| View previous topic :: View next topic |
| Do you want systemd as default on Gentoo? |
| I <3 systemd!! I want Gentoo to switch!! |
|
12% |
[ 26 ] |
| Get that horse-crap away from Gentoo as far as possible! |
|
87% |
[ 186 ] |
|
| Total Votes : 212 |
|
| Author |
Message |
steveL
Watchman
Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery
|
 Posted: Fri Sep 19, 2014 9:29 pm Post subject: |
 |
|
| steveL wrote: | You keep coming up with how we "need" to keep up with systemd (in the land where users "just want things to work"), and sorry but we don't. Far more effective is to rewrite the DEs to use a library as opposed to a daemon (which is totally unnecessary) so that code can be shared across projects without inviting the spaghetti-monster in to the parlour.
|
| depontius wrote: | | Because for the moment, systemd is the strategic direction, and L.P. is the heir-apparent. |
I don't know why you keep repeating the same mantra, like I'm suddenly going to agree with it. As I said before, not imo: if anyone is tying their project to logind, it's simple enough to replace what they're after, depending on what exactly that is, but and this is the critical point, the vast majority of uses within the desktop arena are simply plain wrong.
It's an indicator of a broken design, but thankfully it doesn't infect that much, since the supposedly "vital" uses of it are all about DE-specific "control-centers" and the like, ie what should be considered privileged operations, in one sense or another. Requiring a daemon for this is frankly pathetic, since we have a DM and DE/WM pair by definition.
| Quote: | | Right now systemd is in its "honeymoon phase", so that's where developers are taking their upstreams, and we're not likely to have much luck talking them out of it. |
I have no interest in talking anyone out of anything; that's a recipe for hours of debate and little progress.
| Quote: | | Once the systemd "disillusionment phase" we'll have much more luck with that, but it will be important to be working, , and that's where the shims come in. |
I'm not sure who you're quoting there, but I'll read that as emphasis instead of a quote, since otherwise it makes no sense.
Again, shims are pointless, if what they're shimming is something you simply shouldn't be doing in the first place.
| Quote: | | After that, hopefully people will accommodate systemd, but not hitch their wagons to it. |
Again, you seem to be presenting a strategy, but it's not one I recognise as effective at all, since all it does is cede the design ground, with some sort of acceptance that we must shim "essential functionality," when it's horse-manure.
And I don't want to go round in loops again, to repeat the same point to you yet again, so please don't make the same point to me with slightly different words again. To be blunt, your strategy sucks afaic, and I for one have no interest in playing "catch-up" with a crappy concept that is badly-implemented.
| C. Ray Johnson wrote: | | There is nothing so wasteful as doing with great efficiency that which does not have to be done at all. |
Seems to me all you're talking about is adding loads of unnecessary work on everyone else's shoulders, to validate RedHat's "direction"; iow exactly what they want us to do.
Others might want to waste their time on it, but I'm curious as to how you couldn't see the same point being put to you before. And if so, why you don't answer it directly and substantively, instead of repeating what you said before, again with vague references to nothing much. |
|
| Back to top |
|
 |
steveL
Watchman
Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery
|
| Posted: Fri Sep 19, 2014 9:51 pm Post subject: |
|
|
| NeddySeagoon wrote: | | saellaven wrote: | | Start with William Hubbs (williamh)... he's the one that made it his mission to use his newly gained Council seat to mandate initramfs because systemd was going to drop support for a separate /usr. Despite being the openrc lead, he willfully ignored patches to openrc that had already existed for a year |
I don't belive thats quite correct. I assume that you refer te the patches submitted by steveL?
If so, reread the bug. There were other base system devs involved in the decision. |
Hmm what bug is that? For my side, I'd say saellaven is right, wrt WilliamH. And I was there with those patches, obviously, but also hung out in #openrc trying to contribute, and helped with quite a lot of shell, over a period of 9-12 months or so.
The real issue there is that WilliamH is a terrible lead, likely because he's out of his depth, as I found out when I finally reviewed the git commit history. I lost count of the number of people I saw get the frozen shoulder and leave in a silent huff, never to return, after waiting ages for WilliamH to decide on "use-cases" and then simply blank everyone.
Very different to #-portage, where zmedico regularly turned visitors into contributors.
IME WilliamH is much more interested in "leveraging" his position as openrc maintainer to curry favour with other projects, typically outside of Gentoo, than he is in either coding or in improving openrc. Frankly I found the git commit history nauseating, and had to chase it back to the beginning, to assure myself that the context I was seeing was not Roy's code.
| Quote: | | saellaven wrote: | | ... members on the Council are also on QA, ComRel, PR and Foundation.. |
Thats not correct. |
Perhaps not for trustees, but if there's an issue with someone like WilliamH, then you're on a hiding to nothing, since he's a Council member, so appealing a comrel decision about him is basically a waste of time, even if comrel aren't themselves part of the problem.
OFC, the actual issue is that Council, which is supposed to be strictly about technical matters, has no mandate whatsoever to oversee anything else, and most especially not social matters, which they've strictly abrogated from, as the basis for their existence.
That's a long way of saying: they stand purely on their technical ability, so it's completely inappropriate to expect them to handle non-technical matters with any sort of effectiveness, however much anyone might wish to pride themselves on their emotional nous.
Comrel should be led by proctors as originally mandated by the community in the process that mandated the Code of Conduct: that process mandated proctors as the enforcement mechanism, so separating the two is simply unacceptable in any sort of democratic due process. Without proctors, the CoC doesn't have any mandate over the whole Community: it's pure fiction to argue otherwise.
All CoC matters must come under the purview of Trustees, perhaps delegating to another body (eg a reconstituted comrel with actual proctors in line with the Community decision), but never to Council, which as stated is strictly for technical matters: that's the basis for people standing on it in the first place.
So, when are Trustees going to stand up and take responsibility for what they're supposed to, instead of simply tsking in the background when "developers" act like vipers? |
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 54237
Location: 56N 3W
|
| Posted: Sat Sep 20, 2014 10:55 am Post subject: |
|
|
steveL,
I fail at searching bugsie. I cant find the bug with your patches attached :(
I would really like to respond to | steveL wrote: | | So, when are Trustees going to stand up and take responsibility for what they're supposed to, instead of simply tsking in the background when "developers" act like vipers |
but as a trustee, I can't. At best, I can point you to my 2014 (re)election manifesto.
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
saellaven
l33t
Joined: 23 Jul 2006
Posts: 646
|
| Posted: Sat Sep 20, 2014 1:40 pm Post subject: |
|
|
| NeddySeagoon wrote: | steveL,
I fail at searching bugsie. I cant find the bug with your patches attached 
I would really like to respond to | steveL wrote: | | So, when are Trustees going to stand up and take responsibility for what they're supposed to, instead of simply tsking in the background when "developers" act like vipers |
but as a trustee, I can't. At best, I can point you to my 2014 (re)election manifesto. |
To save others the need to look,
| NeddySeagoon wrote: |
I'm in for a fourth team, if anyone wants me.
During my fourth term, if re-elected, I would like to work on
reorganising the Gentoo meta structure more in line with that of a
normal corporation. Mostly because what we have today works only as
long as we are all good friends.
It will break horribly if that changes and there have been a few near
hits since early 2008. When it breaks, not if, the Foundation will be
held both accountable and responsible for the fallout. The Foundation
cannot point to other groups within Gentoo and say it was nothing to to
with them.
At the very least, any group that makes decisions on behalf of Gentoo
needs to have its members as officers of the foundation. Moreso if the
decisions are sometimes controversial.
The groups I have in mind are Council, Pr and ComRel. There may be
others.
This is my own re-election manifesto. Its not been discussed with the
present board. If you think its a bad idea, don't vote for me.
There will need to be changes to the Foundation bylaws and probably a
GLEP or two but its all for the long term good of Gentoo. |
But I'd really like to know what we can do with regards to the Council NOT doing their job in an appropriate fashion, particularly when the same members exist as members of PR and ComRel too. And wouldn't making them Foundation officers spread their influence so that there is no truly independent oversight body?
For two years, I've been dying to see the Council's supposed technical discussion regarding the changes to /usr, but it doesn't seem to exist anywhere outside of the forced vote. The Council claims their decisions cannot be appealed (ergo, they are completely unanswerable) and the members sit on all of the key projects which one might be able to appeal to, meaning they are unlikely to hear an appeal in the first place and even if they do, their loyalties to each other will be questioned. It's part of the very core of what kept me away from becoming a dev for the 8 years I've been using Gentoo after rolling my own systems from scratch for the 8 years prior to that, the holier than thou "devs are better than users and users will have to eat our dog food and like it" mentality. |
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 54237
Location: 56N 3W
|
| Posted: Sat Sep 20, 2014 2:31 pm Post subject: |
|
|
saellaven,
All my manifesto really says is lets restructure Gentoo formally in the way the laws of New Mexico and the USA hold us to anyway.
The buck stops with the Foundation. Thats the law of the land.
If someone gets pissed off enough to invoke the law, the present setup will break, as it depends on goodwill. Goodwill is the first thing out the window when lawyers letters arrive.
The Foundation does have a postal address - thats a legal requirement too.
Whether the other groups like comrel, council and so on like it or not they may not think they are accountable to the Foundation and the Foundation cannot 'interfere' but the reality is otherwise.
Under the present setup, the independence of the various groups continues to be fostered as there is no formal accountability nor responsibility to the laws of New Mexico and the USA via the Foundation.
As I've said, that makes me nervous. It should make the individual post holders nervous too as the Foundation may not be a buffer between them and the law. (I am less sure of that).
I did write a potted history of how Gentoo got to where it is today. Its incomplete and could do with some editing.
I need to get back to it.
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
saellaven
l33t
Joined: 23 Jul 2006
Posts: 646
|
| Posted: Sat Sep 20, 2014 2:47 pm Post subject: |
|
|
| NeddySeagoon wrote: |
Whether the other groups like comrel, council and so on like it or not they may not think they are accountable to the Foundation and the Foundation cannot 'interfere' but the reality is otherwise.
Under the present setup, the independence of the various groups continues to be fostered as there is no formal accountability nor responsibility to the laws of New Mexico and the USA via the Foundation.
As I've said, that makes me nervous. It should make the individual post holders nervous too as the Foundation may not be a buffer between them and the law. (I am less sure of that).
|
So, how do I go about asking the Foundation to review the process by which the Council made the /usr decision and whether or not that process was valid and thus, should be overturned? Moreover, should the facts bear me out, I'd like to see WilliamH removed and barred from any top level projects for what I believe is malfeasance on his part.
I'd be glad to write up a formal request, complete with documentation and evidence of their lack of discussion and apparent lack of desire to actually do the job to which they were appointed. I wanted to do it a year+ ago but the situation with the co-mingling of devs among the top level projects implied there was nowhere to seek recourse for their actions. |
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 54237
Location: 56N 3W
|
| Posted: Sat Sep 20, 2014 4:00 pm Post subject: |
|
|
saellaven,
Under the present structure, you would need to find something illegal to draw to the Foundations attention, as given the history of Gentoo and the current structure, that is the only basis on which the trustees are likely to intervine.
You also need to understand that in Gentoo, the council is not and was never intended to be a leadership body. It was always intended to be a technical disputes resolution body.
As it was the only visible active body at the time of its creation, it picked up everything else too. When drobbins left, the technical leadership was not replaced.
Gentoo devs pretty much do as they please until they upset other projects, then the council gets involved. With respect to the /usr issue, its driven by $UPSTREAM moving things that should be on / to /usr.
Thats the same $UPSTREAM that gave us systemd. There is a limit to how much effort devs can/should put into resisting the /usr move in Gentoo. The council vote left it up to the individual devs.
Now, suppose the vote had gone the other way, Gentoo is an all volunteer distro, the council would have no way to make the vote stick - its not them that maintains the increasing large patch set to reverse the /usr move.
I will deliberately not respond to your opinions of WilliamH.
If you wish to take things further, comrel is the way to go today.
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
saellaven
l33t
Joined: 23 Jul 2006
Posts: 646
|
| Posted: Sat Sep 20, 2014 4:32 pm Post subject: |
|
|
| NeddySeagoon wrote: | saellaven,
Under the present structure, you would need to find something illegal to draw to the Foundations attention, as given the history of Gentoo and the current structure, that is the only basis on which the trustees are likely to intervine.
You also ned to understand that in Gemtoo, the council is not and was never intended to be a leadership body. It was always intended to be a technical disputes resolution body.
As it was the only visible active body at the time of its creation, it picked up everything else too. When drobbins left, the technical leadership was not replaced.
I will deliberately not respond to your opinions of WilliamH.
If you wish to take things further, comrel is the way to go today. |
aka, it's a waste of time and nothing will be done. The devs are free to use and abuse their power without regard or concern for anyone else.
| Quote: |
Gentoo devs pretty much do as they please untl they upset other projects, then the council gets involved. With respect te the /usr issue, its driven by $UPSTREAM moving things that should be on / to /usr.
Thats the same $UPSTREAM that gave us systemd. There is a limit to how much effort devs can/should put into resisting the /usr move in Gentoo. The council vote left it up to the individual devs.
Now, suppose the vote had gone the other way, Gentoo is an all volunteer distro, the council would have no way to make the vote stick - its not them that maintains the increasing large patch set to reverse the /usr move. |
That is factually incorrect.
Separate /usr can and is mounted during early boot before anything else is. SteveL's patches are marked proof of how simple and easy the solution is.
The /usr issue was driven by systemd initially planning to drop support for a separate /usr, WilliamH's questionable dedication to OpenRC as project lead with is dual loyalty to systemd, as well as the power offered WilliamH's then new position as a Council member. Regarding the Council, there was no public technical discussion of the merits at all, nor any discussion of the potential downside of their solution (initramfs has broken many systems since and will continue to cause new, entirely random breakages in the future). Ergo, the Council's decision was not one of a technical nature at all, given no technical merits were discussed, but a purely political decision driven by WilliamH whom not only refused the proper solution in the first place, but engineered the entire scenario. It's also worth noting that the Council went even further, stating that it is ok for Gentoo devs to intentionally break a separate /usr even if their upstream supports it. Again, that reeks of politics, not technical consideration.
The entire point was to hold back OpenRC to the arbitrary LP ego-imposed limitations of systemd, so that proponents could continue to propose that systemd was actually more robust than OpenRC, paving the way for future forced Gentoo adoption of systemd due to the entirely PR driven "limitations of OpenRC." Again, I would like to reiterate my request to see whatever discussion was had by devs over SteveL's patches and why they were rejected. Given that they work and don't affect anyone not using a separate /usr, there was no technical reason not to include them.
And given that /usr is mounted very early on, only a small handful of packages are potentially affected by an early mounted separate /usr, all of which are easy to manage - essentially, it comes down to lvm2, coreutils and util-linux being installed in / where they are supposed to be anyway. Everything else gets loaded after /usr is mounted. If it was more complicated than that, we'd have 200 MB initramfs files instead of 2 MB ones. |
|
| Back to top |
|
|
steveL
Watchman
Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery
|
| Posted: Sat Sep 20, 2014 4:33 pm Post subject: |
|
|
@saellaven: I don't think we should bother with that discussion personally; the decision happened, and we've dealt with the consequences for our use-cases. Focussing on that, can only take us backwards, though at least we've proved that the "separate /usr is broken" myth is exactly that: a complete load of myth. ;) It's exactly the same situation as 15 years ago: if you need modules to get access to your rootfs, you need an initramfs (pka: initrd.)
To my mind, it's clear that Council has no possible mandate over anything like ComRel; what used to be devrel and userrel, and run with much better ability by fmmcor (sadly missed) and jmbsvicetto.
As time, experience, and they've shown, technical leads are usually terrible at handling social matters; that's why Council is, and always has been, strictly about technical matters. The developers always wanted it like that.
So this is not contentious, unless someone is politicking. |
|
| Back to top |
|
|
saellaven
l33t
Joined: 23 Jul 2006
Posts: 646
|
| Posted: Sat Sep 20, 2014 4:41 pm Post subject: |
|
|
| steveL wrote: | @saellaven: I don't think we should bother with that discussion personally; the decision happened, and we've dealt with the consequences for our use-cases. Focussing on that, can only take us backwards, though at least we've proved that the "separate /usr is broken" myth is exactly that: a complete load of myth.  It's exactly the same situation as 15 years ago: if you need modules to get access to your rootfs, you need an initramfs (pka: initrd.)
To my mind, it's clear that Council has no possible mandate over anything like ComRel; what used to be devrel and userrel, and run with much better ability by fmmcor (sadly missed) and jmbsvicetto.
As time, experience, and they've shown, technical leads are usually terrible at handling social matters; that's why Council is, and always has been, strictly about technical matters. The developers always wanted it like that.
So this is not contentious, unless someone is politicking. |
And should the Council take the Debian approach, where politicking is used to force a political decision regarding mandating systemd? Would you rather wait for that day to come, the decision to be surreptitiously dropped on us, and then not be able to be appealed, ala /usr?
I would rather expose and correct the corrupt dealings of the Council now before we're in a complete position of weakness, as happened with /usr. |
|
| Back to top |
|
|
mrbassie
l33t
Joined: 31 May 2013
Posts: 772
Location: over here
|
| Posted: Sat Sep 20, 2014 4:55 pm Post subject: |
|
|
| saellaven wrote: |
And should the Council take the Debian approach, where politicking is used to force a political decision regarding mandating systemd? Would you rather wait for that day to come, the decision to be surreptitiously dropped on us, and then not be able to be appealed, ala /usr?
I would rather expose and correct the corrupt dealings of the Council now before we're in a complete position of weakness, as happened with /usr. |
You could organise an online petition to not follow Debian and asking for the patchset to be accepted and for no further breakage in acquiesence to freedesktop.org, I'd sign it, I'd hope the rest of the 89% would too.
What do you think? Maybe something like that showing up on the mailing list/lists would make more of an impact than a forum poll. |
|
| Back to top |
|
|
tld
Veteran
Joined: 09 Dec 2003
Posts: 1816
|
| Posted: Sat Sep 20, 2014 5:43 pm Post subject: |
|
|
Excuse the fact that this has little to do with the current flow of this topic, but it is related to systemd. I'm dumbfounded by the reply I just got to the following soylentnews comment I'd made earlier, when I mentioned the travesty that is the binary log format:
http://soylentnews.org/comments.pl?sid=3971&cid=95578
He acts as though he's making the case for binary formats over text formats. Correct me if I'm wrong but he does nothing of the sort. I just don't get were some folks are coming from. Stunning. |
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 54237
Location: 56N 3W
|
| Posted: Sat Sep 20, 2014 6:45 pm Post subject: |
|
|
mrbassie,
I would not sign, since I'm not able to support my signature by contributing to the workload.
You can have all the signatures in the world but what counts is the code and documentatian to make it useful.
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
saellaven
l33t
Joined: 23 Jul 2006
Posts: 646
|
| Posted: Sat Sep 20, 2014 6:51 pm Post subject: |
|
|
| NeddySeagoon wrote: | mrbassie,
I would not sign, since I'm not able to support my signature by contributing to the workload.
You can have all the signatures in the world but what counts is the code and documentatian to make it useful. |
The code has existed essentially from the beginning of Linux... I have always* had a separate /usr without an initramfs (the consequence of which means all of my necessary kernel drivers have to be compiled in).
The code to ensure a separate /usr works with OpenRC already exists but a hostile Gentoo dev upstream refuses to accept it because his loyalties are not to Gentoo or the project he leads.
Nothing else needs to be patched at this point... and if it does need to be patched, it is precisely because of a hostile upstream (which could be as far up as Gentoo according to the Council) trying to break things that already work. That is simple enough to deal with given the code to make things work already exists... and if things become so convoluted with the sole intent to make it hard to patch that code back in, then we need to consider whether or not that upstream source is stable enough to use at all.
The entire /usr situation was of one person's making and, ultimately, that person wasn't even Lennart, as I'm told systemd ended up keeping the separate /usr functionality after all. That person is WilliamH.
* Edit to add: I first started with Linux in 1993, got serious about it by 1995 and was using it exclusively by 1998, at which point, I was rolling my own distro from scratch, so I have long known the fundamental components of getting it up and running and separate /usr has always worked. It only "broke" when Lennart arrogantly wanted to break it but backed down, only for WilliamH to abuse his positions to both hold back OpenRC and pave the way on Gentoo should Lennart ever change his mind again... and he did so at the expense of breaking many existing systems and randomly breaking future systems should something like a non-backwards compatible on-disk format change when upgrading lvm2 but not updating the initramfs (that has already happened and will happen again... and won't break during the upgrade, but might break weeks down the road when the person finally reboots and least expects it).
Last edited by saellaven on Sat Sep 20, 2014 7:28 pm; edited 1 time in total |
|
| Back to top |
|
|
Roman_Gruber
Advocate
Joined: 03 Oct 2006
Posts: 3846
Location: Austro Bavaria
|
| Posted: Sat Sep 20, 2014 7:14 pm Post subject: |
|
|
The most thing that annoys me recently is that nothing is in the place as it should be. and tons of useless symlinks to establish a new structure. ebuilds break and asks for symlinks which i find than on bugs.gentoo.org as workaround
They had a reason why usr is usr and so on. Some stuff was optional to mount later read only and now it is writeable usr and other stuff.
That gives me a headache especially when you want an easy, less junk free system. Just an example: mount works fine and why do i need a gui?
At the end of the day it only matters on how fast something is done and very often bash rules over anything. Simple and beautiful
They more I see recently about pam / systemd / udisks / policykit / webkit I really wonder how things came.
End of rant |
|
| Back to top |
|
|
szatox
Advocate
Joined: 27 Aug 2013
Posts: 3135
|
| Posted: Sat Sep 20, 2014 7:42 pm Post subject: |
|
|
| Well, the best thing about gentoo is that it's not usable by default. We have to tune it to our needs, we more or less know how to do that. If things go wrong enough, it will be relatively easy to fork. Probably easier than in case of any other distro. I'm not a big fan of this idea, but almost all things that would be needed are already there. |
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 54237
Location: 56N 3W
|
| Posted: Sat Sep 20, 2014 7:46 pm Post subject: |
|
|
saellaven,
Thank you for the correction.
I always used a separate /usr with no issues and no initrd until I built my current raid5 based box in 2009.
At first I used kernel raid autoassemble for the raid sets and had root outside LVM. I wimped out of root on LVM at that time as I wanted to avoid an initrd.
I still have that install but its unmaintained as of kernel 3.8.0. It aso has an early Gnome3 desktop. It got harder and harder to keep gnome without systemd and I was not too pleased with the way Gnome was going anyway.
I made a new install with eudev and Xfce ... but I had to learn about root in LVM and initrds now, as I wanted to keep the old install and the only space I had left was LVM on raid5.
Since then I have always used LVM for everything.
When the udev tarball was sucked into systemd, it was clear that the writing was on the wall for udev. Its not clear to me how long eudev can survive the assimilation of udev, so I started looking at static /dev again, with no auto anything. It mostly works too. I have carried ovor all the bad habits from earler initrds here. The initrd mounts both /var and /usr but doesn't really need to. it also tries to umount /dev, which of course fails :)
What of the rest of the /usr move, which on Gentoo has hardly started.?
More and more things that should be in / are moving to /usr. Ultimately the initrd will become the new /. Hold that thought a moment.
An initrd is just a throw away root filesystem. There is actually no requirement to use an initrd at all. The initrd can just be a throw away root filesystem on disk. It serves the same purpose in the same way as an initrd, its just not compressed into a file. It can make the /usr move irrelevant too, as it can safely be an all in one partition throw away root filesystem.
I have been tempted to test the above setup but I would need an initrd to start my raid5 and lvm to be able to get at my throwaway root filesystem which would then start the real system.
--- Edit ---
Hmm. What are USB sticks for :)
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
Anon-E-moose
Watchman
Joined: 23 May 2008
Posts: 6098
Location: Dallas area
|
| Posted: Sat Sep 20, 2014 9:31 pm Post subject: |
|
|
| NeddySeagoon wrote: | | When the udev tarball was sucked into systemd, it was clear that the writing was on the wall for udev. Its not clear to me how long eudev can survive the assimilation of udev, |
I don't think that eudev needs to keep up with udev.
It was only a few months ago that I moved to eudev from udev 171-r6 (which still worked perfectly fine)
the biggest problem was that the devs were starting to put version limits instead of just having a dependency on "udev" in ebuilds
and it was requiring me to copy any new ebuilds to my local portage and modify it just to take away their stupidity.
I've found that stupidity in lots of places other than udev though,
it's almost like "well we don't have an old version of yada, yada around so you shouldn't either",
it's more of the same systemd arrogance BS.
So I bit the bullet and moved to eudev.
But from what I've seen any old version of udev will work for 99+% of users,
in other words chasing after LP and his cabal is an exercise in futility.
_________________
PRIME x570-pro, 3700x, 6.1 zen kernel
gcc 13, profile 17.0 (custom bare multilib), openrc, wayland |
|
| Back to top |
|
|
Ottre
Tux's lil' helper
Joined: 23 Dec 2012
Posts: 129
|
|
| Back to top |
|
|
Ottre
Tux's lil' helper
Joined: 23 Dec 2012
Posts: 129
|
| Posted: Sat Sep 20, 2014 11:03 pm Post subject: |
|
|
| Anon-E-moose wrote: |
But from what I've seen any old version of udev will work for 99+% of users,
|
Not for people like me, who got a Raspberry Pi when it was released in 2012 and installed Gentoo with gcc 4.5 on it.
I've been applying security updates for a couple of years but I recently had to wipe the SD card and re-install. Udev doesn't compile with gcc 4.5 and too many packages depend on a recent version of udev for me to workaround. The same bug applies to eudev. |
|
| Back to top |
|
|
steveL
Watchman
Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery
|
| Posted: Sat Sep 20, 2014 11:40 pm Post subject: |
|
|
| Ottre wrote: | Not for people like me, who got a Raspberry Pi when it was released in 2012 and installed Gentoo with gcc 4.5 on it.
I've been applying security updates for a couple of years but I recently had to wipe the SD card and re-install. Udev doesn't compile with gcc 4.5 and too many packages depend on a recent version of udev for me to workaround. The same bug applies to eudev. |
Blimey, that is one idiotic bug:
| M Ogilvie wrote: | | They replaced the decades-old common "do { ... } while(0)" local scope construct with something that depends on a new (gcc-only, 4.6 or greater) _Pragma() construct to hide warnings about declaring variables after statements. |
While I don't understand how/why you're reinstalling and not upgrading gcc, the code issue (that and static_assert) really is pathetic, imo.
I'm sure eudev would take patches to correct it: have you asked blueness? |
|
| Back to top |
|
|
krinn
Watchman
Joined: 02 May 2003
Posts: 7470
|
| Posted: Sun Sep 21, 2014 1:37 am Post subject: |
|
|
| NeddySeagoon wrote: |
What of the rest of the /usr move, which on Gentoo has hardly started.? |
Not seeing it doesn't mean they aren't trying to make it already NeddySeagoon, i think you underestimate attacks made over our distro a bit.
Look : https://bugs.gentoo.org/show_bug.cgi?id=517688
If the bug was fix as the user asked first, we would have endup with rpc.statd install in /usr/sbin, making anyone using openrc and nfs broken if /usr wasn't early mount...
The problem with the "no-systemd" and "pro-systemd" groups, is that one group is trying to keep what they have for years, and the other group is working against the first one.
Look at this one too : how much times a file revert should takes for you? https://bugs.gentoo.org/show_bug.cgi?id=510036 |
|
| Back to top |
|
|
Naib
Watchman
Joined: 21 May 2004
Posts: 6051
Location: Removed by Neddy
|
| Posted: Sun Sep 21, 2014 8:44 am Post subject: |
|
|
Uselessd looks interesting to see how far back they can strip it but still provide an interface other applications are choosing to depend upon
_________________
| Quote: | | Removed by Chiitoo |
|
|
| Back to top |
|
|
mayak
n00b
Joined: 16 Jul 2013
Posts: 26
|
| Posted: Sun Sep 21, 2014 11:13 am Post subject: |
|
|
| Naib wrote: | | Uselessd looks interesting to see how far back they can strip it but still provide an interface other applications are choosing to depend upon |
This project looks indeed very interesting. I just came across the following article (written in german):
http://www.computerbase.de/2014-09/erste-abspaltung-von-systemd/
And was going to paste the following URL:
http://uselessd.darknedgy.net/
Good name by the way
I also like the acronym "BSD - Ban System D"
SCNR |
|
| Back to top |
|
|
depontius
Advocate
Joined: 05 May 2004
Posts: 3509
|
| Posted: Tue Sep 23, 2014 12:26 am Post subject: |
|
|
| steveL wrote: |
I don't know why you keep repeating the same mantra, like I'm suddenly going to agree with it. As I said before, not imo: if anyone is tying their project to logind, it's simple enough to replace what they're after, depending on what exactly that is, but and this is the critical point, the vast majority of uses within the desktop arena are simply plain wrong.
|
I'm thinking of 3 things specifically.
1 - GNOME - don't use it except on my work machine, where it's the official desktop. For my own purposes I do without, but others don't feel that way. However I'm willing to consider GNOME a loss, but unfortunately I've also heard about KDE sucking up to systemd. I don't use KDE myself either, but I know others do.
2 - Wayland - right now it's not systemd-exclusive, and probably is far from going that way. However they have added one feature that I really like, running the server as non-root, and that does require systemd.
3 - Misc services - I read a blog by an nfs maintainer, and he seemed kind of tired of multiple init-scripts for different distributions, and has at least sipped on the kool-aid that systemd will simplify and unify it all. For something like nfs he's not about to drop regular init scripts any time soon, but it's a disturbing frame of mind to see from a service maintainer.
Perhaps the vast majority of deskop users are wrong, but by volume they can drag servers with them.
Perhaps I keep up the same mantra because I've been on the losing side of too many battles over the years, and I've generally been technically correct. In the long run it has seldom mattered, no matter how hard one fights. In the real world numbers tend to matter more than technical superiority. My attempt is to recognize that tilting at the windmill will probably not work this time either, but laying groundwork to pick up and rebuild afterward might. Believe me, I've fought this type of thing before.
As for "disillusionment phase"... http://en.wikipedia.org/wiki/Six_phases_of_a_big_project (I'd previously heard of "honeymoon" instead of "enthusiasm".)
_________________
.sigs waste space and bandwidth |
|
| 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
|
|
Gentoo Chat
