 3 times in total |
|
Back to top |
|
|
khayyam Watchman
Joined: 07 Jun 2012 Posts: 6227 Location: Room 101
|
Posted: Fri Oct 05, 2012 9:06 pm Post subject: |
|
|
ulenrich wrote: | You clearly have a big problem with humour. Ryaos original was:
Quote: | If Gentoo ever switches to systemd, it is quite likely that P = NP will have been proven long before the switch. |
|
ulenrich ... and you clearly have a big problem following a line of thought. Such decisions should not be imposed on the user, and certainly a "switch to systemd" would be such a impossiton, or did you miss "plays nicely with everything else"? So, I don't see this as a joke, but a statement in line with gentoo's attempt to leave such decsions to the user.
Otherwise, nice of you to really focus in on the real guts of the question at hand, and leave the easy stuff for the birds.
best ... khay |
|
Back to top |
|
|
ulenrich Veteran
Joined: 10 Oct 2010 Posts: 1480
|
Posted: Fri Oct 05, 2012 9:14 pm Post subject: |
|
|
khayyam wrote: | ... and you clearly have a big problem following a line of thought. Such decisions should not be imposed on the user |
Upps, should I have missed a kind of second level humour?
I totally regret my statement about you! |
|
Back to top |
|
|
khayyam Watchman
Joined: 07 Jun 2012 Posts: 6227 Location: Room 101
|
Posted: Fri Oct 05, 2012 9:32 pm Post subject: |
|
|
ulenrich wrote: | - catch daemons, which behave badly
- having this you will be able to provide better virtualisation features
- replace consolekit without its flaws and its hundreds of dead entries
- having this you can offer automatic multiseat features
- implement dbus better integrated
- implement inetd ondemand capabilities
- offer dependend service starting facilities for users (kde) |
ulenrich ... while some of these features seem reasonable, none of them are necessarily solved by systemd, and if potencially solved by systemd should none the less be optional. After all, its not just a question of systemd but that systemd is going all out to control init via a coup de main, as higher level functions will depend on systemd, and systemd only.
Additionally, this is a list of proposed benifits, but what is lost in all of this is something much more valuable than the proposed advantages ... so I say down with fancy pants!.
ulenrich wrote: | The complete list of todo: |
They omited to state: "abandon for the next greastest thing from the great idea fairy in two years, so we can put our mistakes behind us".
best ... khay |
|
Back to top |
|
|
ulenrich Veteran
Joined: 10 Oct 2010 Posts: 1480
|
Posted: Fri Oct 05, 2012 9:46 pm Post subject: |
|
|
khayyam wrote: | and if potencially solved by systemd should none the less be optional. After all, its not just a question of systemd but that systemd is going all out to control init via a coup de main, as higher level functions will depend on systemd, and systemd only.
Additionally, this is a list of proposed benifits, but what is lost in all of this is something much more valuable | Yes, a kind of Coup d'Etat. My previous statement: Quote: | freedom, bazaar, week concepts
- versus -
strong principles, software engineering |
was NOT meant negative in the sense of weak = bad programming
No. Weakness of principle has a freedom possibility for the user!
I thankfully appreciate in this special case the lost freedom!
Rationale: Decision in this case is an opportunity to develop further!
By the way: khayyam wrote: | while some of these features seem reasonable, none of them are necessarily solved by systemd | I consider all these listed features solved, necessarily there will be further optimizations... |
|
Back to top |
|
|
khayyam Watchman
Joined: 07 Jun 2012 Posts: 6227 Location: Room 101
|
Posted: Fri Oct 05, 2012 10:16 pm Post subject: |
|
|
ulenrich wrote: | Yes, a sort of Coup d'Etat. My: freedom, bazaar, week concepts - versus - strong principles, software engineering |
ulenrich ... except that complexity is not a "strong principle" nor a good principle in engineering. Also, if you engineer out the engineers, that is, your engineers are vertical integrators, and introduce wider fragmentation (linux is not the only *nix) then your ecosystem will exhibit the same effects produced by invasive species, monocultures, etc.
ulenrich wrote: | I thankfully appreciate in this special case the lost freedom! Rationale: Decision in this case is an opportunity to develop further! |
I really don't think you understand what it is your advocating quite frankly ...
best ... khay |
|
Back to top |
|
|
M Guru
Joined: 12 Dec 2006 Posts: 432
|
Posted: Fri Oct 05, 2012 10:50 pm Post subject: |
|
|
Unix philosophy: Write programs that do one thing and do it well. Write programs to work together. Write programs to handle text streams, because that is a universal interface.
systemd violates this and much more. They merged udev, and soon it will depend on systemd? Consolekit is deprecated and is now part of systemd? But every DE now has support for consolekit. Binary log files!? Gnome will depend on systemd? Not to mention XML config files, etc, etc...
I don't see nothing better there, a little faster boot, features nobody needs, bla bla, and in the meantime they polluted everything, and want everything to depend on systemd.
Quote: | - replace consolekit without its flaws and its hundreds of dead entries |
And who created consolekit? And why was it needed anyway?
Quote: | - implement dbus better integrated |
And why is dbus needed on a server?
I use Gentoo because I can stay away from *kits and similar nonsense. For auto mount I use sys-apps/uam, there is no need for a daemon! |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|