I wish I were better with words. There’s thoughts that strike a note in your heart and mind, but I would not be able to express these thoughts on paper so that they deliver the needed punch. That was my first thought when I read this open letter on the Linux Kernel Mailing List (LKML): https://lkml.org/lkml/2014/8/12/459 . The text is written by a longtime Debian user who feels deeply betrayed by its board of leadership. The emotions he penned down are exactly mine. Thank you, Christopher Barry. This was of course not the first eloquently written rant, but I hope it sparks a discussion in Kernel Land about what is happening in User Land, and whether they can afford to keep looking the other way (with the public exception of Linus and some others).
One word. One demon. systemd.
What relation does Christopher’s rant have to Slackware? After all, it’s Debian that got the flak, and in the comments section people indicate they intend to switch to Gentoo… forgetting that Slackware is a good systemd-free alternative (but hey! this automatic dependency resolution thingie that makes life so comfortable in Gentoo is not part of Slackware either).
Last week I asked the SDDM developers to reconsider their decision no longer to support ConsoleKit because Slackware does not have systemd or logind and thus we need to keep using ConsoleKit. The answer could be expected: “answer is no because ConsoleKit is deprecated and is not maintained anymore” and therefore I had to patch it in myself.
Of course, the ConsoleKit successor systemd-logind, written by the same team that gave us all the *Kit crap, depends on PAM which we also do not have in Slackware. One of the fellow core developers in Slackware, who is intimately familiar with the KDE developers community, has heard from multiple sources that KDE is moving towards a hard dependency on systemd (probably because they are going to need the functionality of systemd-logind). We all know what that means, folks! It will be the day that I must stop delivering you new KDE package releases for Slackware. That’ll be the day.
Eric
I deeply hope that all those who are expressing themselves against the general trend (systemd and other things) will keep up their commitment, and go on developing alternatives and mantaining patches to avoid the dependencies that many big projects are introducing.
And if there’s need of financial support, I’m not refusing to offer contribution.
Hi, long time lurking reader here… a first here of some sort.
Well, I also share the sentinement of Christopher Barry and though I currently don’t use Slackware, I’m tinkering with it every so other year in hope of being vaguely “better” than last time.
Currently, for my desktop home usage, it’s too much work, but I’m derailing. The point is, that I’m really fearing that Slackware is not one of the oldest and traditional linux distributions, but also being one of the last “real” linux distributions.
Dunno, what Patrick will be doing in the future, but will be the future of Slackware leaving linux and joining the BSD crowd?
Heh. Last year I think it was, the team have jokingly discussed a Slackware transition to BSD, leaving behind all the GNU and systemd politiks and focusing on what really matters: an OS that empowers its user without compromises…
Hey Eric, your words are *excellent*, as usual.
At the moment systemd has three subsystems that significantly affect all Linux distros – logind, udev, and cgroups (ignoring kdbus for the moment).
There is hope for a solution to the logind problem — Debian has systemd-shim (https://packages.debian.org/sid/systemd-shim) and OpenBSD has a GSOC project that looks promising.
The udev problem is more worrying. Gentoo’s eudev would be a suitable solution, except that systemd seems recently to have found a way to torpedo it (see http://lists.freedesktop.org/archives/systemd-devel/2014-May/019657.html and http://lists.freedesktop.org/archives/systemd-devel/2014-May/019664.html)
I don’t know enough about the cgroups situation, but reading http://lwn.net/Articles/601840/ it does not seem to be hopeless (please correct me if I’m wrong).
As for kdbus, we will just have to wait and see.
shakeyourbunny is IMO right that there are only a few distros keeping Linux true to its roots — I would add LFS (see in particular Bruce Dubbs three months ago at http://lists.linuxfromscratch.org/pipermail/lfs-dev/2014-May/069284.html which is a nice companion for Eric’s post).
Meanwhile, systemd’s quest to eat the world continues. Someone pointed out (possibly trolling) that systemd-networkd needs a full set of ipv6 tools, so they are happily adding their own radvd, dhcpv6 etc. And it’s now got its own special tools for backlight, rfkill and an ntpd replacement… what next?
Warning: I have not had coffee yet.
While normally I wish Debian wouldn’t air their dirty/dysfunctional laundry in public, I’m glad they are this time. SystemD is almost as worrying as SecureBoot, and it’s something the Linux community needs to take a good, hard look at. I’m glad the Slackware higher ups are more sober in their consideration of it. (Truth be told, I’m more worried about SecureBoot. I trust Slackware’s devs, and if systemd is inevitable, I’m hopeful that it’ll be implemented in the best way possible.)
On the one hand, I’m glad distros like Ubuntu bring in people who might not use Linux. On the other hand, I dislike how distros like Ubuntu try to force their design philosophy on everyone else.
My big worry is that these policy decisions from the big distros force the smaller ones out of existence. In some respects, it feels like the end of the 1980’s, when all the smaller (that is, non-Apple, non-Windows) computer manufacturers were forced out of business. Conversely, before Ubuntu was Red Hat, and Red Hat didn’t take over the world. So maybe there’s hope?
But, ugh. I’m with all of y’all. Newer isn’t necessarily better.
KDE. My fist DE in Mandrake moons ago. Was nice knowing you.
What next. GTK? I guess, since gnome is already there. That means Firefox and tons of other stuff.
None of this makes sense. The sheer beauty of Linux is its diversity. Circumventing that will kill it. Maybe not today, or tomorrow but soon enough.
Linus, take note.
The sad news for me is the idea of not being able to run future KDE releases on FreeBSD. It will be haldaemon all over again but of a much greater magnitude with porting issues. I doubt systemd will make it onto FreeBSD, however i can see that by the time it did, systemd would suddenly become deprecated, leaving the BSD guys with another deprecated waste of time!
Well, I’m just watching all these systemd thing for a while now.
Do I like KDE? Yes, I do. Can I live without it? Sure I can.
I can live without Slackware too. The thing is that, despite almost all other Linix distro out there, it respects my wishes. So even if I can live without it, I don’t want to.
The same is not true for KDE or every other Slackware package.
I wish KDE team reconsider their path. SDDM is great, but it’ll get out my rc scripts next after this. KDE could have the same fate…
> Comment from alienbob
> Posted: August 14, 2014 at 11:28
> Heh. Last year I think it was, the team have jokingly
> discussed a Slackware transition to BSD, leaving
> behind all the GNU and systemd politiks and focusing
> on what really matters: an OS that empowers its user > without compromises…
Sounds like a good idea!
🙂
It’ll be interesting to see what Patrick and team will do when that happens.
I never use consolekit, which seemed mostly pointless to me, the only real benefit was for multiseat, which was never fully implemented in consolekit, but could be made to work very painfully with some other stuff. Multiseat is an edge case, and one that doesn’t make sense anymore from a TCO of administrating it plus additional hardware, with the prevalence of small cheap powerfull very energy efficient computers.
Fast user switching was always possible without consolekit, I have always used xdm, and just set it up to have the login on all my virtual X consoles, so a quick Ctrl-Alt-F8 etc get me to another xdm login. If you have say 6 users, using a machine, just assign them a specific virtual console and put it on a notecard attached to the monitor so they won’t forget.
My main wm is openbox, but I also will play around in KDE on a second x console and a different user, and I’ve removed the consolskit stuff from my xinitrc and rc.consolekit is disabled. It works fine, you won’t have the default shutdown and restart and switch user available. It’s possible to either add shutdown and restart button to xdm, to do these after logout, or you could make scripts and menu items to do this within KDE, as far as user switching, just stick with good old console switching.
As a software developer, I’ve been considering getting into contributing to the open-source community for a while now, but I’ve always been intimidated because there’s so much I don’t know, and I’m not sure where to start. Perhaps helping maintain those projects that were discarded in favor of systemd would be something to look into . . .
Hello Eric,
I read the open letter and could make it my words. I saw that Christopher Barry is now looking for alternatives to replace Debian. Did you had a chance to suggest Slackware as an alternative?
It looks like that people don’t realize how easy and simple it is to manage a Slackware box with slackpkg+ and sbopkg.
Now is the Time for All Good Slackerz to come to the aid of their POSIX. Now is the time for All Good Slackerz to come to the aid of REALLY FREE software. Now is the time for All Good Slackerz to RAISE HELL WITH ANYONE showing symptoms of caving in to this filthy corporate Trojan Horse called “systemd.” Hats off to Eric and Chris Barry!
Spread the Word, “CHOOSE SLACK! and Don’t look back.”
Now to get to work on making Erics ARMv7h stuff run nice on my Chromebook!
By the way…
https://lh3.googleusercontent.com/-bZId5j2jREQ/U-vlysklvCI/AAAAAAAACrA/B4JggkVJi38/w426-h284/bd0fb252416206158627fb0b1bff9b4779dca13f.gif
I’m so happy that this discussion has surfaced once again since it really scares me when things are forced on me.
Wrote some more on google+
https://plus.google.com/115702648539892354883/posts/3jW1x7Piwcn
LONG POST ALERT!!!!
My knowledge on how Linux works is not very broad, but i’m a longtime Slackware user by now, and although i’ve been tempted numerous times by other distros because of features such as automatic dependency solving, fast boot, availability of popular software packages, nice desktop integration of apps, supported out of the box multilib compatibility, etc. etc., every time and despite it lacks many of those features, i keep returning to Slackware.
This is almost for an unknown reason to me… almost. The reason is because of Slackware’s simplicity, and the fact that it doesn’t “betray” you after an update, so it really always works after you sometimes go through the corresponding investigation to do whatever it is you need. And besides, you learn in the process.
I don’t really know how is it that for instance, systemd differs from ConsoleKit, but by reading you, Eric, and other programmers who seem to be veteran, and that effectively know what they are talking about (because Slackware demonstrates it), it sounds to me that modularity, security and other Unix principles adopted as good software development practices are being neglected by these new trend of software. I’m truly amazed at how there is people out there like you who understand and are able to recognize these issues in such a clear way. By the way, I’m a programmer, currently working for a videogame dev. company, but I haven’t developed a vision such as yours for these things.
Ok, this became a long post indeed, but i wanted to express these things, and my last point: I believe Linux has reached a level where it is gaining more and more attention from commercial companies who don’t want to depend on Microsoft, but still need vertical solutions, and are able to take advantage from this now very mature OS as a base. Maybe this is one of Linux’s final contributions to the commercial world, and many popular commercial apps are moving along this change, teaching us that maybe, the commercial world will never be compatible with concepts such as efficiency, security, stability, but instead, with “convenience” in favor of profit (I hope as a gamer, that this doesn’t mean Steam will someday follow that path as well). The sad part is that even non-commercial products such as KDE are also leaning that way, and so even smaller user/developer groups will have to handle the generation of alternatives that keep the old philosophy alive, almost like starting over (Like a Linux revamp, or something alike), and trying to get support from the commercial side into this new reborn OS once again (… yes, like Steam…).
Crazy times.
I think the battle over the linux distributions which init system to use is more ore less over and systemd is the winner, marginalising the “normal” linux distributions.
I disregard for now the fanatical religious systemd zealots and Poetterings ego for now, the main reason for systemd is not really that bad (unification to more or less one linux system), but the execution is riddled by the prospect of wresting control over the system to systemd, feature creep and – my personal main reason, why I dislike systemd: binary log files. Just try to find something in the Windows Event Log, especially if the system is toast. After that endeavour you got gray hair and you will be full of hate.
Another problem is the “not invented here” attitude of the more recent linux programmer who discard anything what they did not create, labelling these things as “deprecated” and defending aggressively their own stuff while forcing the software into all others throats.
The prospect of faster boot times for the end-user is a fake trojan horse: this is dictated by the hardware (mainly storage) and how many services you start at boot time.
Eric, bad news. I just read this post from Martin Gräblin – Breaking News: kwin_wayland will hard runtime depend on logind?
https://plus.google.com/115606635748721265446/posts/Pw31EE21BXV
Yeah, he really means it when he says Breaking News. Well, now we need to know if they will maintain kwin_x11 without the need of systemd or if we can run KDE with another window manager.
“It will be the day that I must stop delivering you new KDE package releases for Slackware.”
For me, KDE is in the same ilk as systemd. I stopped even looking at KDE when NeoPunk and AnaKonda were introduced.
Eric, has the team looked at OpenWRT’s procd?
http://wiki.openwrt.org/doc/techref/procd
Hi Xsane
It’s good that Slackware leaves you the choice 😉 There are several alternatives as part of Slackware and in 3rd-party repositories.
We have never looked at procd – we do not need a replacement for our current init system.
https://plus.google.com/+MartinGräßlin/posts/Pw31EE21BXV
Is this the final nail in the coffin?
Hi Eric,
Yes, choice is a beautify thing. My comment was not intended as a stab or complaint. I know many people appreciate your kde work, as I appreciate all of your time and effort as well!
I don’t want Slackware’s init to change either. I just wondered if y’all were aware of the procd/hotplug2 project in case the_hand_is_forced somehow.
Back in the pre-automation 2.0 days, installing any distro on then-current hardware required tweaking configuration files and even driver level components, especially for network, disk controller and graphics, to even get a successful boot. udev and the like were not even a concept.
I may have lost those particular skills, but the ability to think critically, what pulled us all to linux in the first place, has not changed. Personally, rather than fretting about the disaster ‘systemd’, I can spend time now in code review with consolekit enough so that I can, once again, tweak things for myself if necessary. I’d rather stick with the last pre-systemd Slackware and do subsystem upgrades by hand if it comes to that.
Eventually a ‘last pre-systemd Slackware’ fork might be a necessary path. This is how it all started anyway, when some uni student forked minix for essentially similar reasons.
Cheers,
@Tim: Linus didn’t fork Minix. He started to code from scratch on his 30386. There’s no Minix code in the first Linux kernel.
… forked minix …
Linux was a response to Tanenbaum’s refusal to accept changes to minix. My point was as a bifurcation point. Thanks for the clarification Niki.
@Eric: does Martin Gräßlins announcement mean that there’s no future for KDE on Slackware in the long run (that is, without systemd)?
Niki,
What I said was in reaction to his sarcastic and somewhat ridiculous remark about systemd haters.
I tthink we will try as hard as we can to keep KDE in Slackware but I do not think we will bring in systemd to keep KDE functional.
One interesting option would be to keep the KDE applications but get rid of the window manager and the session manager when those start requiring systemd.
There are alternatives like LXQT that provide a Qt based desktop environment in which the KDE Applications run perfectly.
Also think of Plasma being re-written so that in theory it can run on devices that range from high-end desktops to mobile phones… have you yet encountered a mobile device (phone, laptop) with systemd? No!
Hi Eric,
this is bad news indeed. I do hope Slackware remains systemd-free, especially because I think systemd might become obsolete in a few years anyway (just like Upstart, which was new and shiny, or HAL).
I wanted to ask exactly that: can’t we keep the KDE apps like gwenview, okular, kate, kwrite (maybe dolphin too) but use LXQT? This could work, couldn’t it? I really don’t want to switch away from Slackware, I love it and its community, so many helpful and knowledgable people, such a nice KISS distro.
As for a phone running systemd (and wayland): check out the Jolla phone running Sailfish OS, or Samsung’s Tizen. I think they run systemd. But of course from a popularity point of view, they are no match for Android, which uses its own init (I think).
lems
By the way, do you know anything new about KDE 5’s planned modularization towards Qt5-Add-Ons?
(I wasn’t fond of the Plasma Shell towards the user perspectives and the entire semantic desktop project, this announcement gave me some relief, but that’s not my main matter.) Modularization of KDE to Qt may benefit and even be the last resort for “saving” some apps and hard dependencies for several distributions once KDE stops providing any init interaction but systemd. (Remark: The Jolla Sailfish OS Phone uses systemd, however not Android, and through busybox, it may not be necessary to change that fast in the future. Some efforts of porting busybox to desktop machines are already taken.)
For login managers, niches will continue to exist, allthough maybe limited to minimalistic – already existing ones like qingy (has support, but does not depend on pam), SLiM don’t offer noticeably much less features than sddm – pure Qt attracts init agnotic developers anyways. LxQt isn’t going to accept systemd-only implementations in the near future, maybe a fork will see the light one day.
So much more for KWin, I fail to see what makes KWin so useful outside of the Plasma Shell, unless it keeps staying a hard dep.
We’re not alone, I see it rather positive how much effort is already taken through entire communities to stay indedependent from systemd, even more how many may switch to Slackware once people become entirely fed up with systemd. It may become a hard time of abandonment,
but I can see where fractions want to go. In the end, the applications people will use are the one they want, it’s Linux making it possible.
I was a long time Arch user myself until I left with a breakage incomparable to any breakage I was used to introduced by systemd – so much more with developer’s direction in general from that point on. Gladly I’ve been using a VM, I could only choose Slackware current for my main desktop, and I’m not going back unless maybe Gentoo on a higher powerful future PC.
So long, systemd zealots can have fun with indebuggable, yet simultaneously nonstop restarting and crashing binaries.
Even rescue CD chroots are useless if the error isn’t either to spot in plaintext configuration/service files or if it uses an incompatible version of systemd itself.
Speaking of Arch, there is a relativley new distribution called Void Linux; it’s a rolling release distro with binary packages, like Arch. XBPS is its package manager, mainly written by Juan RP, a former NetBSD developer. I just wanted to say that Void Linux recently switched _from_ systemd to runit. It’s not a big distro, but anyway, just thought I add another distribution that at least does not default to systemd (systemd is still available as an option, I think).
lems
Eric,
Thanks for your detailed heads-up. LXQT with a bunch of KDE apps sounds like a *very* interesting alternative.
Let’s hope that at one point Linus steps in (like he did with Git) and writes his own init system. He will probably call it “Boot Initialization Through Configured Handlers” or something that makes for a nice acronym. :o)
Well, in the end, all this noise appears to be just one douche dev throwing up bs on our faces… Am I right?
Is it possible for someone to write a brief synopsis of the current state of things? Is the bigger problem udev? C. Barry made many good points. I think his post may act as a lightning rod or if you prefer catalyst. Too many good, smart people are fed up with the complete crap that’s being forced for the course to continue.
1) init works fine no need to “fix” it.
2) udev seems to be ok right now but for how long?
3) I use fluxbox so don’t know about kde anymore, sounds lxqt will work fine
4) a summary of the “threats” would be good
Is the threat simply the distros making a bad choice in systemd or is it more than that?
More and more it seems that Systemd et all are trying to build the equivalent of Active Directory for Linux. All fine and dandy for corporate/government/military use, but massive overkill for most home users.
Moving to BSD is not a bad idea after all, but what will happen when BSD world decide to adopt that systemd thing? There should be a solution out there, and maybe linux world should ditch systemd … at least is what i think. :/
Iff FreeBSD replaced its init, they’d rather use launchd, for which initial porting work already exists. As for OpenBSD, some developers thought about implementing os-agnostic replacements for systemd functions (de Raadt hates Linux actually). I also don’t see how systemd can be easily ported to a BSD; I think systemd is very linux-centric, which was Poettering’s main motivation. I feel sorry for the guy(s) who ported HAL to FreeBSD … it was all for nought.
please excuse the bad english.
If there’s something good about this piece of crap called systemd is this one: in time it will force a clear separation between “i want to be windows” linux distributions and true *nix based distributions. First ones, burn in hell you heretics, second ones, welcome back home!
It turns out that logind is only used (runtime dependency) for kwin_wayland, not kwin_x11, so i guess KDE can still be used without having to add systemd and/or PAM (at least for KDE 5).
See here: https://plus.google.com/115606635748721265446/posts/79uY8LQ3yNx
lems already hinted at it but I’ll post the URL just in case:
http://www.openbsdfoundation.org/gsoc2014.html#systemd
I don’t know if someone is already working on that, but it might be worth keeping an eye on it.
I just realized Dave Spencer aldo mentioned it, sorry for no reading all the comments before posting.
An aside question. I’ve never used LXQT. Is it possible to install single KDE apps like Dolphin, Okular or K3B within LXQT? Do they still need kdelibs as build and/or runtime dependencies? I prefer to ask before hours of possibly fruitless experiences.
Of course they will need the kdelibs and other packages installed. The KDE applications are not self-contained.
Lets just hope that projects like systemd-shim and the OpenBSD one do not end up endlessly chasing a moving target like Wine has.
it’s possible to run lxqt with kwin as the windows manager which produces an interesting hybrid.
Plasma 5 is already here and running. kwin has requirement for logind only with Wayland.
So when is crunch time for KDE on Slackware? A few releases away?
I would be surprised if the fine folks from KDE did really intend to leave so many people standing in the rain: Slackware, Gentoo, Crux, all the BSDs… there’s a follow-up post by Aaron Seigo on the subject.
So to sum this up. We have a new monolithic monster of a program replacing the old init systems to which a major, perhaps _the_ major desktop is building in a hard dependency. Worse still Wayland looks like the future of graphical desktops is making it a hard dependency.
Whether this is a good idea seems irrelevant as there seems to be no choice. You can’t substitute one component.
Hmm. Will even LXDE be runnable under wayland with no systemd? Lets just hope that someone decides that alternatives are a good idea.
Though I’m not a great bash/shell programming fan I can see the power of a modular scripted init system. Surely the only dependencies for a desktop environment / windowing system is that all necessary drives are mounted, all config files are in sane places and all essential services are running. It should not make one jot of difference how you get there.
I’m not sure that this adds anything to the conversation save a bit of a rant.
As far as I have read, systemd may not be as monolithic as some think. I think it can be possible to just install logind without the rest. But the thing is that it seems to require PAM, which has always been out of Slackware. On the other hand, in the commentaries of the last blog from Martin Graesslin, it seems that there may be a solution in the future for them without logind at all, with possible additions to the kernel of some functions related to file descriptors.
Another solution is to create a standalone daemon that would implement the logind API, as I understood. I suspect it can be possible if indeed so many are left in the cold.
I found the commentary by Aaron Seigo on the initial blog by Martin Graesslin quite sensible, BTW.
The reason I started to use Slackware a few years back and keep using it until this day is that it stays true to original Linux principles. I came from Windows (through Ubuntu) and never before have I enjoyed such a freedom and simplicity as in Slackware. Love it and wouldn’t leave it. Recommend it to everybody I meet.
That said I trust the Slackware team will do their best to cope with the situation (as they always do). I’m not really an expert OS developer but from what I understand there are still ways to have usable Linux without this systemd craziness for there’s a lot of sensible people able to continue developing the original solutions.
If at some point all hope is lost, switching to BSD also seems like an interesting idea. As long as Slackware is healthy at its core (meaning the team and the community) I believe we’ll make it no matter how hard the obstacles will be. But perhaps this whole systemd monster will fall and break under its own weight before it comes to that.
And sorry for not adding any technical value to this thread, I just wanted to show some support to Eric and the rest of Slackware team to whom I hold greatest respect.
I’ll start this by saying that I was an early implementer of systemd on my Arch Linux installation when it first became available. At the time, I had Arch on my main system as a tester Linux. I had already been running it for a couple years when I switched it to systemd. It was interesting. I could see its advantages and disadvantages, but my main dislike for it was the fact that it seemed to violate the KISS principle, which Arch was very vehement about following. I felt that Arch sold out that principle when they went to systemd officially.
I don’t run any tester Linuxes on my systems these days. I do have slots available on my shop system, but they’re all empty. Slackware has been my primary OS for 8 years now. I don’t need anything else.
Sad about KDE, but it wouldn’t really affect me much. My love of KDE faded with the release of the buggy 4.0. I switched to Xfce and never looked back. I ran KDE for a while back on 14.0 (Slackware), when it first got released. When the 14.1 upgrade came out, I went back to Xfce. KDE is too pretty for me; and presents too many options for the user. I prefer to keep it beautifully simple… Xfce.
So, if Eric will not continue maintaining KDE for Slackware in the future, it would not be bothersome to me at all; and probably free up quite a bit of time for Eric to do other things he enjoys, like sleeping or maybe playing golf?
I don’t fear systemd, but I don’t embrace it either.
Regards,
~Eric L. (aka Nocturnal Slacker, vtel57
Playing golf??? I will be found dead before I play golf. My form of recreation is listening to music and reading a book. MY work on Slackware is subservient to being able to listen to my on-disk music collection and managing my e-book collection. The rest is just by-products 😉
Alien Bob said:
“Playing golf??? I will be found dead before I play golf. My form of recreation is listening to music and reading a book.”
A man after my own heart! 🙂
The plan, I believe, is for all Wayland compositors to use systemd-logind’s capabilities to provide user access to the console display without requiring root or setuid. This is a feature in recent Xorg versions too, but I think with Wayland you don’t have a choice whereas you can still run Xorg setuid and not rely on logind.
I can totally see Slackware sticking it out with X even as Wayland becomes the new standard. And that’ll last for a few years yet, maybe. The problem is that, like ConsoleKit, developers up-stack from the graphics server will deprecate X support in major toolkits and apps and it will bitrot.
So like I said — Wayland, systemd, and PulseAudio are the emerging standards. You either support them, accept responsibility for maintaining your own local forks, or watch things break.
The path of least resistance is simply for Slackware to support all three in the coming years (and PAM, it’s about time that happened).
More criticism from a Linux veteran. https://plus.google.com/+TheodoreTso/posts/4W6rrMMvhWU
I’ve just realized it’s not gonna be that bad, nobody is forcing us to switch to systemd, that udev thing is probably going to be bypassed sooner or later and as for apps depending on systemd – well, there always were alternatives in Linux ecosystem and there will be (as long as not everyone migrates to BSD). Let systemd stay on servers (and embedded), just like they intended.
My vote goes to the Enlightenment as the default interface of Slackware GNU/Linux.
http://www.enlightenment.org/ss/e-53fcc2302b22a7.86659975.jpg
We cannot desert Richard Stallman.
Near as I can tell, Rasterman’s intentions are to eventually abandon X and switch E entirely to Wayland. Similar writing is on the wall for the big DEs. With Wayland on Linux, systemd is an absolute must: there is no support for running Wayland compositors as root to give them access to graphics hw. Such access must be negotiated through logind.
Over the long term, the flaky “virtual console” system is bound to be removed from the kernel and replaced with a user-space console multiplexer (perhaps Wayland) meaning that systemd-logind will be the only way to log in to your Linux system.
Systemd isn’t just a startup service, it’s a comprehensive rearchitecting of the entire user-space infrastructure of Linux into an integrated whole, rather than a loose kitbag of standalone components. One does not simply decide not to use systemd and expect that your use case will be in any wise supported by major software products’ upstreams.
“Playing golf??? I will be found dead before I play golf.”
My dad was narcoleptic. I think he just watched too much golf. 😉 Theory doesn’t hold for my sister, though.
“there always were alternatives in Linux ecosystem and there will be (as long as not everyone migrates to BSD).”
This is my thinking, too. In a sense, it feels like the 90’s again: hardware changes, lot of forks, people arguing about everything.
Add my vote for Enlightenment. I used it for years, and it never used to be that hard to build from source.
The biggest problem is that the BDS not possibly survive all this. Using BSD solution will be difficult.
Perhaps not all is lost. There is also the possibility of someone implement a solution to break these dependencies, we know to be dificel but should not be impossible.
So today the next step in the “systemd cabal” Grand Plan has just been published.
http://0pointer.net/blog/revisiting-how-we-put-together-linux-systems.html
“Revisiting How We Put Together Linux Systems” … “This is a reinvention how distributions work, and hence needs great support from the distributions”
It’s not necessary to list arguments against the Grand Plan here, because Hacker News is doing a great job 🙂
see https://news.ycombinator.com/item?id=8251288
@alienbob
> we do not need a replacement for our current init system
This is not a STRATEGY, not one that would work against systemd
and whose *pulsions* are more related to the clinical psychology
(compulsivity, e.g.; oral phase; etc.) than with reason.
BTW (nota bene, BTW) Supervision suites should^1
not wake up unless notified
be bug-free, lightweight and easy to understand
provide a basis for high-level service management
When init lacks on these.
Systemd’s *poison* (sort of self-harm, to understand us) needs to
be taken out with the sting inclusive. Think also at some
Pythagoras theorem *demonstration* written in 300 pages of
formulae: such a thing would rather belong to an Art Brut museum.
Show how it’s done properly instead.
1. http://skarnet.org/software/s6/why.html
Why another supervision suite?
Adendum
http://boycottsystemd.org/
Funny (façon de parler), PREDICTABLE systemd *unfolding*:
[Twitter:]
grsecurity@grsecurity 2:23 PM – 16 Sep 2014
Lennart’s response to systemd bug report on grsec kernel: “use a
supported kernel”.
grsecurity@grsecurity 2:24 PM – 16 Sep 2014
His response when it’s pointed out that the same feature exists
upstream: “we don’t support the kernel”
https://bugs.freedesktop.org/show_bug.cgi?id=65575#c8 …
grsecurity@grsecurity 2:27 PM – 16 Sep 2014
Funny that it’s “not compatible with how operating systems work
these days” but worked fine for a decade and a half before he
came along…
grsecurity@grsecurity 3:18 PM – 16 Sep 2014
I think the time wasted by the open source community in general
on his code churn will far outweigh any supposed benefits
[Retweets 4 Favorites 5]
[Reply ex.: someone @grsecurity I wish I could retweet this 1000 times.] 🙂
Errare humanum est, perseverare autem diabolicum:
Uselessd: A Stripped Down Version Of Systemd
http://www.phoronix.com/scan.php?page=news_item&px=MTc5MzA
Much better still: systemdrop, systemdesist, systemdiscard,
systemditch, …
In accordance with the basic engineering principle: [Antoine de
Saint-Exupéry] perfection is achieved, not when there is nothing
more to add, but when there is nothing left to take away.
The matter did not end sytemd debian project. Recently this subject was:
http://www.phoronix.com/scan.php?page=news_item&px=MTgxNjM
and they are not kidding:
http://debianfork.org/
Go, Alien go!
Good to see all these folks here…
No dconf or gconf, no ConsoleKit… oh heck, I can’t keep track of the stuff I dump off my computer. No predigested desktop or office suite and wicked fast. With a 64-bit kernel and 9″ screen thank you very much.
And, if something breaks I know who to blame, me, yea!
What do I care about other people that store every file on the desktop? As well as sending all their information to google… what a joke, google, made big money with a web page that has one type in box… OMG.
Go Slack and don’t look back!
pleas
i not understand how i can make a usb live with your iso
it’s 64 bits only, or 32 bits machines ???
i can use dd if=(name.iso) of=/dev/sdX bs= (kb,M,GB)
your scrip, as it used?
Mauricio Yaman what does this have to do with the article above?
Re-post your question in the comments section of a relevant article please.