Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Gentoo folks, this is your wakeup call(sysd-udev)
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3  Next  
This topic is locked: you cannot edit posts or make replies.    Gentoo Forums Forum Index Other Things Gentoo
View previous topic :: View next topic  
Author Message
Anon-E-moose
Watchman
Watchman


Joined: 23 May 2008
Posts: 6095
Location: Dallas area

PostPosted: Tue Jul 08, 2014 11:59 am    Post subject: Reply with quote

TobiSGD wrote:
Exactly what I said: Sievers patches are not allowed into the kernel until he fixes his behavior when it comes to bugfixing. Only that they may also be allowed when a distro has tried them and they are deemed as stable enough /kdbus patches, not systemd).


Since systemd and kdbus are in lockstep and have to be able to talk to each other properly
and systemd has to be stable to use a stable kdbus, as its the only client of kdbus right now,
it's effectively the same as djdunn said.

This has a little more detail and why djdunn said what he did.
http://igurublog.wordpress.com/2014/04/03/tso-and-linus-and-the-impotent-rage-against-systemd/
And yes I realize that the blog is inflammatory but the facts that he brings our are true and that's the meat of it.
_________________
PRIME x570-pro, 3700x, 6.1 zen kernel
gcc 13, profile 17.0 (custom bare multilib), openrc, wayland
Back to top
View user's profile Send private message
TobiSGD
n00b
n00b


Joined: 14 Apr 2014
Posts: 36

PostPosted: Tue Jul 08, 2014 5:41 pm    Post subject: Reply with quote

Anon-E-moose wrote:
TobiSGD wrote:
Exactly what I said: Sievers patches are not allowed into the kernel until he fixes his behavior when it comes to bugfixing. Only that they may also be allowed when a distro has tried them and they are deemed as stable enough /kdbus patches, not systemd).


Since systemd and kdbus are in lockstep and have to be able to talk to each other properly
and systemd has to be stable to use a stable kdbus, as its the only client of kdbus right now,
it's effectively the same as djdunn said.
If they are in lockstep or not doesn't matter. What Linus said was: No patches from you for the kernel until you fixed your behavior. Distros can test those patches (again kdbus, not systemd) and if they deem them stable it might be considered to let them into the kernel. That is not what djdunn said.
But anyways, this is nitpicking and to me it is clear that kdbus will come anyways and if people adopt it systemd will be de-facto the only init system you can run, unless someone comes up with his own kdbus userspace. Sadly, I am not a developer, otherwise I would start to help with that right now.

Quote:
This has a little more detail and why djdunn said what he did.
http://igurublog.wordpress.com/2014/04/03/tso-and-linus-and-the-impotent-rage-against-systemd/
And yes I realize that the blog is inflammatory but the facts that he brings our are true and that's the meat of it.
Sorry, but I stopped reading when the author claimed that systemd's configuration files are in XML and Javascript. It is one thing to be inflammatory, but being outright misinformed and then being inflammatory is nothing but trolling.
Back to top
View user's profile Send private message
TomWij
Retired Dev
Retired Dev


Joined: 04 Jul 2012
Posts: 1553

PostPosted: Tue Jul 08, 2014 6:20 pm    Post subject: Reply with quote

khayyam wrote:
TomWij ... how is that a "turnaround of words"? [..]

Upon a more thorough read that includes your response, it is not that; an ambiguous sentence was misread, which gives a false impression that you were denoting upstream as the community.

khayyam wrote:
You had written "these questions depend on how much we as a community contribute" to which I replied that some of us, in the wider scheme of things, have already contributed an awful lot ... and having done so find this equity owned by a "community" in which we are expected to see the "choice" of others to pollute the water supply as the same as our "choice" to drink it or not. If this really is a *community* then I expect that this community has more respect for my choices, and doesn't actively work toward the diminution of that choice. I pointed to that previous discussion because it underscores the issue ... "[...] the choice of systemd involves repercussions that are not relevant in the choice of openrc (systemd will similarly effect those who don't choose it) so they are not relationally comparable choices as one has negatives that the party not choosing will none the less have as an outcome." The analogy I used at the time was clear and setup to counter the idea that "systemd is a choice that will coexist with other choices". All of that was writ large at the time, and as I wrote back in Oct 2012 "no one needs a crystal ball to read upstreams intentions". None of this seemed to make any difference to "the community" because the entire issue was framed as though systemd was something that was entirely benign, users could choose it or not ... everyone gets what they want.

As to what "we as a Gentoo community want", I'm sorry but this very same community wasn't the least bit concerned about adopting systemd as a "choice" even when it was perfectly clear that this choice would effect those who didn't choose it. If this idea of community means that I can pollute the water supply without regard for others in my community then it isn't much of a community.

There is a lot of contribution that makes other choices able to exist in the Portage tree, but there could be more contribution to further ensure the future; the equation is simple, if someone stops supporting something then someone else needs to start supporting it for it to remain usable in the future.

You appear to wonder about limitations of choice, but given the way Gentoo works I don't see such limitations. Gentoo does provide choice in a different way than how you describe it, given that limitations (and thus implied choices) are kept to a minimum; as GLEP 39 specifies that "Note that this GLEP does not provide for a way for the community at large to block a new project, even if the comments are wholly negative.", as well as "Global issues will be decided by an elected Gentoo council.". This effectively boils down to that most of the limitations that can exist are those that are voted on by the Gentoo Council. Looking at the history of what has been decided on; there are none that involve enforcing systemd, apart from perhaps the need for an initramfs due to the lack of manpower to support systems without it.

What you label to be framed as coexistence ideas is what Gentoo is all about at its core, as besides GLEP 39 can also be read about in its about and philosophy page; contrary to other distributions, we generally leave the choice to the users instead of picking it instead of them. Given this, accepted a "non-systemd-only" or "systemd-only" Gentoo is something I don't ever see the Gentoo Council do; such choices have a lot of implications, which among other things could result in a community split.

khayyam wrote:
TomWij wrote:
Given the volume of these piss pointing discussions, as well as those already working on other options; there is definitely enough equity to accomplish that, if they step it up...
"piss pointing"? Your mixing things that were clearly delineated ... [...]

It is a way to write it short, you can very well drop the words if you don't agree on the shortened version as the sentence part is about the volume of them; thus there is enough equity to accomplish it given a step up.
Back to top
View user's profile Send private message
miket
Guru
Guru


Joined: 28 Apr 2007
Posts: 483
Location: Gainesville, FL, USA

PostPosted: Tue Jul 08, 2014 7:22 pm    Post subject: Reply with quote

TobiSGD wrote:
Quote:
This has a little more detail and why djdunn said what he did.
http://igurublog.wordpress.com/2014/04/03/tso-and-linus-and-the-impotent-rage-against-systemd/
And yes I realize that the blog is inflammatory but the facts that he brings our are true and that's the meat of it.
Sorry, but I stopped reading when the author claimed that systemd's configuration files are in XML and Javascript. It is one thing to be inflammatory, but being outright misinformed and then being inflammatory is nothing but trolling.

Well, not quite. In place where he makes those mentions,
Quote:
His discussion is not a tirade against systemd, but he too warns of the complexity of the system and its XML (even javascript-based?) configuration files
he's summarizing what Theodore Ts'o wrote about the experience he had with the system of systemd, polkit, and Network Manager. He wrote of this system in the paragraph just before the one I quoted. Sure (at least as of today), systemd does not use either of XML or Javascript, and Network Manager doesn't either. PolicyKit is a different matter: it uses XML and (I roll my eyes every time I think of it) Javascript.

It is possible to run systemd without PolicyKit, but in most distros that's a tricky thing to do. I don't think the article is off-base on this point at all.
Back to top
View user's profile Send private message
Tractor Girl
Apprentice
Apprentice


Joined: 16 May 2013
Posts: 159

PostPosted: Tue Jul 08, 2014 8:44 pm    Post subject: Reply with quote

krinn wrote:

Don't you think it might be time to make eudev the default virtual provider instead of udev?

+1
The Doctor wrote:
Taking it one step farther would there be any loss dropping udev from the tree completely? I mean what does udev do that eudev doesn't do?

++1
Back to top
View user's profile Send private message
Anon-E-moose
Watchman
Watchman


Joined: 23 May 2008
Posts: 6095
Location: Dallas area

PostPosted: Tue Jul 08, 2014 8:57 pm    Post subject: Reply with quote

miket wrote:
I don't think the article is off-base on this point at all.


No it wasn't and I did state up front that it was somewhat inflammatory.

Nonetheless if one reads things properly, then what djdunn said was correct in essence and TobiSGD wasn't more than technically correct.

TobiSGD wrote:
If they are in lockstep or not doesn't matter. What Linus said was: No patches from you for the kernel until you fixed your behavior. Distros can test those patches (again kdbus, not systemd) and if they deem them stable it might be considered to let them into the kernel. That is not what djdunn said.


The only reason that kdbus was even being asked to be put in the kernel was because of the systemd people.
It's pretty clear from that and other things that Linus said about systemd at around the same time that he thought of them as a cojoined pair,
regardless of what that one quote said.

"you should look for better sources of what you hear" your words, and maybe you should take them to heart also.

Anyway, not to sidetrack the tread anymore....
_________________
PRIME x570-pro, 3700x, 6.1 zen kernel
gcc 13, profile 17.0 (custom bare multilib), openrc, wayland
Back to top
View user's profile Send private message
khayyam
Watchman
Watchman


Joined: 07 Jun 2012
Posts: 6227
Location: Room 101

PostPosted: Tue Jul 08, 2014 9:12 pm    Post subject: Reply with quote

TomWij wrote:
khayyam wrote:
TomWij ... how is that a "turnaround of words"? [..]

Upon a more thorough read that includes your response, it is not that; an ambiguous sentence was misread, which gives a false impression that you were denoting upstream as the community.

Tom ... ok, no problem.

TomWij wrote:
There is a lot of contribution that makes other choices able to exist in the Portage tree, but there could be more contribution to further ensure the future; the equation is simple, if someone stops supporting something then someone else needs to start supporting it for it to remain usable in the future.

Yes, I'm well aware of that equation ... however the the point I was attempting to outline: who owns "the communities" contribution when the show is being run by upstream? I have pointed to this in the past: if upstream rides on all the free labour (labour which had it not been provided would have made linux nonviable), and once having acquired the power to control the course of development, *how* else can the community respond? Work harder, provide more of their free time, etc, etc? The problem as I see it is that linux is no longer about users, its about finding new sources of revenue for the big players. That is an issue that ties in directly with why things are unfolding as they are, its not about development its about gaining market share, and finding new sources of revenue ... and to do this you need to tightly control the "product". Users will be sacrificed on this alter as after all they are dead weight, non-contributors, whiners, "systemd-haters" (and other such politically motivated clap-trap). I'm reminded of this (viewer discretion advised) ... "'cause we don't need you around here any more ..."

TomWij wrote:
You appear to wonder about limitations of choice, but given the way Gentoo works I don't see such limitations. Gentoo does provide choice in a different way than how you describe it, given that limitations (and thus implied choices) are kept to a minimum; as GLEP 39 specifies that "Note that this GLEP does not provide for a way for the community at large to block a new project, even if the comments are wholly negative.", as well as "Global issues will be decided by an elected Gentoo council.". This effectively boils down to that most of the limitations that can exist are those that are voted on by the Gentoo Council. Looking at the history of what has been decided on; there are none that involve enforcing systemd, apart from perhaps the need for an initramfs due to the lack of manpower to support systems without it.

I'm quite aware of how decisions are made, the governance structure, etc, I've read all of the GLEP's including the one cited above, but that doesn't really change my point, someone (more to the point, some group of persons) within that "community" took those actions, and some section of that community are tacitly supporting those actions. Now, fair enough no-one can "enforce" a decision either way, but let us be clear about the repercussions and stop pretending that these "choices" will have no effect on other sections of the community. Let us have no "fading away of the idea of a default", or claims that the choice of systemd as init is commensurate with the choice to run openrc (or some other init system) because these claims are entirely false ... upstream has it as "very strict policy [...] to gently push the distros to standardize on the same components for the base system".

TomWij wrote:
What you label to be framed as coexistence ideas is what Gentoo is all about at its core, as besides GLEP 39 can also be read about in its about and philosophy page; contrary to other distributions, we generally leave the choice to the users instead of picking it instead of them. Given this, accepted a "non-systemd-only" or "systemd-only" Gentoo is something I don't ever see the Gentoo Council do; such choices have a lot of implications, which among other things could result in a community split.

I already stated some months back (link) re this "split" ...

khayyam wrote:
IMO, the best that gentoo could do (and I mean gentoo in the widest possible sense) is to recognise the facts and be absolutely upfront in expressing those facts. It should recognise that by systemd entering the tree gentoo was effectively forked (by the act, not necessarily by design). It needs to recognise this so that it can deal with the repercussions. We can no longer speak about gentoo as one entity (at least, one entity ... made of various bodies ... working toward a common goal) it has two conflicting drives pulling it in two separate directions. If this isn't addressed (or made more visible/actual) then the conflict between those drives will only increase until one or other gains dominance and reduces the other to a footnote in history.

... and I don't think this "split" is something yet to happen, its now history.

As to "coexistence" ... that was the very point I was drawing out in the "polluting the water supply" analogy.

khayyam wrote:
We can agree to tolerate each others choices, but this does not extend to those choices I make that effect you in ways that you haven't consented to. If we say all choices (my choice to pollute the water, and your choice to drink it or not) are equal then we overlook how one party is externalising the negatives, in short, refusing to take into account the repercussions of their actions ... respect, tolerance, etc, are not involved (except perhaps as a smoke screen).

best ... khay
Back to top
View user's profile Send private message
TomWij
Retired Dev
Retired Dev


Joined: 04 Jul 2012
Posts: 1553

PostPosted: Tue Jul 08, 2014 11:36 pm    Post subject: Reply with quote

While upstream takes control of downstream, the opposite is possible; it is bidirectional. Downwards, the motivator is market share; upwards, the motivator is user experience. Assuming open source with appropriate free licenses and no underhand contributor agreements, there is no ownership of contributions; unless you explicitly think in market share, instead of user experience. Granted, it is a harder motivator to work for; because when given the choice between implemented software (low cost and reliability) and to be implemented software (high cost and reliability), people pick the former as long a it works for them and pick the latter when it doesn't.

Your points are clear, but these repercussions alone becoming the final outcome is unacceptable to me; which is why efforts for these "choices" are needed, for a divergent final outcome (and not by pretending or claiming, but rather by doing it). In a broader view the split indeed has always been there due to the bidirectionally acting nature, in a narrow view it could happen when the meta entity realizes that it needs to split; however, given what Gentoo is at its core, it isn't history unless you consider Gentoo to split every time an alternative for a package is added. Before we get to discuss the dot on the letter 'I', please note that whether it is split or not doesn't matter much in the full picture; it is rather about what each split part bidirectionally does with it, as that altogether determines the future of both parts.
Back to top
View user's profile Send private message
Navar
Guru
Guru


Joined: 20 Aug 2012
Posts: 353

PostPosted: Wed Jul 09, 2014 1:04 am    Post subject: Reply with quote

khayyam wrote:
TomWij wrote:
There is a lot of contribution that makes other choices able to exist in the Portage tree, but there could be more contribution to further ensure the future; the equation is simple, if someone stops supporting something then someone else needs to start supporting it for it to remain usable in the future.

Yes, I'm well aware of that equation ... however the the point I was attempting to outline: who owns "the communities" contribution when the show is being run by upstream? I have pointed to this in the past: if upstream rides on all the free labour (labour which had it not been provided would have made linux nonviable), and once having acquired the power to control the course of development, *how* else can the community respond? Work harder, provide more of their free time, etc, etc? The problem as I see it is that linux is no longer about users, its about finding new sources of revenue for the big players. That is an issue that ties in directly with why things are unfolding as they are, its not about development its about gaining market share, and finding new sources of revenue ... and to do this you need to tightly control the "product". Users will be sacrificed on this alter as after all they are dead weight, non-contributors, whiners, "systemd-haters" (and other such politically motivated clap-trap).


I've previously stated similar along with others. I haven't seen a perfectly eloquent and succinct t-shirt/banner variant yet.

Regardless, +1 and something that needs a good amount of thoughtful consideration to attempt refute. I'm all ears Tom if you really want to attempt tackling that. By that I mean a serious response on the issue instead of hitting the same drum beat and ignoring. The people with very substantial funding pushing The Product have not shown themselves to entertain polite discourse, constructive criticism, or even semblance of reasoning to support their rationale. There is no RFC. There was no peer review. There's just take it and like (sic) it.

Quote:
very strict policy [...] to gently push the distros to standardize on the same components for the base system.


Perhaps one of the most damning statements I recall. Who writes something like that without entirely meaning its clearly stated intent? I started a sig for a time just due to that and to avoid any confusion on support threads which side I stood, thereby allowing me to respond as neutral as possible.

adoption_rate = (gentle) ? merit() : PoS_push(severity); /* unfortunately, it's not PoS_push(int i) {;} */

The whole crux of the matter reminds me of the Milgram experiment.

khayyam wrote:
... and I don't think this "split" is something yet to happen, its now history.

As to "coexistence" ... that was the very point I was drawing out in the "polluting the water supply" analogy.

Unfortunately true and yet has happened time and time again from drives to => interrupt the status quo => politics => rapid change => control => faster sustainable revenue.
Back to top
View user's profile Send private message
saellaven
l33t
l33t


Joined: 23 Jul 2006
Posts: 646

PostPosted: Wed Jul 09, 2014 2:01 am    Post subject: Reply with quote

khayyam wrote:

I'm quite aware of how decisions are made, the governance structure, etc, I've read all of the GLEP's including the one cited above, but that doesn't really change my point, someone (more to the point, some group of persons) within that "community" took those actions, and some section of that community are tacitly supporting those actions. Now, fair enough no-one can "enforce" a decision either way, but let us be clear about the repercussions and stop pretending that these "choices" will have no effect on other sections of the community. Let us have no "fading away of the idea of a default", or claims that the choice of systemd as init is commensurate with the choice to run openrc (or some other init system) because these claims are entirely false ... upstream has it as "very strict policy [...] to gently push the distros to standardize on the same components for the base system".


Mind you, the Council is up for election... now's a chance for the devs to vote out the people that started forcing the requirements of systemd's technical inferiorities (ie, initramfs) into Gentoo out. Of course, you need to be a dev to do that. FWIW -

Quote:

<ulm> "The Council agrees that all preparations for dropping support for
separate /usr without an initramfs or similar boot mechanism are
complete. A news item will be prepared, and users will be given one
month to switch after the news item has been sent."
<ulm> one month variant
* WilliamH votes yes
* ulm votes no
* rich0 yes
* scarabeus yes
<dberkholz> yes
* blueness yes
* creffett abstains


All the yeses need to go IMO
Back to top
View user's profile Send private message
mackal
Tux's lil' helper
Tux's lil' helper


Joined: 04 Aug 2011
Posts: 87

PostPosted: Wed Jul 09, 2014 3:04 am    Post subject: Reply with quote

I like systemd, piss and all, I guess.

I'm pretty sure that requirement for /usr was done as the best solution to keep OpenRC and systemd, if you don't like that decision, maybe you should of came up with a better one.
Back to top
View user's profile Send private message
saellaven
l33t
l33t


Joined: 23 Jul 2006
Posts: 646

PostPosted: Wed Jul 09, 2014 3:25 am    Post subject: Reply with quote

mackal wrote:
I like systemd, piss and all, I guess.

I'm pretty sure that requirement for /usr was done as the best solution to keep OpenRC and systemd, if you don't like that decision, maybe you should of came up with a better one.


If systemd is inferior, let the people that want to use it feel the pain of its limitations... don't force them on people that choose not to use systemd.

The patches exist to make OpenRC work fine with a separate /usr but the lead of OpenRC, WilliamH, whom also currently sits on the Council and is running for re-election, knew about those patches, refused to apply the patches, was the one that called for the vote to cripple Gentoo in favor of systemd's limitations and refused to tell his fellow Council members about the patches. He's also on the ComRel team, so good luck appealing somewhere for his transgressions. He, more than anyone, needs to go.

THAT is what got me into the fight against systemd. I don't care what you want to use on your system, more power to you. But don't break mine because of your choices.
Back to top
View user's profile Send private message
Yamakuzure
Advocate
Advocate


Joined: 21 Jun 2006
Posts: 2280
Location: Adendorf, Germany

PostPosted: Wed Jul 09, 2014 10:32 am    Post subject: Reply with quote

saellaven wrote:
The patches exist to make OpenRC work fine with a separate /usr
Nothing works fine with a separate not early mounted /usr. For years. Too many things have gone from /lib and /bin to /usr/lib and /usr/bin over the years. I took the short route: Merged back /usr into /, because there is absolutely *no* advantage in separating usr from rootfs. (Unless you use ZFS and want different compression variants. But then, you need an initramfs anyway...)
_________________
Important German:
  1. "Aha" - German reaction to pretend that you are really interested while giving no f*ck.
  2. "Tja" - German reaction to the apocalypse, nuclear war, an alien invasion or no bread in the house.
Back to top
View user's profile Send private message
depontius
Advocate
Advocate


Joined: 05 May 2004
Posts: 3509

PostPosted: Wed Jul 09, 2014 12:05 pm    Post subject: Reply with quote

I'm of mixed feelings on the whole /usr matter. Personally I don't have separate /usr on any of my systems any more, but can understand the need in some circumstances.

But the real problem goes back to the Linux kernel itself. The real source of this problem is one of how smart the system/kernel is at what stages of the boot process. The original root/usr separation anticipated that the system could be booted, but not yet have enough smarts to mount /usr without "something more". That "something more" was contained in the root filesystem, and could for instance be nfs-utils for mounting an nfs /usr, for instance.

Then Linux added more capabilities, and under some circumstances it became impossible for the kernel to even mount root unaided, for instance root on non-self-assembling RAID. The solution to that was "early userspace", or initrd/initramfs.

So really in a way, "early userspace" is equivalent to root-without-/usr, and if you've got to resort to using it, you've got enough smarts at hand to fully mount the system.

I suspect the real issue here is the same "If it's good enough for me, it's good enough for everybody, no alternatives needed" attitude that causes so much annoyance around systemd.

To be honest, I use genkernel, simply because it lets me walk away from the whole process and indulge in the real world while it crunches. I manage my own kernel configs, so I'm happy. I also get and use the initramfs it generates.

But at least I recognize that this is just the way I work, and others like to work differently. I'm happy to offer my operating method to others curious, but feel no need to force it on anyone as the "One True Way".

However I feel that in the current discussion some history of root-without-usr vs initramfs is informative, though it probably won't change anyone's attitude.
_________________
.sigs waste space and bandwidth
Back to top
View user's profile Send private message
khayyam
Watchman
Watchman


Joined: 07 Jun 2012
Posts: 6227
Location: Room 101

PostPosted: Wed Jul 09, 2014 12:41 pm    Post subject: Reply with quote

TomWij wrote:
While upstream takes control of downstream, the opposite is possible; it is bidirectional.

Tom ... I do not dispute that power is distributed, to a greater or lesser extent, throughout society, but you can not ignore the centralisation of power and how the control of production, information, etc, functions to tip the scales in favour of certain parties. In the case of upstream/downstream its not a level playing field, upstream have more resources at their disposal having found a method of harnessing all the free labour provided by the community and converting it into revenue.

TomWij wrote:
Downwards, the motivator is market share; upwards, the motivator is user experience. Assuming open source with appropriate free licenses and no underhand contributor agreements, there is no ownership of contributions; unless you explicitly think in market share, instead of user experience. Granted, it is a harder motivator to work for; because when given the choice between implemented software (low cost and reliability) and to be implemented software (high cost and reliability), people pick the former as long a it works for them and pick the latter when it doesn't.

As the nature of the problem is political, I would say the upward motivator is liberty and the downward motivator is the inverse. When I posed the question "who owns" I mean this to illustrate what the contribution entails, in law we say this entails "rights and duties", these are not so much transfered as in effect right accross the social landscape. If someone contributes to a community they do not necessarily acquire "ownership" of some part of that community but "rights and duties". For instance, I might bequeath a gift of rare books to a library, in accepting this gift the library does not have a right to burn these books because in accepting them they take on certain duties, part of which limits their right to dispose of the books in whatever manner they see fit, it also imposes a duty to make them available to users of the library. It was these "rights and duties" I was pointing to not the "ownership" per se. In other areas of acquisition we see similar obligations, and the question I was asking, somewhat indirectly, is that when "we contribute" does this not impose/imply similar "rights and duties"?

TomWij wrote:
Your points are clear, but these repercussions alone becoming the final outcome is unacceptable to me; which is why efforts for these "choices" are needed, for a divergent final outcome (and not by pretending or claiming, but rather by doing it). In a broader view the split indeed has always been there due to the bidirectionally acting nature, in a narrow view it could happen when the meta entity realizes that it needs to split; however, given what Gentoo is at its core, it isn't history unless you consider Gentoo to split every time an alternative for a package is added. Before we get to discuss the dot on the letter 'I', please note that whether it is split or not doesn't matter much in the full picture; it is rather about what each split part bidirectionally does with it, as that altogether determines the future of both parts.

This seems to ignore the fact that these "repercussions" are not "bidirectional", and it was this fact I was drawing out in the "polluting the water supply" analogy. The comparison between systemd (with its "policy") and other "packages" being added is not equatable (for reasons I've already outlined). Now, you can argue that what matters is what some part "does with it" but this seems to ignore the question of relations (of power, of coexistence, etc) so while true, it side steps the fact that, as I said, if the "community" (you pointed to) behaves in this manner then "rights and duties" are the issue, not how the effected party responds.

best ... khay
Back to top
View user's profile Send private message
khayyam
Watchman
Watchman


Joined: 07 Jun 2012
Posts: 6227
Location: Room 101

PostPosted: Wed Jul 09, 2014 1:09 pm    Post subject: Reply with quote

mackal wrote:
I like systemd, piss and all, I guess.

mackel ... if you are going to contrbute to this discussion then please make an attempt to understand the argument your interlocutors are presenting. The "piss" in the above is not "systemd" but the effects on others created by systemd (the "poluting the water supply" analogy).

mackal wrote:
I'm pretty sure that requirement for /usr was done as the best solution to keep OpenRC and systemd, if you don't like that decision, maybe you should of came up with a better one.

This "if you don't like it" argument is obnoxious, you overlook the fact that many of us have contributed to this OS *you* use, supported it with hours of free labour, and all we have asked for in return is that this is accounted for, and that we are not treated as "haters" because we make criticisms of how upstream are imposing *their* decisions on *us*.

best ... khay
Back to top
View user's profile Send private message
Anon-E-moose
Watchman
Watchman


Joined: 23 May 2008
Posts: 6095
Location: Dallas area

PostPosted: Wed Jul 09, 2014 1:12 pm    Post subject: Reply with quote

depontius wrote:
I suspect the real issue here is the same "If it's good enough for me, it's good enough for everybody, no alternatives needed" attitude that causes so much annoyance around systemd.


Indeed

Quote:
But at least I recognize that this is just the way I work, and others like to work differently. I'm happy to offer my operating method to others curious, but feel no need to force it on anyone as the "One True Way".


The whole one true way is the annoying t*rd in the punchbowl.

I don't expect others to have their systems set up the same way as me, and I'm also happy to help if they ask.
But like you I don't think that anyone's one true way should be forced on others.

It's arrogant and very annoying.
And before the fanbois chime in about no one is being forced,
yes the subtle forcing is there whether some want to acknowledge it or not.

Edit to add: I suppose the thing that most disappointing, to me, is all the lies told by various parties,
distro council, devs, upstream. And yes, omissions of truth count as lies. And when it becomes clear
that there were lies being told out comes the finger pointing, "oh it's not my fault, it's theirs"

What I do know is this, if gentoo allows itself to be led to a systemd only omniverse then I'll become
one of those distro-less people and go back to rolling my own. Yes it's more work, but at least I can
decide what patches to apply whether some package is really needed to support another, etc.
I may be in a minority, but that's ok, as I don't care about things like being a majority supporter.
_________________
PRIME x570-pro, 3700x, 6.1 zen kernel
gcc 13, profile 17.0 (custom bare multilib), openrc, wayland
Back to top
View user's profile Send private message
khayyam
Watchman
Watchman


Joined: 07 Jun 2012
Posts: 6227
Location: Room 101

PostPosted: Wed Jul 09, 2014 1:55 pm    Post subject: Reply with quote

saellaven wrote:
All the yeses need to go IMO

saellaven ... I simply can't agree with you on that. I can agree that this is a crisis of governance, that to some degree users (or the more general "community") have been let down, and that the whole issue has not been clearly articulated and thought through, etc, but I don't see this as by design but by effect. As I wrote in the post linked above "by systemd entering the tree gentoo was effectively forked (by the act, not necessarily by design)" and I'm inclined to give the benefit of the doubt. Yes, there is perhaps some level of spin, overlooking, etc, but I'm far less inclined to see this as wilful, people are fallible and often don't carefully consider outcomes, if we did then we would never err.

best ... khay
Back to top
View user's profile Send private message
krinn
Watchman
Watchman


Joined: 02 May 2003
Posts: 7470

PostPosted: Wed Jul 09, 2014 4:31 pm    Post subject: Reply with quote

khayyam wrote:
saellaven wrote:
All the yeses need to go IMO

saellaven ... I simply can't agree with you on that.


And i agree with him.
All those are council members, if anyone abuse the Council providing false arguments to backup his wanted change. It is every members responsibility to make sure the vote initiator doesn't abuse them.
You cannot be a council member and vote to follow other members. If you don't know what it is about, it's your responsibility as council member to ask clarification and delay to get the whole picture.
Clearly only ulm has vote no, only ulm seems to know what he was speaking about. Not even williamh that introduce the vote knows what he was speaking of, or worst, he knows and was abusing other. ulm tell him it's not FHS compliant to do so, williamh answer with a link to FHS doc, ulm answer he knows the FHS doc well and that why it is not FHS compliant ; next answer from williamh was the crappy well know systemd propaganda link.

http://www.gentoo.org/proj/en/council/meeting-logs/20130813.txt
Quote:
<WilliamH> ulm: /usr merge is fhs compliant though.
<ulm> WilliamH: it isn't
<ulm> no way
<WilliamH> ulm: they aren't mutually exclusive.
<WilliamH> ulm: wait a sec [21:27]
<ulm> "The contents of the root filesystem must be adequate to boot, restore,
recover, and/or repair the system."
* WilliamH is getting the link
...
<WilliamH> the /usr merg has symlinks in / for bin, sbin, lib* that go to
their counerparts in /usr
...
<WilliamH> ulm hang on I'm getting the fhs link for you to show what I was
talking about.
<WilliamH> ulm: http://www.linuxbase.org/betaspecs/fhs/fhs.html [21:30]
...
<ulm> WilliamH: yes, and that says what I've quoted earlier
<WilliamH> blueness:
http://www.freedesktop.org/wiki/Software/systemd/TheCaseForTheUsrMerge/


Council members are taking decisions like /usr merge that could be critical to Gentoo. If anyone is a member and just vote randomly, he have nothing to do in the Council. I expect Council members to protect Gentoo : and from whom Gentoo should be protect if it's not Gentoo devs? So Council is there to protect Gentoo from its own devs.
They have clearly fail there.

And i don't fucking know how it can be change, but i'm so upset to see again the next Council vote is not going to change that.
Council members can only be Gentoo devs. And i suppose only some GLEP might change the rule ; and guess who vote on GLEP...
No dev leading a team should be allow to be Council member.
If police is there protect citizen from bandits, it's so crazy to see policemen can be a bandit chief and nobody see a problem in that.
Any Council members SHOULD drop any team leading when elect. Power separation is common in all democracies, for a reason...
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 54096
Location: 56N 3W

PostPosted: Wed Jul 09, 2014 4:49 pm    Post subject: Reply with quote

krinn,

The /usr merge is being driven by assorted upstreams.
Gentoo devs can only stand in its way while they are prepared to carry patch sets to make things work the old way. That's a lot of work and growing so I'm not surprised gentoo is moving slowly down the /usr merge route. Its the path of least resistance.

It just means that the initramfs is the new root. However, the initamfs need not be in a file, or even in RAM.
Consider an install that was partitioned as

Code:
sda1 boot
sda2 initrd
sda3 root
sda4 usr
sda5 var ...

boot loads the kernel without an initrd and the kernel mounts /dew/sda2 as root.
The int script there mounts root, usr and whatever you like, then does a switchroot.

Ugly isn't it but it will work.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
TomWij
Retired Dev
Retired Dev


Joined: 04 Jul 2012
Posts: 1553

PostPosted: Wed Jul 09, 2014 6:28 pm    Post subject: Reply with quote

khayyam wrote:
Tom ... I do not dispute that power is distributed, to a greater or lesser extent, throughout society, but you can not ignore the centralisation of power and how the control of production, information, etc, functions to tip the scales in favour of certain parties. In the case of upstream/downstream its not a level playing field, upstream have more resources at their disposal having found a method of harnessing all the free labour provided by the community and converting it into revenue.


That only goes as far as customers are interested, as they can choose their upstream; they can try elsewhere, or in this case with the lack of alternatives try it by themselves. As you and others have said, that's easier said than done; but at the very least wake up calls are made like this very thread, which acts like a nagging reminder for those that could be interested.

khayyam wrote:
As the nature of the problem is political, I would say the upward motivator is liberty and the downward motivator is the inverse. When I posed the question "who owns" I mean this to illustrate what the contribution entails, in law we say this entails "rights and duties", these are not so much transfered as in effect right accross the social landscape. If someone contributes to a community they do not necessarily acquire "ownership" of some part of that community but "rights and duties". For instance, I might bequeath a gift of rare books to a library, in accepting this gift the library does not have a right to burn these books because in accepting them they take on certain duties, part of which limits their right to dispose of the books in whatever manner they see fit, it also imposes a duty to make them available to users of the library. It was these "rights and duties" I was pointing to not the "ownership" per se. In other areas of acquisition we see similar obligations, and the question I was asking, somewhat indirectly, is that when "we contribute" does this not impose/imply similar "rights and duties"?


"Contribute" is a too big word, as you're not necessarily "giving away" something; code you give you still have yourself to use freely (unless there is a CLA in place), as opposed to giving a book to the library. Gentoo itself is as meta as it can get and barely has any limitations, but nobody says this contribution necessarily needs to happen to the Portage tree.

Thus ... Outside the Portage tree, there are no such "right and duties"; an overlay would work out well. Inside the Portage tree, the before mentioned GLEP allows you to add new packages and projects with ease; the "rights and duties" would be the same as for any other Portage tree activity. You cannot touch another maintainer or project's packages, without prior discussion and an agreement; but I don't think that serves any problem, given the efforts would be appreciated. If unsure, a discussion on the mailing lists can make people's viewpoints more clear.

khayyam wrote:
This seems to ignore the fact that these "repercussions" are not "bidirectional", and it was this fact I was drawing out in the "polluting the water supply" analogy. The comparison between systemd (with its "policy") and other "packages" being added is not equatable (for reasons I've already outlined). Now, you can argue that what matters is what some part "does with it" but this seems to ignore the question of relations (of power, of coexistence, etc) so while true, it side steps the fact that, as I said, if the "community" (you pointed to) behaves in this manner then "rights and duties" are the issue, not how the effected party responds.


They are bidirectional; if you draw a ton of people away from upstream's product and/or service, because of an alternative or own product and/or service; that will affect upstream's revenue. The dependencies (by GNOME, now UPower, etc...) however make this relation strongly favor upstream's side; so, I do agree on the pollution but a cure for it is still possible.
Back to top
View user's profile Send private message
krinn
Watchman
Watchman


Joined: 02 May 2003
Posts: 7470

PostPosted: Wed Jul 09, 2014 7:19 pm    Post subject: Reply with quote

NeddySeagoon wrote:
krinn,

The /usr merge is being driven by assorted upstreams.
Gentoo devs can only stand in its way while they are prepared to carry patch sets to make things work the old way. That's a lot of work and growing so I'm not surprised gentoo is moving slowly down the /usr merge route. Its the path of least resistance.

No, "assorted upstreams" is just redhat.

That usr merge is pushed by redhat, as they broke fhs compliance to symlink /sbin with /usr/sbin for no sane reason.
Everything start because of that. They try to get as close as Oracle and unix. How ironic to break fhs compliance in order to have your distro be more compatible with unix (for ones that don't get it, Linux is GNU/Linux made with GNU tools, and GNU mean GNU's not Unix!)...
Of course you can guess why, there's valid reason to be as close as your competitor in the market to play against him. Making more money. That's valid reason, not a sane one, as they broke fhs for their only benefit.

Now that /usr/sbin is no more than just a symlink to /sbin, anyone making a program that is critical to boot your system inside /usr/sbin will still works as expect in redhat/fedora, but in all other fhs compliant distros, the program is broken as /usr might not even be mount when booting. And the result : critical to boot program now is broken, system cannot boot.

When dealing with any kay and lennart and friends programs, Gentoo Council must keep in mind, they work and they are redhat devs, any decisions made by them in any of their program should be question with that kept in mind. Redhat start symlink /sbin... and surprise! systemd now declare /usr is broken if not mount early...

I think it's justify to have Council rules by devs only, as most decisions are technical ones, but when some political decisions drive technical ones, we clearly see the weakness of Council, all the members that vote yes shown how stupid or candid they were.
I'm not a dev, so i cannot be Gentoo dev, and so not a Council member. But clearly i think i would really do a better job at protecting Gentoo.
They vote yes and broke many systems to comply with upstream decision, a decision upstream made to comply with their own employer politic. I hope some have vote yes but redhat is giving them money, at least, they wouldn't be so stupid.

Thanks for the initrd partition trick, i'm not finding it ugly but nice :)
Back to top
View user's profile Send private message
mackal
Tux's lil' helper
Tux's lil' helper


Joined: 04 Aug 2011
Posts: 87

PostPosted: Wed Jul 09, 2014 8:12 pm    Post subject: Reply with quote

krinn wrote:
for ones that don't get it, Linux is GNU/Linux made with GNU tools, and GNU mean GNU's not Unix!


Okay cool, so we can throw away the do one thing well idea.

krinn wrote:
but when some political decisions drive technical ones


It was actually a technical decision. http://0pointer.de/blog/projects/stateless.html

I'm all for standardization, and systemd is using it's presence to push through some. The most sensible would probably standardization of the hostname file to /etc/hostname (SuSE used to be /etc/HOSTNAME, RH/Fedora was /etc/sysconfig/network, so it's not like they are just going with what RH was doing)
Back to top
View user's profile Send private message
khayyam
Watchman
Watchman


Joined: 07 Jun 2012
Posts: 6227
Location: Room 101

PostPosted: Wed Jul 09, 2014 8:36 pm    Post subject: Reply with quote

TomWij wrote:
khayyam wrote:
Tom ... I do not dispute that power is distributed, to a greater or lesser extent, throughout society, but you can not ignore the centralisation of power and how the control of production, information, etc, functions to tip the scales in favour of certain parties. In the case of upstream/downstream its not a level playing field, upstream have more resources at their disposal having found a method of harnessing all the free labour provided by the community and converting it into revenue.

That only goes as far as customers are interested, as they can choose their upstream; they can try elsewhere, or in this case with the lack of alternatives try it by themselves. As you and others have said, that's easier said than done; but at the very least wake up calls are made like this very thread, which acts like a nagging reminder for those that could be interested.

Tom ... no, its a political fact, that is why we have laws that attempt to prevent abuse of power, absolutism, monopoly, etc, and a legal system where we enshrine rights and duties. We do this because we understand that it isn't simply a matter of "choosing" one product over another but that principles underscore jus.

TomWij wrote:
khayyam wrote:
As the nature of the problem is political, I would say the upward motivator is liberty and the downward motivator is the inverse. When I posed the question "who owns" I mean this to illustrate what the contribution entails, in law we say this entails "rights and duties", these are not so much transferred as in effect right across the social landscape. If someone contributes to a community they do not necessarily acquire "ownership" of some part of that community but "rights and duties". For instance, I might bequeath a gift of rare books to a library, in accepting this gift the library does not have a right to burn these books because in accepting them they take on certain duties, part of which limits their right to dispose of the books in whatever manner they see fit, it also imposes a duty to make them available to users of the library. It was these "rights and duties" I was pointing to not the "ownership" per se. In other areas of acquisition we see similar obligations, and the question I was asking, somewhat indirectly, is that when "we contribute" does this not impose/imply similar "rights and duties"?

"Contribute" is a too big word, as you're not necessarily "giving away" something; code you give you still have yourself to use freely (unless there is a CLA in place), as opposed to giving a book to the library. Gentoo itself is as meta as it can get and barely has any limitations, but nobody says this contribution necessarily needs to happen to the Portage tree.

No, because time, effort, etc, are also "things" and so have some material value in the circulation of goods. It doesn't matter if the contribution is made up of words, carrots, or bytes, they all require physical inputs and so are a resources that involve expenditure in one form or another. The gifting of books was there to illustrate "rights and duties" not the transfer of goods (which you are still focused on ... strange as this was what I was countering with that particular paragraph/example). It absolutely doesn't matter what form the contribution takes "rights and duties" are still involved (if, that is, our principles are to have any meaning).

TomWij wrote:
Thus ... Outside the Portage tree, there are no such "right and duties"; an overlay would work out well. Inside the Portage tree, the before mentioned GLEP allows you to add new packages and projects with ease; the "rights and duties" would be the same as for any other Portage tree activity. You cannot touch another maintainer or project's packages, without prior discussion and an agreement; but I don't think that serves any problem, given the efforts would be appreciated. If unsure, a discussion on the mailing lists can make people's viewpoints more clear.

These "rights and duties" are operative wherever there are social arrangements, even in animals. They needn't be codified as they often function without the need to be so. You would find it a violation of the most basic if these if I lied about you, or caused you intentional harm, you may even consider it your duty to retaliate, or prevent me from doing the same to others, should this occur. It has nothing to do with ebuilds, overlays, or what-have-you, you're completely speaking past the point I was making.

TomWij wrote:
khayyam wrote:
This seems to ignore the fact that these "repercussions" are not "bidirectional", and it was this fact I was drawing out in the "polluting the water supply" analogy. The comparison between systemd (with its "policy") and other "packages" being added is not equatable (for reasons I've already outlined). Now, you can argue that what matters is what some part "does with it" but this seems to ignore the question of relations (of power, of coexistence, etc) so while true, it side steps the fact that, as I said, if the "community" (you pointed to) behaves in this manner then "rights and duties" are the issue, not how the effected party responds.

They are bidirectional; if you draw a ton of people away from upstream's product and/or service, because of an alternative or own product and/or service; that will affect upstream's revenue. The dependencies (by GNOME, now UPower, etc...) however make this relation strongly favor upstream's side; so, I do agree on the pollution but a cure for it is still possible.

Again you're ignoring the central point of what the "polluting the water supply" analogy is all about. When I pollute the water upstream from you I do not bare the negative effects of my action, you do. So, its not "bidirectional". Remember where I started with this ... it was about "community", and what the "rights and duties" of a community involve, polluting *your* water supply, lying about you, or doing you intentional harm, violate principles of jus.

best ... khay
Back to top
View user's profile Send private message
krinn
Watchman
Watchman


Joined: 02 May 2003
Posts: 7470

PostPosted: Thu Jul 10, 2014 12:37 am    Post subject: Reply with quote

mackal wrote:

It was actually a technical decision. http://0pointer.de/blog/projects/stateless.html

Where is there any technical decision for Gentoo there? It's a systemd blog, so if any technical decision is present in that link, it's technical decision for systemd. And despite all claims by systemd fans : Gentoo and Linux are not systemd.

mackal wrote:

I'm all for standardization, and systemd is using it's presence to push through some. The most sensible would probably standardization of the hostname file to /etc/hostname (SuSE used to be /etc/HOSTNAME, RH/Fedora was /etc/sysconfig/network, so it's not like they are just going with what RH was doing)

It doesn't matter where you store the hostname of your host, the standard is to call hostname to set your hostname or call hostname to read it.

So systemd doesn't define any standard to set your hostname, "hostname" is the program that do that already, at best systemd define a standard file to store your hostname configuration.
That's again a totally stupid wanking example of systemd change.

If you standardize where to store your hostname configuration, you might think it's good so you can use that file to actually get your hostname...
In real world, you are making jewel for your fans out of NOTHING!
Let's see what systemd has solved in real.
Gentoo store it in /etc/conf.d/hostname
Code:
> hostname
myhostname
> sysctl kernel.hostname
kernel.hostname = myhostname
> cat /etc/conf.d/hostname
hostname="myhostname"
> echo 'hostname="systemd_is_so_awseome"' > /etc/conf.d/hostname
> hostname
myhostname
# here you should ask yourself : WTF it doesn't change my hostname as i wish!
> hostname systemd_is_again_solving_no_problem
> hostname
systemd_is_again_solving_no_problem
> sysctl kernel.hostname
kernel.hostname = systemd_is_again_solving_no_problem
# Now we will get into the awesome level only systemd can reach...
hostname $(cat whatever_file_config_name_and_location_you_like)

Now because of systemd so awesome standardization of hostname config file location and name, everyone might think, yeah ! thanks to systemd i have now a way to get my hostname in any distros !
You just need to type "hostname" in SuSe, fedora, Gentoo, whatever sane distro that provide hostname utility to actually really get your hostname easy.
And even without hostname utility, you can ask kernel to get it. Whatever solution you use, they will gives you your hostname, while reading any configuration file, will just return what the user as set as default hostname ; something that is not reliable as it might have change.
systemd solved another non existing problem... It actually add one, as many users might think /etc/hostname is reliable and if present actually reading it will return the hostname ; just like you prove yourself...
Back to top
View user's profile Send private message
Display posts from previous:   
This topic is locked: you cannot edit posts or make replies.    Gentoo Forums Forum Index Other Things Gentoo All times are GMT
Goto page Previous  1, 2, 3  Next
Page 2 of 3

 
Jump to:  
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