Updated Info On SpaceFM
Greetings! Just thought I’d check in from my extended hiatus and offer a few info items on SpaceFM.
My development work on SpaceFM and my other projects is still currently suspended, so no change there, but mostly they are still running as they were. I’ve been working elsewhere and have only been a user on Linux lately. I can’t tell you much about my plans, except that I am that much more determined to not ever run a system that includes systemd, especially seeing the direction it’s going (IP forwarding, etc), growing way beyond a safe and stable init system. Clearly many people aren’t happy with it, but they never were, and I doubt major distros are going to listen to their users. So I’ve been giving things some time to bubble, seeing what falls out of this mess as options.
When I have some free time, I may try gentoo without systemd, or I may try one of the BSDs. Let me know below if you’ve found a promising road away from systemd. My only real hesitation is my Brother MFC-4720 printer, which is a good printer but always hell to install, and I never could get it working on gentoo or BSD last I tried. But I’m told desperate times call for desperate measures. Once I find my next OS direction, then I will decide what if anything I want to do in the area of software dev. For now I’m just using SpaceFM on a retro Debian system, nice and quiet while all hell is breaking loose in Linux, but I think my days of using Debian are soon done.
A few notes on SpaceFM…
My thanks to previous SpaceFM contributor BwackNinja, who has been maintaining a maintenance fork of SpaceFM with a few bugfixes, plus he has added the ability to have transparent desktop backgrounds. Nice work there, so if you want to use that feature, you can grab the source, and if you have an urgent issue with SpaceFM, you might want to politely bring it to his attention. You are also still welcome to post issues to the main SpaceFM issue tracker, so others can review them, offer possible fixes, and I may eventually see them.
If you encountered an error in the console saying Attempt to unlock mutex that was not locked some months ago when starting SpaceFM, or were unable to start it, this was caused by an update in glib 2.41 which broke many GTK apps, especially when used with GTK 2.24.24. This problem was corrected upstream in the release of GTK 2.24.25, but still may affect some older versions of GTK2, as well as GTK3. BwackNinja’s fork includes a fix for this, and you can read more details here. Thanks to everyone who helped troubleshoot that in my absence!
Also, those using the IgnorantGuru PPA should have noticed a key expired error on my key. Rather than replace the key at this time, I have simply removed the expiration date from my public key (0x01937621), so it’s no longer expired, and have re-uploaded it to keyservers. You can get and add the updated key with these commands, and the PPA should work again:
gpg --keyserver keys.gnupg.net --recv-keys 0x01937621 # If you receive an error, try again later. # Then, add the keys to apt-key: bash -c 'gpg --export -a 01937621 | apt-key add -'
Alternatively, you can use the keyserver at keyserver.ubuntu.com, and it should migrate to others in time.
I haven’t been keeping up with Linux or SpaceFM discussions much, so if there’s something you want me to know (keeping in mind that I’m not currently working on these projects), some thoughts or resources you’d like to share with other SpaceFM users (the homepage directs them here and many users are subscribed), etc., now is your chance to leave your comments, links, etc. I’ll leave this thread open for comments for a few weeks. Also feel free to give any thoughts on anti-systemd migration – I’d like to know what people are using. Thanks and best wishes!
29 Comments
Sorry, the comment form is closed at this time.
IgnorantGuru's Blog 
You can try Slackware – Pat decides that
Sorry for my previous post – I have pressed wrong button on smartphone.
In a 2012 interview, Slackware’s founder Patrick Volkerding also expressed reservations about the systemd architecture:[35]
Concerning systemd, I do like the idea of a faster boot time (obviously), but I also like controlling the startup of the system with shell scripts that are readable, and I’m guessing that’s what most Slackware users prefer too. I don’t spend all day rebooting my machine, and having looked at systemd config files it seems to me a very foreign way of controlling a system to me, and attempting to control services, sockets, devices, mounts, etc., all within one daemon flies in the face of the UNIX concept of doing one thing and doing it well.
From https://en.wikipedia.org/wiki/User:ScotXW/systemd
So you can try Slackware or better Porteus – based on it. No need to install – only download, add correct entries to bootloader and reboot. If you like clean, really quick and simple system – try it.
Yay, glad to hear a sign of life from you! :)
I’m still using Debian sid/unstable, without systemd installed (no Gnome, pulseaudio, etc or other such systemd-ish stuff, just using LXDE quite contentedly for my desktop, happy with a more minimalist simple approach, including enjoying your software, thanks for it!) It seems so far not hard to avoid systemd for me.
I gather that for people who want a more “Windows”-esque “fancy” desktop and so on, it is much harder to avoid systemd (but then most of those people to view systemd as a feature rather than as a bug…)
If it ever becomes hard/impossible for me to avoid systemd with this Debian sid/unstable installation of mine, I expect that will probably be a sign that Linux has been fully converted to the dark side, and I will be tempted to jump to a BSD, probably Open or Free. Unless I’ve first taken the initiative to explore Minix 3, which seems potentially intriguing and even farther from the systemd-bloat-philosophy. :) Gentoo also remains a possibility if I decide I don’t want to have to learn as much new stuff, and if it’s still easily systemd-free…
Bug hell, it’s a virus. :) I’m surprised you’ve been able to keep systemd out of your sid. Multiple things on my sid system wanted to pull in systemd, but I didn’t have a chance to look at the particulars. Sounds hopeful. I thought Jesse was going toward systemd for it’s init, so I had given up on it, but maybe I got that wrong. I was hoping they would release Jesse stable without systemd required, so we’d have one more usable Debian stable. But I’ve heard some good things about Devuan, might be enough people to make it stick.
All of these are stop-gap measures for now, but it’s good to know there’s some momentum away from the systemd route. Not sure if BSD will evade it either though, lots of politics and pressure there as well.
I use CRUX Linux with SpaceFM, and I warmly recommend it: http://crux.nu
It’s small, it’s BSD-like with its port system (everything is compiled, including the kernel), it’s simple, and fast. And of course, no systemd, never.
I’ve been quite happy with Void Linux. http://www.voidlinux.eu/
Seconding Void Linux here! It’s incredibly easy to write packages for the distro. Submitting a package for their repos is as simple as submitting a pull request to the developers. As soon as it’s approved, the buildbot will get to work making a binary package.
The only reason why I haven’t swapped out my desktop machine with is that I haven’t packaged Compiz nor Fusilli (a rather exciting Compiz 0.8.x fork) for Void Linux yet.
I’ve switched to Porteus (portable Slackware), happily using v3.0.1 LXDE on a USB stick.
http://forum.porteus.org/viewtopic.php?f=35&t=3716
I can summarize the general state of affairs, if you’re interested. Systemd is the accepted solution most everywhere, and only increasing in scope. The BSDs are in a bit of a poor state because of the trouble in updating things like Gnome because of the dependence on systemd dbus apis and there is thus the project systembsd implementing those interfaces, though it’s still far from complete with regards to logind support. There’s also been a bunch of fighting in Debian and a fork Devuan that also has projects like ConsoleKit2 continuing the development of that as well as LoginKit which implements the logind api on top of ConsoleKit2. There is certainly a way moving forward from systemd without missing out on the new development of projects, but it’ll still be a while before it fully matures.
I’m still working a bit on a Linux distro of my own, which I’ve said a little about here: http://bwackninja.github.io/2014/11/16/motivations-and-blogging.html and am avoiding systemd as well because it has become a bit too big for me to be comfortable with and hides too much of how the your system works. Progress is slow, but I’ve hit some a couple milestones and it’s a self-hosting environment. Eventually, I plan to port everything to a DragonflyBSD kernel, but that’s also waiting on support for my graphics card as well as evdev support.
Also, hi everyone! Hopefully I can manage to keep SpaceFM running, but there haven’t been too many new issues reported, and nothing major (other than the glib 2.41 issue).
Thanks for the general update. The BSDs do have a chain around their neck because of their traditional use of Gnome, and I haven’t yet ruled out the possibility that the systemd virus will infect them too – a lot of influence is pushing systemd hard. At any rate it’s nice to see you’ve kept an open and functional mind wrt to systemd, as I was wondering what your latest input on it would be, since you know it so well. It’s sad for me to see what’s happening to Linux, me being admittedly old school, but perhaps some good growth will come out of it. I think mainstream Linux and maybe even BSDs are ultimately sunk. They’ll go on in name/brand, but it won’t be the open/KISS/UNIX ecosystem it was. That’s why once I jump ship, I want to jump onto something progressive yet with a reasonable future. As you say, maturity will take time, just as it did for Linux. For now it’s a tough bet to make. Between hearing from you and OmegaPhil, I’m glad I’ve stayed out of the game for the last few months. :) He seems to be heading toward Devuan.
Congrats on the distro! That’s a remarkable project, and I like your willingness for novel choices. You’ve gained a lot of background in that with LFS, and I like your attitude toward it. That’s the only way to approach it, as a very long-term investment. We’re certainly ready for some new home-grown approaches.
Glad to know that you’re well!
I’ve been using sid without systemd on my laptop without any problems.
Debian community is a problematic space, but the distro is not yet fully rotten. Reproducible builds is ultimately one of the great things that they’re making progress on.
Been considering migrating to Gentoo, but I’m wary of maintenance overhead.
Looking forward to seeing Devuan+reproducible builds.
Also, I think Qt and KDE people are making progress in making their tools modular and ultimately becoming an even stronger contender to replace GTK and associated Gnome tools and libraries.
Good luck!
It is still possible to use Debian Stable without systemd by installing “systemd must die” by the Debian developer wookie:
deb http://wookware.org/software/repo/ sid main
There is also http://alpinelinux.org which uses OpenRC for init & which also does not have pulseaudio (although it is in the unmaintained repo if you really need it – ‘apulse’ is a better solution). It also has a grsecurity kernel as standard.
Installing an anti-systemd package to prevent systemd from installing seems a bit like silly/dramatic overkill, doesn’t it? :)
http://wiki.linuxquestions.org/wiki/Prevent_systemd_installation
You can simply directly configure apt to not install systemd.
I don’t know what the anti-systemd package does, but it’s not as simple as configuring apt. eg cups now depends on libsystemd0. If you blacklist systemd stuff it doesn’t go well – I just tried it again. Unless Devuan works it out, I think Debian is lost.
Alpine looks interesting, I might give that a spin. Due to upstream deps on systemd, I think any distro will have a hard time avoiding it, but maybe with shims they can do it. And as I noted below, FreeBSD looks like it’s going that way too in their Next Ten Years plan, probably sooner.
As I’ve said before, I think we’re in a position of no-good-OS, nothing genuinely free of corporate/govt control. And if you think systemd is open, try auditing the source and keeping up with the changes. It’s effectively closed and so complex that it’s unauditable, by design, probably with manufacturer binary components next on the agenda, as happened with X.
A few Arch users went to the trouble of documenting Arch-without-systemd:
https://wiki.archlinux.org/index.php/Init
Though I know you’ve used Arch, and of any distro Arch embraces systemd the most, so …
Manjaro has a community spin using OpenRC, latest ISO is 2 days old.
https://forum.manjaro.org/index.php?board=50.0
https://wiki.manjaro.org/index.php?title=Openrc
http://sourceforge.net/projects/manjaro-openrc/files/upcoming/xfce-openbox/
You can see FreeBSD lead developers acclimating their users to systemd here already, and trying to turn it too in a mobile thing. Even Lennart Poettering gave it his ‘okie dokie’. Obviously all these allegedly free OSes are deeply controlled, and there is a lot of pressure to push people into Google/MS, and to poison what remains with super-viruses like systemd. I wonder what OpenBSD will do – maybe the same story there.
It’s good to see a post from you again.
I became concerned about systemd, ironically, right after the Debian Technical Committee made its decision, which at the time I thought was long past due and the correct decision. mea culpa.
It was about that time I read a link to this blog and not long afterward you announced your hiatus and while I was disappointed that one of the voices that had provided much clarity to me about several subjects inclusive of GTK3, GNOME, Systemd, et. al. was now going silent. However, I respect your decision as these sorts of battles are wearisome.
I am hopeful that a new direction is taking shape. There are enough people concerned that I think we will have some viable alternatives in the near future. I have been following Devuan (pronounced dev-one) since its inception and some good things are happening with some fresh ideas. In some ways it feels a bit like 20 years ago again. I hope that seeing an alternative path here on the Linux side will give our *BSD friends the confidence to put on the brakes and continue on as before and not succumb to the invasive culture that is freedesktop.org.
I help maintain a library for communicating with amateur radio hardware. It now concerns me that the next scope of the system that the systemd developers will subsume is access to the system hardware. In other words, I suspect that they will deem it “dangerous” and “deprecated” for applications to access a serial port directly and will use udev to obfuscate things to the point that such access will need to go through the morass that is systemd at this point (maybe I should not be giving them ideas). I fear that project knows no boundaries and certainly has no clearly defined limit of scope (except, perhaps, the universe) nor stable API. Apparently, the systemd project places no value on backward compatibility as snippets from its developers suggest.
It should be obvious to anyone paying attention that systemd and its related projects are a grave threat to those of us who value freedom in our computing software. Also, it is disappointing that whenever some project chooses to depend on systemd that support for previous methods is simultaneously dropped. I hope I am wrong about that point as it seems such functionality should be able to coexist. I really have no problem if some desktop component adds support for systemd so long as support for non-systemd methods isn’t removed at the same time. After all, the C library is full of deprecated functions that remain for backward compatibility but with advisories that new code should not use those functions, etc. That is the way responsible developers would write system level software.
I hope you will consider continuing to offer your opinions on these matters. The light needs to be continuously shown on these all encompassing projects that seek to subsume the entire Linux stack as we have known it. I, for one, will be walking away from them.
Nate,
You have a pretty clear view there. For my part, it’s not so much that I’ve gone silent, as that I’ve said what I had to say. I unraveled it all to my satisfaction and shared my results. I’m not one to keep repeating it (too much). I’d rather devote my energies to something productive.
Above I called systemd a super-virus, and that should say a lot. It’s important to put all of this into context. We have to understand that wars today are largely fought in cyberspace, and military types view this as their battleground. Social media and its hi-tech manipulation, economic and industrial espionage, etc. – it all revolves around software and hardware. So various entities seek supremacy, kill switches, intrusion, manipulation, etc. Control.
I see systemd as a very carefully engineered method (social and technological) for owning Linux/BSD, free and open software in general, and specifically centralizing control of development. Freedesktop is just a front for Red Hat, which is just a front for the military, and their hands are very deep in Linux and BSD (and Microsoft and Google and Apple and Intel, etc). Anyone who doesn’t grasp that is not connecting with reality. You need to look at who is writing this stuff, as well as the people behind the people. It’s very carefully designed to do what it’s doing, and that includes the social war that has accompanied it. This is just the latest attack on Linux, depending on a growing group of users who are not genuinely freedom/activist oriented. They are easily deceived.
The people behind systemd definitely don’t honor things like backwards compatibility. Rather, their intention is to undermine the distributed nature of Linux and destroy the ecosystem. They don’t want what we want – a stable ecosystem that allows lots of people to participate freely. They want a complex system that is continuously changing (see Microsoft), bug-ridden, where the only ones who can keep up with it are those in charge of writing the latest version of it. Secrets (backdoors, etc) are easy to hide, brute force works, developer resources are exhausted through continuous change and breakage, and options are limited. How do you round up and contain something as diverse as Linux? systemd.
Unfortunately people in the military think they’re working for ‘their side’, but they can’t see that the only thing which can really set people free of tyranny is distribution of power and resources, rather than centralization. This we learned in Linux long ago – that’s why Linux is being attacked to its core. Nor is it just computers or software – everything in the world, from media to politics to medicine to food to water, is under the same hammer, the same push for centralized, dictatorial control. This all dates back to WWII and beyond (in case you’re wondering how ugly it can get).
Every generation seems to need to learn this anew the hard way. In the 80s, geeks certainly didn’t trust corporations or governments! Today you see many happy little Nazis, for lack of a better word. They don’t really want to mass murder, but they don’t understand the implications of what they’re supporting, they just follow the money offered and the latest shiny bait. They’re digging their own graves. One day they will see where it has brought them, and humanity will have a mighty deep hole to climb out of, if it survives at all.
So to really understand this, you get into some very serious stuff – basically war. For us, it merely means our distros and OSes are being consumed by being engineered away from us, turned into tools of centralized control of people, rather than creative tools for people. Forks like Devuan are already coming under sophisticated attacks from trolls (a well-organized campaign – this isn’t just a few fans of systemd, these are people (or AI) very determined to eliminate any form of software freedom. If they merely wanted systemd for themselves, they’ve already got that – all the major distros are onboard.) Devuan may find themselves under attack by “hackers” too – always large-scale attacks are blamed on some little-known hacker group, never the big three-letter agencies, as if they never hack or disrupt anything. It’s more like organized crime that’s running things – government/military has become a tool, very corrupt. Social degeneration. It’s not one nation that is responsible, they’re just pawns too, often witless.
Typically with things like this, they hammer it down so that most of the mainstream stuff is under control (and Linux has become mainstream). Usually they don’t try to completely eradicate all the little offshoots, so long as they don’t feel particularly threatened by them. For example, Devuan users are still largely using Debian, which is seriously pwned at deep levels. So even if there are a few innovative people being creative, it doesn’t threaten the mainstream trends. This is about the only freedom we can look forward to, a few homegrown creative projects to enjoy. As long as it doesn’t change the world. But you can kiss mainstream Linux goodbye as anything usable in years to come, and BSD is essentially the same, just slower and less mainstream.
As far as your OS being pwned, I think it’s important to at least acknowledge it honestly, rather than fooling yourself. It’s all tainted and broken from a strong security standpoint, and growing worse. Anyone who tells you otherwise is a fool or is trying to convince you of a lie (they think ‘their side’ is winning and want you to remain naive). They’ll probably tell you your cell calls are private or sell you snake oil to pretend-encrypt them too. We’re in the middle of a battleground, even if we’re just trying to share some private email and develop a ham radio app. It’s always the innocents who pay the real toll of war. And much of the world is sliding toward a very nasty brand of tyranny. It’s not just Linux.
I have been following your posts related to Systemd and have switched to Gentoo. I did check out OpenBSD but my preference lies in Linux not the BSDs. I first came across your post related to your switch from Arch Linux to Aptosid and was a bit disappointed of your hiatus period as I wanted to know more about why you thought systemd was / is a bad idea.
I do not have the technical skill in understanding why systemd is “bad” but I have come the conclusion that Lennart Poettering is rather arrogant in the way he has handled the implementation of systemd. This arrogance spills out in the form of “putting” down the Distros that have decided not to move ahead with systemd.. yet..
You described Systemd as a “virus”. Please can you give working examples of why you think that this the case. If it is a virus then why is it that Linus does not mind it? This is not meant to be a challenge but a simple question of understanding by using examples..
For me linux is all about “There’s more than one way to skin a cat.” approach. I mention this as there is a method to read systemd’s binary logging to text logging using syslog-ng. So if this is the case, savvy users will be able to navigate through the complexities of systemd and find ways to exploit it.
I am a user of the Jolla phone which uses Sailfish OS.. this implements and used systemd.. why would the developers of this use systemd over openrc or other init system? What I am saying is that systemd is an essential in mobile technology due to its speed.. Please feel free to correct me.
I am excited about Devuan and personally I feel that it is a project that will succeed over time. I also sense a trend that OpenRC is being adopted as an alternative to systemd and that there are enough people who are passionate about linux because of the level of control that linux allows the user who will find ways to graft in alternative init daemons for the future..
I look forward to your reply and to learn more from you..
Hi. Briefly, I view systemd as a form of carefully engineered bait – it has some features that are welcomed in themselves, so some of the technical debate is beside the point, but it takes Linux more toward being a closed, Windows-like OS in various ways, and will grow moreso that way. And the people introducing it aren’t just independent devs, it’s a Red Hat product. Red Hat, a corporation with ties to the military (their largest customer) and large corporations like Google, has dictatorial control over most key components in Linux (xorg, udev, udisks, gnome, gtk, the *kits, systemd… the list is long – everyone runs Red Hat), and they’re taking it in directions not in tune with its origins. Linus has had much friction with them over this. Linux is not an open and free OS, nor is it secure – that’s nonsense to anyone who looks at who really develops most of it and why. Nor can just anyone participate in that development. It is centrally controlled by Red Hat, and when they put on the pressure for systemd to be adopted, all the major distros obeyed. (All of these statements are based in facts, so don’t be swayed by people who just say “that’s ridiculous” and don’t provide facts.)
I think avoiding systemd and its libs on principle is a great way to avoid being dragged along with much of this, as that pivotal decision will force you more into the wilderness, and will help avoid the encroaching changes to mainstream Linux. But being less mainstream also limits options, just like using Linux once did. I don’t see any really good choices, but there are a few like Alpine etc that limit the components more reasonably.
I’ve covered the details quite a bit in this blog, so I’ll refer you to some articles below, and the comments contain more discussion. Not all of these refer to systemd specifically, but all of them relate.
Hey IG! It’s great to see you’ve not fallen off the face of the planet!
I switched to Gentoo in 2012 after the Arch debacle, as you probably know. I’ve been able to blacklist systemd and anything else I deem malignant (*kit, avahi, anything Poettering’s had a hand in), and I happily run on OpenRC and eudev. Maintenance isn’t that bad, and the Gentoo folks seem rather dead-set on keeping OpenRC their default, which is great news for us. Funtoo has also rejected systemd, iirc.
The decision made by Debian’s tech committee has resulted in a few devs leaving the distro; maybe we’ll see them fork Debian and retain its more valuable philosophies! It would be nice for people to have a binary distro to use that’s free of systemd. To be frank, I’d like to see a more forceful distro that flat-out will not support packages for software that’s deemed dangerous to user liberties, and anything that relies on them like GNOME, etc. Then again, I’m somewhat of an idealist so I enjoy seeing strong stances taken on things.
I’m in the process of becoming a Gentoo developer and hope to aid in whatever I can to keep choices alive. If I’m lucky, I’ll end up as maintainer of Fluxbox (which I’ve been proxy-maintaining), SpaceFM, and udevil. They’re three of my favorite packages and keeping them running well on Gentoo would be a small but important favor I could do for other Gentoo users.
If I ever get enough programming experience (currently in chapter 6 of K&R2, close to finishing it!), I’d like to devote efforts to projects that are against systemd. OpenRC, last I heard, has plans to support most of systemd’s unit file spec (as a compatibility option, not replacing the bash scripts), and used cgroups since before systemd was around, so I see OpenRC taking systemd’s biggest pro (the unit file) away from it and offering the same sane, dependency-based daemon management that it has for quite a while. I’ll need to browse its code more to make a better-informed decision. Before I jumped from Arch, I used runit, so that’s a possible contender as well.
A guy also wrote a tongue-in-cheek blog entry about creating a socket-activated init with cgroups. He did in in Ruby, in under 500 lines of code! If the biggest technical benefits of systemd can be copied and done better, in a smaller, more performant package, it can be publicly lambasted for being bloatware and it could influence the mindshare. I mean, cgroups and socket activation aren’t anything new. It’s a somewhat clever approach that’s worth copying, but an init system should be anything except ambitious.
I’ve abstained from talking about negative things on my blog, since it was impacting my personal life and I felt bad every time I visited it. I’ve relaunched it as a statically-generated site and focus on the things I think are beneficial.
If you and/or others ever have plans to start a radical new OS that’s outside the clutches of Red Hat, count me in. I’m pretty decent with documentation (working on a replacement for Fluxbox’s currently) and I’m getting better with programming each week.
Glad to you see you again. :)
Hey Daniel, good to see you here. I really like your thinking on this. I too don’t like all the negative commentary I’ve gotten into, but it was an adjustment to be made, and this is a challenging time for the open-source community. I also had the thought that other init systems could read systemd files. What we could really use is a packaging standard for daemons that is init-solution-independent – imagine if daemons were merely started and took care of restarting themselves, etc. But Red Hat has huge influence in these things, and they are making compability and flexibility as difficult as they can. There are technical options, but to have them adopted and used by something like a large distro requires a certain social organization. There seems to be a lot of pressure to dissolve such efforts, even the pressures and trolls Devuan is encountering are an example. I think we’re being played pretty well.
For now I’ve decided to focus on my projects more, and less on the OS scene, kind of letting it simmer and settle. I suspect enough annoyed people will fashion some alternatives. Just as they’ve made systemd central to their plan, I’ve made avoiding it pivotal to mine. Glad to hear of your success with Gentoo – it’s on my list. You sound like you’re getting into some good dev work! And I’m glad to hear SpaceFM’s libudev code is apparently working via eudev – I haven’t looked into the specifics there.
Init-agnostic daemons wouldn’t be too difficult, I don’t think. As you said
though, it comes down to getting the mindshare. I’m picturing daemons working
similarly to conky in that they would respond to signals. For instance, conky
will run in the background if you tell it to, and when it detects its config
file has changed (I have no clue how that’s done), it’ll reload without the user
having to intervene. It’s not even a daemon, either! If a userspace program can
do that sort of rich, responsive behavior, I see no reason for daemons to not at
least look into it. Granted, part of the advantages of init systems is that you
can daemonize everything, but I honestly don’t see the downside of drafting some
sort of “best practices” for creating daemon applications. A few proofs of
concept could be thrown together for common daemon tasks to show others how they
can alter their existing projects to daemonize on their own. I guess the issue
then is standardizing on interfaces, like runtime flags, et al, which I assume
SysV was drafted to alleviate: bash scripts that all assumed a ${1} of
start|stop|restart, etc.
The implications of daemons that can police themselves are interesting, though.
To me, it seems like it could simplify an init system, beyond that of SysV even.
Init systems could then focus exclusively on bringing the system up and down,
and provide a few controls for managing process trees (should the user give a
shit about cgroups on Linux systems). Granted, you’d lose out on daemons gaining
dependency resolution, but a single config file could make that possible. But
all of this is just conjecture: I’ve not played with creating a daemon,
personally, so I don’t have the experience necessary to go into the fun details.
But I think it’s worth exploring.
I don’t blame you for choosing to focus more on your own projects. GNU/Linux is
becoming more of a political battleground between companies trying to wrest
control from each other and some nebulous “standardization”, which will kill the
spirit of FOSS and drive away anyone competent and conscientious. In my current
position, however, I like too many pieces of software that have no dog in the
init wars that I’d rather do what I can in the distros that are still sane than
to drift away and give up. Maybe once I get a laptop again, I’ll feel like I can
mess around with other OSes and maybe migrate away or help others build a new
OS. But for now, my computer still has many jobs to carry out, and I’m not ready
to give up. As long as there are people who don’t want to run systemd, there
will be a way to run GNU/Linux without it. Assuming it doesn’t become part of
the kernel itself, of course. But I think Linus will have an aneurysm the day
that happens.
Eudev is — from what I can tell, anyway — API-compatible with udev. They stay
up to date with most udev updates, but choose to remain init-agnostic and fix
bugs that upstream misses. I’ve not played around with mdev or other device
managers, but they could be another alternative to the monstrosity that
systemd/udev have become.
No guarantees on the dev work, but my mentor seems somewhat satisfied with my
test and I’m hoping to move to the next step soon. I’ve been doing my best to
respond to any bugs on packages I proxy maintain and thoroughly test packages
whenever possible. If becoming a Gentoo dev proves to be too much trouble, I may
pursue a Gentoo-derivative or join another group of developers. It’s hard to
say, with so many interesting alternatives out there like stali (in
development), morpheus (likewise), Plan9, GoboLinux (seems almost dead; and it’s
a shame since the FS heirarchy is so off-beat), and so on. If all else fails, I
and other people could congregate and develop our own Gentoo-derivative.
Gentoo’s best used as a meta-distribution as it is, so it’s perfect for
branching off.
I wasn’t around for it, but the early days of FOSS seem like they were awesome.
So many hackers exchanging code, discussing things, and building tools they
legitimately thought benefitted the world, free of corporate influence.
Community-owned software, essentially. It’s by no means perfect, but I don’t see
corporate involvement being good for something that’s built on solid social
values like independence, transparency, cooperation, and liberty. It’s a shame
to see something become tainted by money and power.
Devuan seems pretty interesting. I’ll need to give them a look and figure out
what they’re all about. Do you know if any of the main devs are ex-Debian devs?
That would be fitting. :)
Funtoo, funtoo, funtoo. It will never have systemd.
Nice to read from you.
On my home machine i switched to CRUX. Simple and easy to use distro with a ports based system.
And I usually use software recommended by the suckless guys to keep things minimal and text-based. So far I’m very happy with it. But you need to be patient, if you don’t have a multicore system to bake software.
Puppy Linux doesn’t use systemd, and it looks to me like the Puppy Linux community overall frowns on systemd.
Here’s a Puppy Linux Discussion Forum thread titled “boycott systemd”, currently 14 pages long: http://murga-linux.com/puppy/viewtopic.php?t=93586
Lots of info and suggestions for alternatives are there.
Barry Kauler, the creator of Puppy Linux, doesn’t like systemd either: http://bkhome.org/blog2/?viewDetailed=00168
Best wishes,
Apollia
I recently found a more recent blog post (from Nov. 14, 2014) about systemd from the blog of Barry Kauler, the creator of Puppy Linux:
http://bkhome.org/news/?viewDetailed=00114
In that more recent post, he sounds like he’s more informed, and much more firmly against systemd.
He wrote: ‘Puppy does rate as one of the “major” Linux distributions, in terms of popularity, so should always be listed as one of the holdouts.’
I’ve been using various forms of Puppy Linux since Feb. 2011, so I’m very glad to read that.
Puppy Linux is currently #14 on DistroWatch’s page hit ranking: http://distrowatch.com/
SpaceFM v0.8.3 is actually preinstalled in the Puppy distro I recently switched to – Lighthouse 64 Puppy Linux 6.02 Beta 2. http://www.lhpup.org/release-lhp.htm#602
I haven’t used SpaceFM very much yet, but, it looks promising, and judging by how fascinating and educational this blog is, I’m guessing SpaceFM is probably brilliant and very useful. At some point, I’ll try to upgrade to the newest version of SpaceFM and try it out more carefully and thoroughly than I have so far.
I like and am very accustomed to Puppy’s usual default, Rox-Filer, but sometimes I have needed something more powerful. Perhaps SpaceFM will turn out to be exactly what I need.
Thank you so much for your fascinating blog, SpaceFM, and other work which I’m not yet familiar with!
Best wishes,
Apollia
Hi, I’m using Voyager X, it’s Xfce 4.12 on Ubuntu 14.04 LTS. When the Devuan guys have a distribution, I’ll jump over. I believe they will be using Xfce 4.12 as well. Here is a nice clip of Voyager X here: https://www.youtube.com/watch?v=k9z43RViHYU