I can understand that it is hard for you to following gezley confused tirades and you lose context when I refere to an earlyer post from him ;)

You're confused, not me. The quote was from Gazl, not gezley.

This topic MUST BE closed ASAP !!! period !! next !

What's it got to do with you? Get back to your BLEEDIN-EDGE MINT DEBIAN. :)

http://without-systemd.org/wiki/index.php/Main_Page

http://www.gabordemooij.com/escape_from_systemd

http://www.reddit.com/r/linux/commen..._to_not_using/

The parallel is that just like Microsoft had the advantage of control of the undocumeted APis, systemd devs have the advantage of control over the interface between components. Of course, this wouldn't be so much of an issue if systemd didn't internalise so much stuff unnecessarily.

There seems to be some confusion in this thread about public/private interfaces in programming. Depending somewhat on language/paradigm choice, public and private objects/functions/classes etc. roughly mean that they are part of internal workings of a certain part of the program and because they are likely to change a lot with development aren't part of the API (relatively static interface for other code to talk to). Its not because they are trying to hide some special features for themselves, its because its likely to break if its changed internally without corresponding changes to the code utilizing it. If you are developing on the project itself you can adjust both an internal function and the code calling it in tandem to keep things working, if it was part of the stable API you wouldn't want to change it in certain ways since you know external code is already using it expecting very specific things to happen when using it. Using the existence of private classes etc as an argument against systemd makes no sense. Linux kernel has private functions also. You could call them if you really wanted, but there is a fair chance they might not keep working they you want in the future.

Yes, but the kernel private functions are reserved functions aimed at developers of the kernel which does not interface with the userland components. These actually are not completely undocumented interfaces though.

The fact of the MS-Windows API functions being undocumented and private were used to give Microsoft an unfair advantage over it's competition. These same style interfaces in systemd likewise are complete unknowns. For one, yes they are undocumented and private, but second, they are publicly available even though they are claimed to be unstable, yet why are they listed as unstable? If they are unstable, why are they unstable?

Just saying "Oh, don't use these because they're unstable because we said so," isn't being forthcoming with why? A good developer with any sense of responsibility would probably say, "the XYZ interface is undocumented because we don't want people using it yet. In tests, it presents a system instability which could trigger a cascade failure of services, and caused some software to issue behaviors similar to memory leaks, so please don't use it until we get it locked down better, or remove it."

The point is this regarding the future of GNU/Linux... If things get out of hand, yes, the BSDs will be probably the next stop in many people's future in UNIX. If GNU/Linux does get to a point where it can be more easily attacked by malware then it will be no better off than OSX and Windows are. In fact, OSX has one of the worst case issues with malware because, not just of the attitude that OSX is virus-free, but the malware can dig into hardware components like the batteries, and other firmware components. Even Windows isn't hit that nastily anymore.

an other desperate try to find something that does not exist
there is nothing secret, they say dont use it and if you have any questions about switches and IPC messages you can ask the developers and look in the code.


maleware FUD, just wild speculations, as usual no facts just bla bla bla :o

Completely agree, but that wasn't the argument I or Richard (who I was responding to) was making. If you go back to post #1600, you'll see that Richard and I were actually arguing that these interfaces should have been private classes/methods/interfaces (delete as applicable), not public ones. My comment with regard to the MS API situation was simply suggesting that this was another case where the public/private line had been blurred for tactical advantage.

classes/methods/interfaces (delete as applicable) = ""
do you now what the difference between command line switch, IPC message and API is?
well, who cares about such details when you can do such competent posts like #1600

You know what. If this is what this forum has come to, with assholes like you being argumentative for the sake of it then, F**k it! I'm done here.

s/forum/thread/

This forum needs sensible and knowledgeable members like you, GazL.
Same stuff is being repeated with increasing heat... about time to close this thread, IMO

The 'stability' of an interface does not necessarily mean the stability of the system -- by unstable I believe they just mean the interface may change in the future, not that it can induce crashes or unexpected behaviour...
I would say about 90% of the people posting in this thread (on both sides) can be included in this. Perhaps myself too. systemd -- ruining internet discussions since 2010.

FUD? How the fsck is THIS FUD?

http://www.extremetech.com/mobile/19...quipped-device

http://www.usatoday.com/story/tech/2...ware/18584589/

http://www.businessinsider.com/apple...lurker-2014-11

Stop with your bullshit already!

Is this the future of GNU/Linux you'd love to see or are you just braindead and blind, claiming ignorance? This is why GNU/Linux has been less prone to malware because the systems vary in kernel, library versions, etc. which makes it HARDER to target a wide range of systems!

It's been proven that when systems vary, it's harder to target things accurately because there's not one single point that can be effectively attacked, but guess what? That's changing. Now not only will the malware authors have a single target specification, but if they were to utilize an undocumented interface that could cause a problem, and these guys range anywhere from mindless script kiddies to 200 level IQ hackers, who know what they know and could utilize?

FUD? Hardly! Why don't you go ask a security analyst about how vulnerable a uniform system is? You might get the education you so plainly lack!