You don't understand the problem with people restricting other people's freedom of choice, and then calls them a**holes if they object?yzT wrote:I really don't understand what's the problem.
This is nothing more than appeal to popularityIf all major distros, Debian included, are either using or planning to use systemd, it can't be as bad as you say.
In that case stay silent.yzT wrote:I really don't understand what's the problem.
Then you either don't understand FOSS or have no clue how to read or use a search engine.yzT wrote:I really don't understand what's the problem.
I'm all for people getting paid, and all for making money of both Linux and GNU and other GPLed code. More power to them, for they are implementing Richard Stallman's vision and refuting Bill Gates and (more recently) Jaron Lanier's position. And, as edbarx always likes to point out, we have the source codeLinadian wrote:Poettering is a PAID employee of Redhat, a corporation who's interest is to make money off of Linux...
Linadian wrote:[...]and it seems, control the rest of Linux world with their code.
We might not like systemd and the wider system (e.g. kdbus) but it is *not* proprietary. Lennart and his chums are 100% transparent and posting their code, unlike, say, nvidia. Their code is all GPLed. Can we try to keep the technical structure argument separate from the licence argument please? Otherwise the systemd people can claim FUD. See my earlier posts about how the the narrative is being constructed.Linadian wrote:If Redhat and Poettering want a proprietary kernel, then let them build their own, not FORCE it on to the rest of us.
You didn't address how it wants to control almost everything on your system including networking and why they're insisting on binary logs. As for the money aspect, I don't care if Redsoft makes money off of GPL, good for them, but I should have the choice whether or not to have their code on my machine (or make it NOT handle networking, I find that very disturbing and alarming), that includes not being ingrained in a FOSS GPL kernel.keithpeter wrote:I'm all for people getting paid, and all for making money of both Linux and GNU and other GPLed code. More power to them, for they are implementing Richard Stallman's vision and refuting Bill Gates and (more recently) Jaron Lanier's position. And, as edbarx always likes to point out, we have the source codeLinadian wrote:Poettering is a PAID employee of Redhat, a corporation who's interest is to make money off of Linux...
Linadian wrote:[...]and it seems, control the rest of Linux world with their code.
Nope, I genuinely think Lennart is just trying to deliver stuff that works for Red Hat and others. I suspect that it is the lack of other companies contributing that has lead to the huge tidal field that the systemd group are exerting.
We might not like systemd and the wider system (e.g. kdbus) but it is *not* proprietary. Lennart and his chums are 100% transparent and posting their code, unlike, say, nvidia. Their code is all GPLed. Can we try to keep the technical structure argument separate from the licence argument please? Otherwise the systemd people can claim FUD. See my earlier posts about how the the narrative is being constructed.Linadian wrote:If Redhat and Poettering want a proprietary kernel, then let them build their own, not FORCE it on to the rest of us.
'Provide services for' rather than 'control' and I believe you can switch off the binary logs and install text based logging (I read that somehwere reliable - I think it was the CentOS forum - but I'm no sysadmin) but yes, I take your point on the mudballification. And also this thing where the systemd authors keep changing the api from version to version so as a byproduct it is hard for the *-shim authors to keep up, and it imposes extra overhead on the various upstream projects. These are valid structural and implementation complains in my (totally unqualified) opinion. I just wanted to box off the licence and the money side of it.Linadian wrote:...
You didn't address how it wants to control almost everything on your system including networking and why they're insisting on binary logs.
You don't have to have any of their code at all, and we have listed various ways that can be achieved. I'm not sure how Mr Torvalds will react when they try to get kdbus back in the kernel but it will be fun to watchLinadian wrote:As for the money aspect, I don't care if Redsoft makes money off of GPL, good for them, but I should have the choice whether or not to have their code on my machine (or make it NOT handle networking, I find that very disturbing and alarming), that includes not being ingrained in a FOSS GPL kernel.
(pardon me for the capital letters)Assuming you meant me.keithpeter wrote:
@twoflowers: GPL is stronger than 'open source' (i.e. BSD, Apache &c) in various ways. Ubuntu Unity Amazon search lens was just removing *one* binary package using apt-get and more recently a single button to click in config. Not sure how that example invalidates my confidence in the unruly mob that is the GNU/Linux world's ability to work round Lennart and his chums by using the source code for their own ends. I have no specific reason for using his first name, he is a very talented programmer and will probably get a really good job with Microsoft one day (and I'll celebrate when he does).
Perhaps this is the central issue? For Mr (Dr?) Poettering, a younger man, this is a career delivering value for his employer. Not a social mission. Times change.
Apologies for getting your user name wrong, I've edited the post.fruitofloom wrote: Assuming you meant me.
I for one really appreciate what you are doing, I really like Debian and will hate having to abandon Debhat. I'm waiting for the results of your very honourable endeavours with bated breath.edbarx wrote:With an open and free to edit and compile source, solutions can be found to avoid using systemd, but like anything this requires some time. In my case, I am trying to use sysvinit on SID even when systemd is pushed down my throat. I want to see what the results will be before I count myself as defeated.
I've got a laptop ready to test the results when you feel ready to release your packages. I'm a bit hazy about how your plan extends to logind and dbus and the bits needed by the various full-fat desktop systems, but if there is enough to sort out automounting of USB sticks, I can use the old-school desktop (see sig) with your init and achieve more than I can at present.edbarx wrote:My approach is to bypass what Debian DDs want to impose as an initialisation system. So, my packages will make no reference to any existing packages to avoid forced rejection by the package management system. If I succeed, my installation of Debian will still continue to assume systemd is the workhorse, but it will not be the case, as a disguised sysvinit will be in control.[...]
Still a few around. Have a look at Jaromil (https://jaromil.dyne.org/) and the goto10 group (http://goto10.org/ alas dormant but the artists themselves pop up now and again). As you like music, try SLUB (http://slub.org/ also 'resting' - having babies I suspect).fruitofloom wrote:I think i am close to the ChaosComputerClub too. (When i started with Linux i assumed it would be a crowd of anarchists, communists, activists ... Well: i seem to have been wrong. I will get over it ...).
That is exactly the point. Some people, VERY rarely, have the chutspa to stick with their values over money and livelihood. This is so rare in the entire human races existence that we have called these people, "heroes". We do not have any of those in the Linux world right now other than arguably, RMS. Right now at this exact moment in time, WE NEED ONE!keithpeter wrote: Perhaps this is the central issue? For Mr (Dr?) Poettering, a younger man, this is a career delivering value for his employer. Not a social mission. Times change.
A) All filesystem-based or filesystem-representable IPC mechanisms (notably pipes), can't be used because of a lack of a world-writable directory, where all processes can mkfifo/create the filesystem/socket representation of their IPC port (/dev/socket notwithstanding, which is used for system processes, e.g. rile, zygote, and their ilk).
B) None of the suggested mechanisms have the capability of "service location" which is required for Android. In UNIX, there's an RPC portmapper, and Android needs similar functionality. Enter: The ServiceManager, which can use binder to register as a context manager, to register/lookup service handles on the fly
C) There is an extensive need for serialization - be it intents, or other messages. Binder provides the parcel abstraction, which can be used for data marshaling by the Parcel.java.
D) SysV has other issues than Mr. Lambada's answer which are more paramount, notably race conditions, and lack of authorization.
E) Message queues and pipes can't pass descriptors. UNIX Domain sockets may, but can't be used because of (A) (again, unless you're root/system, like zygote, rild, installd..)
F) Binder is really lightweight, and has built-in authorization mechanisms. It also has nifty features like waking up the recipient process, as well as memory sharing, which the other mechanisms simply don't have. (and remember, no mmap(2), because of the file problem in (A) for named mappings).
and - let's not forget
G) Binder was started at Palm (ah, nostalgia) (q.v. OpenBinder). Ex-palmers got to Android, and brought their code in with them.
