Hello, On Sat, 09 Aug 2014, Jan Engelhardt wrote:

On Saturday 2014-08-09 09:27, David Haller wrote:

...

I endorse not getting hooked by systemd forever. Maybe it can be just an episode ...

The opensuse-factory mailing list cannot substitute for the advice of a medical professional (for instance, a qualified doctor/physician, nurse, pharmacist/chemist, and so on) -- including, but not limited to, your episodes.

I think you've missed your last years of drugs and therapy sessions. FYSWAC. -dnh -- There are, after all, only three main datatypes: a thingy, a pile of thingies, and a sparse pile of thingies. A thingy is simultaneously multiple more fundamental types until you collapse the wavefunction by using it in a way that requires a more concrete type. -- Peter Corlett -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Henne Vogelsang wrote:

Hey,

On 12.08.2014 06:00, David Haller wrote:

...

[ bickering ]

Guess I have to remind you two:

Here at openSUSE we value the respect for other persons and their contributions, for other opinions and beliefs. We listen to arguments and address problems in a constructive and open way. We believe that a diverse community based on mutual respect is the base for a creative and productive environment enabling the project to be truly successful. We don't tolerate social discrimination and aim at creating an environment where people feel accepted and safe from offense.

Henne

Henne, bit of a pipe dream, isn't it? I've followed this thread, and AFAICT, David was only responding in kind. That doesn't make it any better, but we frequently lack mutual respect, being constructive and addressing problems in an open way. IMHO. -- Per Jessen, Zürich (18.2°C) http://www.hostsuisse.com/ - virtual servers, made in Switzerland. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Henne Vogelsang wrote:

Hey,

On 12.08.2014 21:03, Per Jessen wrote:

...

Henne Vogelsang wrote:

...

On 12.08.2014 06:00, David Haller wrote:

...

[ bickering ]

Guess I have to remind you two:

Here at openSUSE we value the respect for other persons [...] Henne, bit of a pipe dream, isn't it?

You think? I don't think our Guiding Principles[1] are much of a pipe dream.

In as much as we frequently fail to achieve them, yes, I think they are pretty much a pipe dream. -- Per Jessen, Zürich (16.9°C) http://www.hostsuisse.com/ - virtual servers, hosted in Switzerland. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 08/13/2014 12:56 PM, Henne Vogelsang wrote:

and value very much, are with me on that.

- Toujours la Politesse .......................... regards -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

El 09/08/14 a las #4, Raymond Wooninck escribió: I don't think that this should close the door for alternatives And what alternatives will that be ? The only other workable alternative was upstart.. until canonical decided to adopt systemd instead. . -- Cristian "I don't know the key to success, but the key to failure is trying to please everybody." -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 08/09/2014 10:09 AM, Raymond Wooninck wrote:

Hi Robert,

On Saturday 09 August 2014 06:09:14 Robert Schweikert wrote:

...

Also as fellow user and contributor I have to say that continued and constant systemd bashing is not conducive to moving the distribution forward and making the openSUSE project a great project to participate in. Quiet frankly I am getting tired of it.

I fully agree with you that this bashing on systemd is counter productive, however I guess we should also see if from a different point. openSUSE is a community driven distribution, but as far as I remember the community itself never took the decision to switch to systemd. This was mainly driven by the maintainers of the core-system, which are part of SUSE.

I agree that there were a number of breakdowns in the way the transition and the decision process was handled. However those things cannot be reverted. What we can do is learn from what has happened and do better in the future.

And this might be the biggest issue and the cause for the systemd bashing by the community. One of the items mentioned on that website is that more and more programs are starting to have dependencies on systemd and that distro's have to make the switch to the systemd world. Whether they like it or not.

Debian discussed this with its community and based on the feedback it took a decision. As indicated I can't remember to have seen any discussion or questionnaire within the openSUSE community regarding the successor of the sysvinit environment. Of course we can sweep it under the umbrello of "those that do the work decide". But then those have also be prepared for the subsequent discussions that their decision could cause.

I have no problem with the subsequent discussion, but at some point it has to end. The discussion about systemd has now been going on for more than 3 years.

But it is indeed too late to revert back as that we are already depend too much on systemd and it would create a lot of havoc if we would remove it now. However I don't think that this should close the door for alternatives.

Well, we have the best tool in the world to work on alternatives and/or derivative distributions, OBS. I have not seen anyone step up and do the work necessary. However, I do have to admit that I am certainly not aware of every project in OBS. Later, Robert -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQEbBAEBAgAGBQJT50h9AAoJEE4FgL32d2UkL8UH+PF/BjHbGcDhI+3R0HYEtb9p HUaLCKrsum5jqTpPFP8I/05zfHJ3MT/jUUFVV78j8elOYKqz9uaAuuE5INDXHaZW d67z+Fo80OGV84l6FSa/FGuiV3tnyqUfswawLfbdKb7929DKAbasi0gsZ46D998u 1LtHoj5CIvSTmU05wEeK82xYm135kEfMXdU2ftv6rk9V/s0MNkFTQ3zvFSlGvH/t 6jdljNRS6ZnuAVe9Z1Gp2h4Rm6i0v4QXJbf3Mlnyjyu4QxiOZrCuSMbsLb/kRp3M u0rMRbJ19oklhWgHu8jN6118xcT3HhSg6DFfj3Y9kVH7Jokcu7WJ9qZwQD4s9Q== =2sGO -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Content-ID: On Sunday, 2014-08-10 at 14:40 +0200, Carlos E. R. wrote:

On Sunday, 2014-08-10 at 06:25 -0400, Robert Schweikert wrote:

...

I have no problem with the subsequent discussion, but at some point it has to end. The discussion about systemd has now been going on for more than 3 years.

I don't see why... :-)

If people are not happy about a decission, they will probably comment/complain/bitch (maybe for ever) and not just silently accept the decission. Maybe the majority does, but not all.

Oh, I forgot to say that the discussion is, of course, pointless. :-) - -- Cheers, Carlos E. R. (from 13.1 x86_64 "Bottle" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iEYEARECAAYFAlPnavMACgkQtTMYHG2NR9W1FgCdF1xq9eKyPO+MURrGr23S3YCT xKIAoIT/VrTiesVoHZzEqJigwbRkmotq =rrHT -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Sunday 10 of August 2014 06:25:01 Robert Schweikert wrote:

Well, we have the best tool in the world to work on alternatives and/or derivative distributions, OBS. I have not seen anyone step up and do the work necessary. However, I do have to admit that I am certainly not aware of every project in OBS.

Apparently you haven't tried that. If you did (as I did), you would find how many subtle - and sometimes not so subtle - changes have been done to packages not related to systemd at all which were not needed for systemd but only break non-systemd systems. In some cases, just build a package of "wrong" name gives you a hard error from OBS checks... Michal Kubeček -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Dne Po 11. srpna 2014 08:16:12, Stefan Seyfried napsal(a):

Am 11.08.2014 um 07:53 schrieb Michal Kubecek:

...

On Sunday 10 of August 2014 06:25:01 Robert Schweikert wrote:

...

Well, we have the best tool in the world to work on alternatives and/or derivative distributions, OBS. I have not seen anyone step up and do the work necessary. However, I do have to admit that I am certainly not aware of every project in OBS.

Apparently you haven't tried that. If you did (as I did), you would find how many subtle - and sometimes not so subtle - changes have been done to packages not related to systemd at all which were not needed for systemd but only break non-systemd systems. In some cases, just build a package of "wrong" name gives you a hard error from OBS checks...

Or actually "factory-maintainers" -- a bunch of people apparently who have decided they better maintain "my" project than I do -- accept submissions against "my" packages, which remove non-systemd stuff, which I kept in on purpose.

Which leads to me actually mostly abandoning the vdr project right now.

OT: can I just remove the factory-maintainers from the vdr project? I certainly have not added them, so removing them would most likely be futile...

Factory maintainers team work based on maitainers inactivity. You can avoid them doing any work just by reviewing SR#s or stating there that they need to be redone in comment section. Usually the changes are reviewed if they are >1 week older without any activity. In the review we only care for merging patches that work on openSUSE:Factory and to extend if it works for supported openSUSE which are 12.3+ (so systemd only). The blame for having merged SR#s you didn't want is only for you. We simply don't want people to sit and wait for 1 month + without any activity on their patches. It is severly demotivating. And we still have such as for some areas it is simply not obvious and maintainer does not care... And no you can't remove them from your project :) Cheers Tom

On Monday 2014-08-11 10:25, Stefan Seyfried wrote:

It's summer holiday time here in europe, and one week is barely inactivity.

It is, in the context of globalization, which is where the world (not just OBS) has unfortunately been put: Everything has to be done faster, delivered sooner, and in higher quality, and if you cannot do it and do it in time, someone else will be assigned. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Mon, 2014-08-11 at 10:33 +0200, Michal Kubecek wrote:

Not true.

SR 163517: created 2013-04-10, accepted 2013-04-12. SR 211422: created 2013-12-18, accepted 2013-12-22.

Michal, Neither of those SR's were accepted by members of the factory maintainers team. If you have an issue with the decisions made by those individual reviewers, I'd recommend you speak to them Regards, Richard Brown openSUSE Board Chairman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Mon, 11 Aug 2014 10:46:39 +0200 Tomáš Chvátal wrote:

Dne Po 11. srpna 2014 10:33:27, Michal Kubecek napsal(a):

...

On Monday 11 of August 2014 10:04:01 Tomáš Chvátal wrote:

...

Factory maintainers team work based on maitainers inactivity.

You can avoid them doing any work just by reviewing SR#s or stating there that they need to be redone in comment section.

Usually the changes are reviewed if they are >1 week older without any activity.

...

...

The blame for having merged SR#s you didn't want is only for you.

Not true.

SR 163517: created 2013-04-10, accepted 2013-04-12. SR 211422: created 2013-12-18, accepted 2013-12-22.

Michal Kubeček

Neither of these are factory maintainers, these are project mainainers, eg people having perms on the whole serve:database project.

For factory-maintainers it is now just me I think

--- % osc rq show 201553 Request: #201553 submit: home:scarabeus_iv/converseen@2 -> graphics Message: - There is no reason to hardcode libnames as they are detected properly by itself. State: accepted 2013-10-01T07:48:24 scarabeus_factory Comment: Forward. History: new 2013-10-01T07:19:51 scarabeus_iv --- 28 minutes?

- There is no reason to hardcode...

You can found a reason looking at 12.3 Build Log. -- WBR Kyrill

* Tomáš Chvátal [2014-08-11 17:18]:

Dne Po 11. srpna 2014 19:07:35, Kyrill Detinov napsal(a):

...

You can found a reason looking at 12.3 Build Log.

Ah yes, that is one of the craps that were forwareded to factory and I found out right away it was broken. So to keep it not broken on Factory I went right away and forwarded it everywhere to fix it asap. It was said here quite often that Factory has prefference. If you have to break devel project for older distros to have it running for factory, the goal is to keep it run in factory. Building devel project for older distros is just something that is nice bonus, not mandatory thing.

Actually in practice there are devel project maintainers who reject SRs if they break compatibility with SLE11 (and it's prehistoric rpm)... -- Guido Berhoerster -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Mon, 11 Aug 2014 17:18 +0200 Tomáš Chvátal wrote:

Also the code I removed is for 13.1+ so it could not EVER afect 12.3 build...

12.3 build was succeeded in 2013 and didn't need the code before. I don't remember well. Anyway, I _know_ what I do in "my" packages. -- WBR Kyrill

On Sat, Aug 9, 2014 at 4:09 PM, Raymond Wooninck wrote: [...]

I fully agree with you that this bashing on systemd is counter productive, however I guess we should also see if from a different point. openSUSE is a community driven distribution, but as far as I remember the community itself never took the decision to switch to systemd. This was mainly driven by the maintainers of the core-system, which are part of SUSE.

And the maintainers of the core system are also part of openSUSE. I don't think you are trying to make the case that people who contribute to openSUSE, be it during their work time, and/or outside of it and who happen to be employed by SUSE, are not part of the openSUSE community. So why that inflammatory statement with its underlying implicit claim that it was done by authority and outside of the realm of the community?

And this might be the biggest issue and the cause for the systemd bashing by the community. One of the items mentioned on that website is that more and

"might". Or not. I don't see any of that playing a role in boycottsystemd.org

more programs are starting to have dependencies on systemd and that distro's have to make the switch to the systemd world. Whether they like it or not.

And programs also had to have dependencies on the way SysV init works (exit codes, PID file handling, etc...), whether they liked it or not.

Debian discussed this with its community and based on the feedback it took a decision. As indicated I can't remember to have seen any discussion or questionnaire within the openSUSE community regarding the successor of the sysvinit environment. Of course we can sweep it under the umbrello of "those that do the work decide". But then those have also be prepared for the subsequent discussions that their decision could cause.

That is a fair statement, but a project like this is an evolving thing, and learns from mistakes, hopefully, and adapts. See it as natural selection: if the right decisions and changes are made in such a project for a majority of the people in it, it will thrive. If not, it will die. Mistakes are made all the time, in this project too, almost always with the best of intentions. Arguably though, if you recall the context at that time... 1. would it have helped to have an open questionnaire and go by majority of votes? 1.1. who would get to vote? who would have been seen as having sufficient expertise in the domain to have an informed opinion? 1.2. should a decision like that be taken by majority, or by qualified opinions? 2. the discussion was very much of a hypothetical, as systemd was still quite new, and it was mostly fear of things that might happen, or not happen, in the future 2.1. how many of the claims made then against systemd have proven to be right? and false? 2.2. is it doomsday and has it technically killed the distributions that use it? If you get to the core of it, at least from my personal opinion, it boils down to resistance to change. Some people want to keep it as it has always been, arguing that this automatically means stability and robustness, and predictability. (I don't believe that is true necessarily) Some people want things to change in order to improve, have new features that will allow better functioning. (personally I believe Linux has grown up and taken enough market share to go its own way and stop being held down by the burden of ancient UNIX compliance, stand on its own feet, just look at containers and docker -- but that is merely my personal opinion, given for contrast with the previous ones) I think that the pretty emotional component of the reactions and statements on the topic does reinforce the idea that it is mostly about that psychological aspect. What I mean to say is: I sincerely do not believe a vote would have provided any further guidance. That is not to say that the whole topic could have been approached in a more transparent manner. But transparency also requires responsible community members. Including not making inflammatory claims, play the vocal minority bully card, threatening to leave (*) or harm, or just downright lies and trolling, etc... (*) bye David. seriously. claiming that in a community driven project, the opinion or preference of a single person should trump everything else is ridiculous.

But it is indeed too late to revert back as that we are already depend too much on systemd and it would create a lot of havoc if we would remove it now. However I don't think that this should close the door for alternatives. As David indicated that he has been contributing a lot in the last years, maybe he is willing to work on the integration of a systemd alternative.

Contributing a lot in the last years does not give you the right to threaten to enforce your opinion upon a community. Watch the poisonous people video again for reference: https://www.youtube.com/watch?v=Q52kFL8zVoM Technical arguments may be valid though. But you can't just list issues with systemd and leave out: - that other solutions have issues of their own, especially SysV init - that systemd does open the door to a wealth of new very useful features, that are starting to be adopted more and more, after the transition phase from init - that the doomsday claims that were made when the discussion came up have proven to be wrong - that whatever the init system is, programs will have a dependency on it, at the very least to its requirement and behaviour, as it was the case for init too - that systemd is now clearly the actively developed solution for the majority of developers and projects, and is best option to go forward with - the many issues going with an alternative would bring -- even Canonical was forced to see the light on that point, despite their extreme Not Invented Here syndrome etc... For perspective. cheers, -- Pascal -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

These discussions and disagreements are part and parcel and nothing new. Before SuSE was born there was glibc5 vs glibc6. Others like KDE3 vs KDE4, pulseaudio, going away from using a root password in favour of "sudo su" and the user password -- many others along the way generated flames and heat until they eventually ran out of fuel and changed nothing. Regards Sid. On 10/08/14 13:22, Pascal Bleser wrote:

...

I fully agree with you that this bashing on systemd is counter productive, however I guess we should also see if from a different point. openSUSE is a community driven distribution, but as far as I remember the community itself never took the decision to switch to systemd. This was mainly driven by the maintainers of the core-system, which are part of SUSE. And the maintainers of the core system are also part of openSUSE. I don't think you are trying to make the case that people who contribute to openSUSE, be it during their work time, and/or outside of it and who happen to be employed by SUSE, are not part of the openSUSE community. So why that inflammatory statement with its underlying implicit claim that it was done by authority and outside of

On Sat, Aug 9, 2014 at 4:09 PM, Raymond Wooninck wrote: [...] the realm of the community?

...

And this might be the biggest issue and the cause for the systemd bashing by the community. One of the items mentioned on that website is that more and "might". Or not. I don't see any of that playing a role in boycottsystemd.org

...

more programs are starting to have dependencies on systemd and that distro's have to make the switch to the systemd world. Whether they like it or not. And programs also had to have dependencies on the way SysV init works (exit codes, PID file handling, etc...), whether they liked it or not.

...

Debian discussed this with its community and based on the feedback it took a decision. As indicated I can't remember to have seen any discussion or questionnaire within the openSUSE community regarding the successor of the sysvinit environment. Of course we can sweep it under the umbrello of "those that do the work decide". But then those have also be prepared for the subsequent discussions that their decision could cause. That is a fair statement, but a project like this is an evolving thing, and learns from mistakes, hopefully, and adapts. See it as natural selection: if the right decisions and changes are made in such a project for a majority of the people in it, it will thrive. If not, it will die. Mistakes are made all the time, in this project too, almost always with the best of intentions.

Arguably though, if you recall the context at that time... 1. would it have helped to have an open questionnaire and go by majority of votes? 1.1. who would get to vote? who would have been seen as having sufficient expertise in the domain to have an informed opinion? 1.2. should a decision like that be taken by majority, or by qualified opinions? 2. the discussion was very much of a hypothetical, as systemd was still quite new, and it was mostly fear of things that might happen, or not happen, in the future 2.1. how many of the claims made then against systemd have proven to be right? and false? 2.2. is it doomsday and has it technically killed the distributions that use it?

If you get to the core of it, at least from my personal opinion, it boils down to resistance to change. Some people want to keep it as it has always been, arguing that this automatically means stability and robustness, and predictability. (I don't believe that is true necessarily) Some people want things to change in order to improve, have new features that will allow better functioning. (personally I believe Linux has grown up and taken enough market share to go its own way and stop being held down by the burden of ancient UNIX compliance, stand on its own feet, just look at containers and docker -- but that is merely my personal opinion, given for contrast with the previous ones)

I think that the pretty emotional component of the reactions and statements on the topic does reinforce the idea that it is mostly about that psychological aspect.

What I mean to say is: I sincerely do not believe a vote would have provided any further guidance. That is not to say that the whole topic could have been approached in a more transparent manner. But transparency also requires responsible community members. Including not making inflammatory claims, play the vocal minority bully card, threatening to leave (*) or harm, or just downright lies and trolling, etc...

(*) bye David. seriously. claiming that in a community driven project, the opinion or preference of a single person should trump everything else is ridiculous.

...

But it is indeed too late to revert back as that we are already depend too much on systemd and it would create a lot of havoc if we would remove it now. However I don't think that this should close the door for alternatives. As David indicated that he has been contributing a lot in the last years, maybe he is willing to work on the integration of a systemd alternative. Contributing a lot in the last years does not give you the right to threaten to enforce your opinion upon a community. Watch the poisonous people video again for reference: https://www.youtube.com/watch?v=Q52kFL8zVoM

Technical arguments may be valid though. But you can't just list issues with systemd and leave out: - that other solutions have issues of their own, especially SysV init - that systemd does open the door to a wealth of new very useful features, that are starting to be adopted more and more, after the transition phase from init - that the doomsday claims that were made when the discussion came up have proven to be wrong - that whatever the init system is, programs will have a dependency on it, at the very least to its requirement and behaviour, as it was the case for init too - that systemd is now clearly the actively developed solution for the majority of developers and projects, and is best option to go forward with - the many issues going with an alternative would bring -- even Canonical was forced to see the light on that point, despite their extreme Not Invented Here syndrome etc...

For perspective.

cheers, -- Pascal

-- Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Senior Staff Specialist, Cricket Coach Microsoft Windows Free Zone - Linux used for all Computing Tasks -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 11/08/14 22:07, David Haller wrote:

Hello,

...

Before SuSE was born there was glibc5 vs glibc6. Others like KDE3 vs KDE4, pulseaudio, going away from using a root password in favour of "sudo su" and the user password *Yikes*!?!?! Is oS shipping a sudoers that allows that? *looks into

On Sun, 10 Aug 2014, Sid Boyce wrote: the 13.1 VM* *ARGH* Albeit with "targetpw".

-dnh

That's the default behaviour in Ubuntu to this day. If you need to do anything that needs root privileges, it's "sudo <command>" upon which it asks for the user password. To set root's password "sudo su" - enter user password then "passwd root". During install it does not prompt you for a root password. The default password for the default user ubuntu is temppwd "sudo su", give password "temppwd" and you have root privileges. It's said to be done that way so as not to confuse new users who by many comments on this list when it was discussed found remembering 2 passwords presented a great difficulty to them. So unless the admin changes the password for ubuntu or explicitly sets up a root password there is that large gaping hole. Regards Sid. -- Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Senior Staff Specialist, Cricket Coach Microsoft Windows Free Zone - Linux used for all Computing Tasks -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

For perspective.

cheers, -- Pascal

Thanks for sharing it Pascal ... A worth to read it. -- Bruno Friedmann Ioda-Net Sàrl www.ioda-net.ch openSUSE Member & Board GPG KEY : D5C9B751C4653227 irc: tigerfoot ~~~Don't take Life too serious. Nobody gets out alive anyway!~~~ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Mon, 2014-08-11 at 23:58 +0200, David Haller wrote:

If you step back and look at the big picture, the whole process of how systemd crept up as a "just a new init" up to what it's now (and what it may gobble up next), you may reevaluate those reactions ...

That is what makes me detest the thing. Do one thing and do it well flew right out the window. Systemd developers seem determined that systemd shall become lord and master of the entire universe. I suspect it does do some things well, it must, else it wouldn't have gotten off the ground, much less to the point where many seem to think "One daemon to bind them" (them being you in my view) is a good idea. I've not seen even ONE tangible improvement in my encounters with it, feel all left out, am just not seeing the "wonderfulness". -Mike -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

* Todd Rme [2014-08-12 11:03]:

On Mon, Aug 11, 2014 at 11:58 PM, David Haller wrote:

...

On Sun, 10 Aug 2014, Pascal Bleser wrote:

...

I think that the pretty emotional component of the reactions and statements on the topic does reinforce the idea that it is mostly about that psychological aspect.

If you step back and look at the big picture, the whole process of how systemd crept up as a "just a new init" up to what it's now (and what it may gobble up next), you may reevaluate those reactions ...

You, and other members of the anti-systemd crowd, make it sound like systemd developers held a gun to the heads of developers and forced them to join the systemd project, forced them to adopt sytemd in their distro, forced them to adopt systemd in the desktop environments.

Yes, that pretty much summarizes it. Since the systemd authors are closely aligned with the u* and *kit folks, basic functions of desktop environments are now totally broken (session tracking which is the basis managing authorization of hotplugging, auto-mounting, system-shutdown etc., system shutdown, suspend, hibernation and inhibition of systemd handling of power buttons and so on and so forth) unless they change the code and add systemd support and workarounds. So yes, DE authors face the choice of having their DE unusable on systemd distros or adopt its usage (unless they are able to compete with the raw engineering power of RedHat and create an alternative middle stack for Linux).

Projects voluntarily chose to join systemd because they thought it would benefit their project. Distros voluntarily chose to adopt systemd because they thought it made handling init easier compared to the alternatives. Desktop environments chose to adopt systemd because it provides features they think are useful that aren't available anywhere else.

Obviously you have no clue what you are talking about. -- Guido Berhoerster -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

+1. Am Dienstag, 12. August 2014, 11:39:25 schrieb Richard Brown:

Okay all, I think this thread has run its course

If there are any productive, practical, efforts underway, I suggest starting a new thread to discuss implementation details

But the discussion regarding the pros/cons of systemd, it's history, its adoption, who said what and whether they are right, provide no further benefit to anyone and can all be stopped now.

Regards,

Richard Brown openSUSE Chairman

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Tue, Aug 12, 2014 at 11:29 AM, Guido Berhoerster wrote:

* Todd Rme [2014-08-12 11:03]:

...

On Mon, Aug 11, 2014 at 11:58 PM, David Haller wrote:

...

On Sun, 10 Aug 2014, Pascal Bleser wrote:

...

I think that the pretty emotional component of the reactions and statements on the topic does reinforce the idea that it is mostly about that psychological aspect.

If you step back and look at the big picture, the whole process of how systemd crept up as a "just a new init" up to what it's now (and what it may gobble up next), you may reevaluate those reactions ...

You, and other members of the anti-systemd crowd, make it sound like systemd developers held a gun to the heads of developers and forced them to join the systemd project, forced them to adopt sytemd in their distro, forced them to adopt systemd in the desktop environments.

Yes, that pretty much summarizes it. Since the systemd authors are closely aligned with the u* and *kit folks, basic functions of desktop environments are now totally broken (session tracking which is the basis managing authorization of hotplugging, auto-mounting, system-shutdown etc., system shutdown, suspend, hibernation and inhibition of systemd handling of power buttons and so on and so forth) unless they change the code and add systemd support and workarounds. So yes, DE authors face the choice of having their DE unusable on systemd distros or adopt its usage (unless they are able to compete with the raw engineering power of RedHat and create an alternative middle stack for Linux).

That is not what I am talking about. For example, for Wayland, KDE developers plan to use systemd socket activation for plasma workspaces to replace the old KDE service bash script. They were not planning to support upstart since its version of socket activation was far too limited (and upstart is a moot point now anyway). That means that, when using Wayland, plasma workspaces will depend on systemd, not because systemd forced them to, but because it provides features they find useful. xfce already supports using systemd socket activation for similar purposes. Things like logind are used because none of the alternatives are maintained. And again, this ignores the fact that projects joined the systemd umbrella because they wanted to, sytemd developers didn't force them to.

...

Projects voluntarily chose to join systemd because they thought it would benefit their project. Distros voluntarily chose to adopt systemd because they thought it made handling init easier compared to the alternatives. Desktop environments chose to adopt systemd because it provides features they think are useful that aren't available anywhere else.

Obviously you have no clue what you are talking about.

Thank you for that detailed rebuttal. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Tue, Aug 12, 2014 at 12:43 PM, Guido Berhoerster wrote:

* Todd Rme [2014-08-12 12:19]:

...

On Tue, Aug 12, 2014 at 11:29 AM, Guido Berhoerster wrote:

...

* Todd Rme [2014-08-12 11:03]:

...

On Mon, Aug 11, 2014 at 11:58 PM, David Haller wrote:

...

On Sun, 10 Aug 2014, Pascal Bleser wrote:

...

I think that the pretty emotional component of the reactions and statements on the topic does reinforce the idea that it is mostly about that psychological aspect.

If you step back and look at the big picture, the whole process of how systemd crept up as a "just a new init" up to what it's now (and what it may gobble up next), you may reevaluate those reactions ...

You, and other members of the anti-systemd crowd, make it sound like systemd developers held a gun to the heads of developers and forced them to join the systemd project, forced them to adopt sytemd in their distro, forced them to adopt systemd in the desktop environments.

Yes, that pretty much summarizes it. Since the systemd authors are closely aligned with the u* and *kit folks, basic functions of desktop environments are now totally broken (session tracking which is the basis managing authorization of hotplugging, auto-mounting, system-shutdown etc., system shutdown, suspend, hibernation and inhibition of systemd handling of power buttons and so on and so forth) unless they change the code and add systemd support and workarounds. So yes, DE authors face the choice of having their DE unusable on systemd distros or adopt its usage (unless they are able to compete with the raw engineering power of RedHat and create an alternative middle stack for Linux).

That is not what I am talking about.

For example, for Wayland, KDE developers plan to use systemd socket activation for plasma workspaces to replace the old KDE service bash script. They were not planning to support upstart since its version of socket activation was far too limited (and upstart is a moot point now anyway). That means that, when using Wayland, plasma workspaces will depend on systemd, not because systemd forced them to, but because it provides features they find useful. xfce already supports using systemd socket activation for similar purposes.

That is incorrect, Xfce has no systemd socket activation support whatsoever. In fact upstream Xfce 4.10 is totally broken on systemd and requires lots of patches to get basic functionality working again. 4.12 will be the first release to optionally support systemd and restore this basic functionality but will not make use of socket activation or any other systemd-specific features. And I'm certain of that because I maintain/have written some of these patches.

I am sorry, you are correct. I was remembering incorrectly, it looks like it is Enlightenment that supports socket activation, not xfce. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Todd Rme wrote:

You, and other members of the anti-systemd crowd,...

Lets call it what it is... the pro-choice crowd, and the no-choice crowd, as that is probably the biggest sticking point with most people. Many people don't like dictatorships and being told what to do. It's not relevant what the features of the no-choice camp are, because historically we've never seen one what has had happy citizens. So please forgive our continued attempts to ask for choice. Freedom of choice, a basic freedom in a free society is one of those things that much constantly be fought for -- and it doesn't come overnight and isn't lost in 1 battle or not (though the 1984-authoritarians would love to perpetuate the myth that there can never be change).

Projects voluntarily chose to join systemd because they thought it would benefit their project.

Some did... some didn't -- those that didn't weren't given a choice...they were absorbed. Several on here have documented active sabotage to remove any possibility of choice in the upcoming release. This is an established fact. There is no argument needed.

Distros voluntarily chose to adopt systemd because they thought it made handling init easier compared to the alternatives. Desktop environments chose to adopt systemd because it provides features they think are useful that aren't available anywhere else.

Providing secure boot is a priority to keep linux in the game as a 'controllable platform' that can be used to supply content and compete with Apple and MS. That's NOT what many use linux for, but it is what those with big $$$ signs in their eyes primarily focus on.

As far as I can tell, you are essentially arguing that because you don't like systemd and don't find it useful, all the groups that do find it useful aren't allowed to rely on it. There is a simple solution: have something else that provides the features these projects are looking for.

---- I had 90% of the features I wanted, and the ability to implement ones I didn't have. With systemd, I won't have that ability, because there will be no place for alternatives.

But you aren't going to get very far trying to tell projects that they can't make use of features they find beneficial at all because you don't happen to like the software that provides those features.

I agree some features are beneficial, but I want the a la carte menu. It could be provided, but part of the systemd design is to eliminate possibilities of ala carte usage and partial use. What was the MS motto -- systemd sure seems to have it down pat: Embrace, Extend, Extinguish... Ms eventually got brought under control because it was a monopoly and used it's monopoly power to aborb the functions of more and more programs, putting alternitives out of business. Not that I think it likely anti-monopoly laws would apply here, but in the business world, those actions were found *unethical* and unfair against other alternatives. Many courts ruled this way. So if you can't see how the same actions in the software community wouldn't cause similar if not equal antipathy, you aren't trying.

Whether you find the features of systemd useful is irrelevant. As long as systemd and only systemd provides the features that projects are looking for, those projects are going to use systemd. The anti-systemd crowd needs to start putting in the work to provide alternatives, or systemd usage is just going to keep growing.

---- The alternatives were there. LP chose to absorb, extend and extinguish them. I still say system auditing would have provided all the process tracking in real time that he needed to provide a service manager -- he didn't need p1. But think he'd change that? being #1 pid gives hime power over the whole system in a way that can be exploited. It will be interesting when he wants to integrate systemd more into the kernel and take it over how Linus will respond.....If we are lucky he'll choose to go off and invent a whole knew OS, and linux can get back to doing what it does best -- provide freedom and choice. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Please, lets just end this conversation here I think all parties involved have had ample opportunity to contribute productive dialogue, and if this thread continues it seems destined to descend into further petty bickering and pointlessness As I started in my earlier mail, if there are productive, actionable, practical efforts born from this discussion, please use a new thread to discuss the implementation of that Otherwise, lets end this conversation now. It's long left the topic of this list (openSUSE Factory) and it's not producing anything of value to anyone. Regards, Richard Brown openSUSE Chairman On 12 August 2014 14:25, Todd Rme wrote:

On Tue, Aug 12, 2014 at 1:53 PM, Linda A. Walsh wrote:

...

Todd Rme wrote:

...

You, and other members of the anti-systemd crowd,...

Lets call it what it is... the pro-choice crowd, and the no-choice crowd, as that is probably the biggest sticking point with most people.

It is also "the crowd that is willing to do the work and the crowd that isn't". That is the biggest sticking point for people who don't oppose systemd.

...

So please forgive our continued attempts to ask for choice. Freedom of choice, a basic freedom in a free society is one of those things that much constantly be fought for -- and it doesn't come overnight and isn't lost in 1 battle or not (though the 1984-authoritarians would love to perpetuate the myth that there can never be change).

People who fought for choice made an effort to make choice possible. Just talking about choice doesn't change anything when no one is willing to put in the effort to implement choice.

...

...

Projects voluntarily chose to join systemd because they thought it would benefit their project.

Some did... some didn't -- those that didn't weren't given a choice...they were absorbed.

Please name any project that was forced to become part of the systemd umbrella against its developers' wishes.

...

...

As far as I can tell, you are essentially arguing that because you don't like systemd and don't find it useful, all the groups that do find it useful aren't allowed to rely on it. There is a simple solution: have something else that provides the features these projects are looking for.

----

I had 90% of the features I wanted, and the ability to implement ones I didn't have. With systemd, I won't have that ability, because there will be no place for alternatives.

Again, this isn't about what you find useful, it is about what the developers of major open-source projects find useful. They are the ones who decide whether relying on a particular piece of software -- systemd, cups, x11, wayland, llvm -- provides benefits to their project or not. And that depends to a large degree on whether the software provides features that the project developers think their project will benefit from.

...

...

But you aren't going to get very far trying to tell projects that they can't make use of features they find beneficial at all because you don't happen to like the software that provides those features.

----

I agree some features are beneficial, but I want the a la carte menu. It could be provided, but part of the systemd design is to eliminate possibilities of ala carte usage and partial use. What was the MS motto -- systemd sure seems to have it down pat: Embrace, Extend, Extinguish... Ms eventually got brought under control because it was a monopoly and used it's monopoly power to aborb the functions of more and more programs, putting alternitives out of business. Not that I think it likely anti-monopoly laws would apply here, but in the business world, those actions were found *unethical* and unfair against other alternatives. Many courts ruled this way. So if you can't see how the same actions in the software community wouldn't cause similar if not equal antipathy, you aren't trying.

Again, there is a simple solution to this: provide an alternative that projects actually want to use. All this talk about whether alternatives should exist or not will not accomplish anything. Unless the people who oppose systemd are willing to put in the work to actually provide alternatives, projects will continue to use systemd.

Free software was not built by people talking about how nice freedom it, it was built by people doing the work to provide the software. If no one is willing to do the work to provide the software, it won't exist. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Michal Kubecek wrote:

On Tuesday 12 of August 2014 14:25:58 Todd Rme wrote:

...

It is also "the crowd that is willing to do the work and the crowd that isn't". That is the biggest sticking point for people who don't oppose systemd.

Repeating this round and round doesn't make it true. We were in no need of a new init system. We had one and it worked well.

I hesitate joining this thread, but Michal has really hit the nail on the head here. My production systems are also slowly migrating towards systemd now, maybe in another couple of years we'll be done, but I have not yet seen _any_ benefit to the migration, only effort. It's a pity we weren't able to have this debate back when systemd was introduced. Yes, we argued back and forth, I was there, but I don't recall any definite community decision (other than that of the do-ocracy). Btw, Michal Kubeček deserves recognition for a clear, concise and non-inflammatory argument on this topic. I'm not so sure I can say the same for the pro-systemd camp, but I haven't read every pro-systemd posting. -- Per Jessen, Zürich (16.1°C) http://www.hostsuisse.com/ - dedicated server rental in Switzerland. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 2014-08-12 14:25 (GMT+0200) Todd Rme composed:

the crowd that is willing to do the work and the crowd that isn't".

That's an invalid dichotomy. It assumes everyone could do the work if willing. Not everyone is capable of learning the necessary programming skills. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Am 11.08.2014 um 07:49 schrieb Michal Kubecek:

The problem is it's not one package. In the name of systemd, the whole system is being redesigned and reworked in order to make the other options stop working. I've found numerous cases where changes have been done to other packages that don't help systemd and systemd systems at all but only break non-systemd systems. And then I keep reading these "if you want to provide an alternative, you are welcome to do so" appeals. It's frustrating and I really feel like I'm being mocked on purpose.

Even though I'm generally in favour of systemd, I have to agree with Michal on this: this makes it e.g. almost impossible for me to use openSUSE on embedded / ARM, because I want to use the same system on all my embedded boards and I have boards that cannot run systemd. Apart from that systemd has the problem that it just boots much too slow on embedded systems that are not from 2014 due to its huge memory and CPU footprint. So no openSUSE for me on ARM, fortunately the yocto project fits just fine for my needs, and it can easily be built even for x86_64 :-) But the premise "you can just maintain your own non-systemd distribution on OBS" is IMO a plain lie, unless you really want to imply everyone should build *everything* of the distribution in his $HOME project. -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Am 11.08.2014 um 10:58 schrieb Richard Brown:

On Mon, 2014-08-11 at 08:35 +0200, Stefan Seyfried wrote:

...

But the premise "you can just maintain your own non-systemd distribution on OBS" is IMO a plain lie, unless you really want to imply everyone should build *everything* of the distribution in his $HOME project.

Please Stefan, tone down the rhetoric

You can maintain your own "non-systemd distribution" in OBS just fine. But, for example with the vdr Project in OBS - that project is assigned as a Devel project. The purpose of a Devel project is to feed packages into Factory

Factory/openSUSE is a systemd distribution

Yes, and "init scripts will always work just fine with systemd". The fact is, that the old-style init script is the way for people to get their old configuration working during update. For new installation, the runvdr-extreme-systemd package is preferred. Now the old script was removed and a symlink rcvdr was added, which breaks the update path. Now I could have done this from the start, because keeping the update path is painful and a lot of work, just removing working things is easy. But I actually tried to care for my users. Something that is apparently totally out of scope nowadays.

So changes that add/provide native systemd support are *good* and *welcome* for Devel projects.. square peg, square hole

Yes. Exactly. That's why I kept the init script *as an option, it was even in its own subpackage*, so that users expecting square pegs (old style /etc/sysconfig configuration) would get a working setup. Now I would not have objected, would the SR have also included a converter from /etc/sysconfig/vdr to /etc/runvdr.conf, but that is not the case AFAICS.

If you want a project in OBS that does not feed into Factory, and isn't co-maintained by the factory-maintainers group, then don't have it as a Devel repo.

So this means I have to file a drop request for VDR from Factory before I go on two weeks vacation next week? -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Richard Brown schrieb:

On 11 August 2014 11:11, Stefan Seyfried wrote:

...

The fact is, that the old-style init script is the way for people to get their old configuration working during update.

For new installation, the runvdr-extreme-systemd package is preferred.

Now the old script was removed and a symlink rcvdr was added, which breaks the update path.

Now I could have done this from the start, because keeping the update path is painful and a lot of work, just removing working things is easy.

But I actually tried to care for my users. Something that is apparently totally out of scope nowadays.

...

So changes that add/provide native systemd support are *good* and *welcome* for Devel projects.. square peg, square hole

Yes. Exactly. That's why I kept the init script *as an option, it was even in its own subpackage*, so that users expecting square pegs (old style /etc/sysconfig configuration) would get a working setup.

Now I would not have objected, would the SR have also included a converter from /etc/sysconfig/vdr to /etc/runvdr.conf, but that is not the case AFAICS.

...

If you want a project in OBS that does not feed into Factory, and isn't co-maintained by the factory-maintainers group, then don't have it as a Devel repo.

So this means I have to file a drop request for VDR from Factory before I go on two weeks vacation next week?

Or alternatively, why don't you reach out to that community member who made the changes to vdr (not the factory-maintainer who accepted it after a week of no activity) and work with them to implement some of the options you propose above?

There is a great opportunity here to work with someone new who is also clearly interested in 'your' vdr package

Instead of complaining that a factory-maintainer accepted an SR which was left idle for 7 days, you have an opportunity to work with someone new who potentially could help you make the vdr package better for all scenarios (systemd, non-systemd, upgrades, fresh installs)

VDR is a rather special beast. Can we please stop here and agree that a) sometimes submit requests are made with the best intentions, look good from the outside but are still suboptimal for the package which only the maintainer can know b) factory-maintainer override is sometimes necessary to keep Factory in shape c) one week inactivity timeout for a leaf package that is not crucial for anything is a bit too tight Ok? cu Ludwig -- (o_ Ludwig Nussel //\ V_/_ http://www.suse.de/ SUSE LINUX Products GmbH, GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 16746 (AG Nürnberg) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 08/11/2014 07:41 AM, Guido Berhoerster wrote:

* Ludwig Nussel [2014-08-11 13:18]:

...

c) one week inactivity timeout for a leaf package that is not crucial for anything is a bit too tight

Agreed, it's just unfortunate thatnot much seems to have changed to address this since last time: http://thread.gmane.org/gmane.linux.suse.opensuse.devel/51703

Yes, Sorry, Stefan you are barking up the wrong tree. We have to find a way to address these what I would call structural issues. Maybe we need some general guidelines and role definitions that go along with those? Of course that should be discussed in a different thread ;) Later, Robert - -- Robert Schweikert MAY THE SOURCE BE WITH YOU SUSE-IBM Software Integration Center LINUX Tech Lead Public Cloud Architect rjschwei@suse.com rschweik@ca.ibm.com 781-464-8147 -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQEcBAEBAgAGBQJT6L89AAoJEE4FgL32d2UkO44H/ivbn6hH1Vaa4yCNSXPpul7t T7nnbl6dY2OMIffLEwQzx3ceTljyXlGfeOKQ7sI23OT7pMg73z4hTSIjgsf7vVQ4 bg/yuOHAher/y+1roxwlbi4KZRZ1+n15SZ3Ec7yIym2AuKUJyQk6cJnD/mhzmJMX w+UdTCBmW2u3ylojrP0eHMcVwHg0oABasoppfsc3/TXAC0MfzgiZr6APe6aYy8g5 0hdg32x0vsSJAPxQ3eGetqlDBilaZHnQu2oMOsoXl1G7sB4hCCa8OQU+z/431ySE LTeO4TM5MRvaUgu7Q53TUl/UCbWPxahnrcBCUkZ3DPghXvYSYKQTIIaXDzwUiXQ= =5UDS -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 08/11/2014 02:35 AM, Stefan Seyfried wrote:

Am 11.08.2014 um 07:49 schrieb Michal Kubecek:

...

The problem is it's not one package. In the name of systemd, the whole system is being redesigned and reworked in order to make the other options stop working. I've found numerous cases where changes have been done to other packages that don't help systemd and systemd systems at all but only break non-systemd systems. And then I keep reading these "if you want to provide an alternative, you are welcome to do so" appeals. It's frustrating and I really feel like I'm being mocked on purpose.

Even though I'm generally in favour of systemd, I have to agree with Michal on this: this makes it e.g. almost impossible for me to use openSUSE on embedded / ARM, because I want to use the same system on all my embedded boards and I have boards that cannot run systemd.

Apart from that systemd has the problem that it just boots much too slow on embedded systems that are not from 2014 due to its huge memory and CPU footprint. So no openSUSE for me on ARM, fortunately the yocto project fits just fine for my needs, and it can easily be built even for x86_64 :-)

But the premise "you can just maintain your own non-systemd distribution on OBS" is IMO a plain lie,

Well, other than the necessary build power I am not aware of any limitation in this area. Please fell free to educate me. Based on the experience I have with OBS there is nothing in the way for me to set up a subproject in my HOME and build all of openSUSE, create iso images from it etc. I can maintain patches that add sysV init support in the project and do whatever. There may of course be issues that I am not aware of as I have never attempted anything at that scale. Again, please share your knowledge. Proclaiming that statements are "lies" is not very constructive. Later, Robert - -- Robert Schweikert MAY THE SOURCE BE WITH YOU SUSE-IBM Software Integration Center LINUX Tech Lead Public Cloud Architect rjschwei@suse.com rschweik@ca.ibm.com 781-464-8147 -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQEcBAEBAgAGBQJT6L3LAAoJEE4FgL32d2UkkCUH/3XC7qQZdJ/ggIwo7BmzBceV Iw61uyOPsgsfe91lpXW9T06YFSlkyNLOHNlin0nFXEe3Z9Lal/Y38VUHKI8kIwA7 ZjvdaKNuJY0ZMkAtgID8z630gajlGMlDVgJ4dn8BrmJchiHidHgwsOqdVhZVJexT Y0ORPXZAr+gDhmUFdhaJ2Fp0uJ4OdEYJ21aKdSdzuEjam0Vr8GfDFcCULFWg1lWH U5IToPHbb5gD4uGyUNYHdp6uPDwjJNnekSVnIo77t8gLYq2wqIcD+HkX5/keH1f+ KKal+zH+NLuczY490DBtnTOF7Hyimv/ZX757E9OXPph+XfxvmMvDhiC6tesB+eY= =FUmR -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Cristian Rodríguez wrote:

El 11/08/14 a las #4, Robert Schweikert escribió:

...

Well, other than the necessary build power I am not aware of any limitation in this area. Please fell free to educate me.

Robert, if this were a rocket going to mars, it already took gravity assists sailed across the solar system.

Systemd won, the alternative lost and that's it. More and more applications will gain systemd specific code over the time rendering the old ways unworkable.

If it were a rocket launched to mars, that'd be great! There's always the NEXT launch that doesn't have to be based on it at all. Just because you screw up something in 1 trip doesn't mean you get all of the future times to come. Either systemd addresses all of the issues people have or those people will, in the future, continue to lobby for change. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Hello Roman, On Mon, 11 Aug 2014, Roman Drahtmueller wrote:

For this purpose, the obs project Base:sysv was created. Please have a look at the project's description: It states that there should be a meta-package (at some point in time in the future) that, if selected, pulls in everything else from the repo and thereby obsoletes systemd and installs glue packages that are needed.

A system like this would easily follow my (personal?) demand for

* separation of boot (and after boot) glue pieces according to their respective purposes * leaving subsystems up to their own businesses (cgroups, network, device naming, ...) * leaving services on their own (0,1,2 on /dev/null, chdir("/"), fork() and maybe setsid(2)) ) * configuration of the framework not dependent on a semi-programming language where changes to configuration have unforseeable consequences. * a reasonable network device naming scheme. This is ridiculous.

I haven't had enough time to dig into this yet, so it's pretty much vanilla yet. With changing circumstances in October/November this could change.

If you want to participate, feel free. Also please let me know if you wish to be included in the prioject.

[x] done ;) I guess the systemd-shim, eudev etc. linked on the boycottsystemd.org page should go in there too ... Oh, and probably there should be a way to build agains other libc... -dnh -- For every complex problem, there is a solution that is simple, neat, and wrong. -- H. L. Mencken -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 08/12/2014 09:51 AM, Stefan Seyfried wrote:

Am 11.08.2014 um 14:57 schrieb Robert Schweikert:

...

...

But the premise "you can just maintain your own non-systemd distribution on OBS" is IMO a plain lie,

Well, other than the necessary build power I am not aware of any limitation in this area. Please fell free to educate me.

Well, I'll not try that, but everybody else probably would have a lot of fun if I did start the effort with

for i in `osc ls openSUSE:Factory`; do case $i in systemd*) continue;; *) ;; esac osc linkpac openSUSE:Factory $i home:seife:factory-nosystemd done

...

Based on the experience I have with OBS there is nothing in the way for me to set up a subproject in my HOME and build all of openSUSE, create iso images from it etc. I can maintain patches that add sysV init support in the project and do whatever. There may of course be issues that I am not aware of as I have never attempted anything at that scale. Again, please share your knowledge.

If the weather is bad on the weekend, I might actually try. But I fear that I'll pull the wrath of all other OBS users on me and I'd not consider this a fair and good use of the shared resources.

...

Proclaiming that statements are "lies" is not very constructive.

Yes, "lie" was the wrong word. But I still consider this impractical, since right now 603 packages depend on systemd, and almost all depend on systemd-rpm-macros, so it is nothing that you can "just fix" by replacing the "faulty" component.

But how many of those 603 depend on systemd conditionally? I have some that depend on systemd but also support sysV init. Yes, the sysV init is for older distros but I am certain someone with better RPM macro magic knowledge than me could come up with some clever condition I could test to install sysV init scripts vs. .system files even on a distribution that is nominally systemd. I personally really do not mind the extra maintenance of sysV init scripts, as I am doing it anyway for older distros. In the end as I have pointed out, from my point of view this is up to each package maintainer whether they are willing to have the additional complexity in their spec files, maintain the sysV init scripts, and possibly carry extra patches to the upstream code.

Probably, instead of wasting the build service resources on this experiment, I'd rather try an installation of yocto built x86_64 on my spare openSUSE partition on my notebook. This will only take a few hours to build and on my private resources.

Anyway, this thread has gone long past its usefulness,

I agree, unfortunately many of the comments just regurgitate things that have been said many times and are not very constructive. However, I think there are few things here and there with valuable input. It would be interesting to understand the numbers better. Later, Robert - -- Robert Schweikert MAY THE SOURCE BE WITH YOU SUSE-IBM Software Integration Center LINUX Tech Lead Public Cloud Architect rjschwei@suse.com rschweik@ca.ibm.com 781-464-8147 -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iQEcBAEBAgAGBQJT6iCuAAoJEE4FgL32d2Uka00H/0osPo5qeeMAYGBvl/W04FbJ HU/i0SqtV8BAMeayegDI4cfk6F2PyEirtGi+SzmXfBs4VAEULlo4ZJYed/bcZr0K KG9R9eo+YNJPSdRAxxWtYT6OIT8q0sazoxo/N8mRc5GzPGKB8MkU1DE9ot0VXb3C pyuDurDEiFfH9d8uUaLlSmEGl0592OUed7WS5fMp7PUp+Upieiji89CX2S36q8rg P+P32OR3wqNCF8Upk6X2pTeYJwyCTxUbixgZh6SJiyo/f/E7zhxLuf7cq2OSCX8E fGelfvkLwFYtoi+GfmJXVZL2z2YTx9/bPZUVkbhMBdx9TQhgR692DUDzYud9hxo= =MB+C -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Roman Bysh wrote:

It's time to wrap it up. This whole thread is a three year old rehash.

----- The issue won't go away until the problems with systemd are dealt with. It's like. If you really don't want to deal with it, work to change systemd to be suitable to its detractors . No one who supports system seems to be willing to touch it or push for compatibility -- to the contrary, you have monotheistic "one-way-ists" claiming that one way is the right way. and it should be one way in everything. Until people like that change their attitude or are replaced, no one wants or is allowed to work to make systemd more compatible with existing systems. To the contrary, instructions seem to be to remove compatibility options to deliberately screw people over who don't buy into one way ism. So, don't claim this is about those who haven't drunk the koolaid and how should do something. -- Change can come with -- what you and others claim is a majority -- and push for more open standards and compatibility plugin points. But I still haven't seen anyone in the systemd (MCP) camp step up and work to try to resolve these issues. Instead, efforts are put on making compatibility worse by removing options. I can't see this issue going away as long as the dominating systemdeists refuse to even try to include compatibility.

We have a lot more pressing issues for the _Factory_ mail list.

---- Oh? Seems like systemd is front and center of factory. Hasn't it absorbed OBS yet? -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Am 11.08.2014 um 17:27 schrieb Cristian Rodríguez:

El 11/08/14 a las #4, Stefan Seyfried escribió:

...

Apart from that systemd has the problem that it just boots much too slow on embedded systems that are not from 2014 due to its huge memory and CPU footprint.

What kind of boards are we talking about .. ? last thing I tried was a raspberry PI with archlinux and it booted in a reasonable amount of seconds considering the hardware in question.

Well, but it boots much faster with a simple init system like busybox init plus simple init scripts which I'm using on all embedded boards. Yocto's sysvinit also boots faster than systemd on my embedded systems. I'm talking about digital STB boards where the latest kernel I can use (because of drivers) is 2.6.26 and I cannot enable cgroups (because the drivers won't load anymore), and so systemd is not an option to even try. There's also 10 year old PPC405 STBs, but I never tried to run openSUSE userspace on them, but with kernel 2.6.12 being the last available one, you won't get lucky with systemd on them :-)

Maybe the problem is related to the amount of parallelism ..do you have the graphs generated by systemd-bootchartd to give an idea what may need to be tuned ?

No, because there is no need to tune it. Simple init scripts work just fine and faster. Note I'm not starting much crap on those boxes: mounting/checking rootfs, setting up network, then starting inetd for telnet/ftp and dropbear, while the main application (GUI for TV) is already starting. It looks like systemd is totally busy deciding what to do and figuring out how to do it, while simple scripts already do it. My second guess would be journald which until today is an example of ineffectivity, even on fast x86_64 systems. It made me buy SSDs for old systems, just to make them boot in a bearable time. But I have not gone through the pain of debugging it all. Fortunately the yocto folks are not too keen of systemd and will probably keep it off the "must have" list for quite some time :-) -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Robert Schweikert wrote:

On 08/11/2014 01:49 AM, Michal Kubecek wrote:

...

The problem is it's not one package. In the name of systemd, the whole system is being redesigned and reworked in order to make the other options stop working.

This implies malicious intend. I doubt that anyone makes any changes with the intent to break something else.

-------------- I've been complaining about this since 12.3 hit. It's just that more people are seeing these "one-way" actions as malicious as time goes forward. It was worse in 13.1, but I was better prepared as I ran off factory. But I expect considerable damage in 13.2 will need to be fixed to make it usable. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

El 11/08/14 a las #4, David Haller escribió:

Guess why "systemd-shim" is called as it is? As a "pretend" systemd, to fool all the stuff requiring systemd. That's how far it has already crept into the infrastructure (and upstreams).

systemd-shim is an horrendous hack, fortunately only intended as a transition measure, originally from upstart.

The difference is: until now, no init precluded another. And especially, no init gobbled up tons of other critical functions like logging, udev, dbus etc.

Yes, until before systemd, a number of system components operated with very limited understanding of each other..now things are starting to make sense. and will get even more integrated in the near future.

And as far as I know LP's work, he's far from done yet. Oh, and BTW:

==== 9. systemd is designed with glibc in mind, and doesn't take kindly to supporting other libcs all that much[10]. In general, the systemd developers' idea of a standard libc is one that has bug-for-bug compatibility with glibc. ====

THAT's a huge honking blaring big major NO-NO for all that want to run Linux on embedded stuff using µlibc, dietlibc, ...

Yes, this is because only one thing is supported, one kernel, one libc, one you name it. it is called making precise, limited, realistic design decisions to fullfil a role in a particular market in order for stuff to be actually supportable and avoid drowning in compatibility hacks. I cannot overstate how much I support this and is an idea that should be widely imitated. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, On 08/11/2014 04:31 PM, David Haller wrote:

Hello,

On Mon, 11 Aug 2014, Robert Schweikert wrote:

...

On 08/11/2014 01:49 AM, Michal Kubecek wrote:

...

On Saturday 09 of August 2014 06:09:14 Robert Schweikert wrote:

...

So, my appeal to all. Please stop the systemd bashing, it leads nowhere. If there is a bug in systemd that annoys you, file the bug or go fix it, be constructive. If because of 1 in the 7000+ packages in Factory the openSUSE distribution no longer meets your needs or your stability expectations then so be it.

The problem is it's not one package. In the name of systemd, the whole system is being redesigned and reworked in order to make the other options stop working.

This implies malicious intend. I doubt that anyone makes any changes with the intent to break something else.

The whole design of systemd reeks of malicious intent (now, that's a bit of bashing *hrhr*). It may be unwillingly, but the effect is the same. Out of whatever actual intent.

...

...

I've found numerous cases where changes have been done to other packages that don't help systemd and systemd systems at all but only break non-systemd systems. And then I keep reading these "if you want to provide an alternative, you are welcome to do so" appeals. It's frustrating and I really feel like I'm being mocked on purpose.

Sorry you feel this way. In the end it is up to each package maintainer to decide which SRs to accept and which ones to decline. If there is a pending SR that removes sysV init support and makes systemd the only support than it is up to the package maintainer to decide if this is OK or not.

It is _NOT_ up to the package maintainer if openSUSE removes sysvinit from the distro and tons of changes require you to install systemd.

Well I guess I shouldn't really respond, but just for clarification. Are you telling me that if I maintain a package that also is in factory there is someone that can force me to remove the code in the spec file that also supports sysV init? BTW, I do maintain a number of packages that have sysV init code and systemd code and no one has approached me yet about ripping out the sysV init support, and yes those packages are in Factory. Anyway, please introduce me to the person that can force me or any other maintainer to get rid of sysV init code in a package, I certainly have not yet made their acquaintance.

<snip>

If systemd just were another init (like upstart, openrc, etc.), I wouldn't care. But IT IS NOT. Far from it!

Well something obviously makes you hate systemd, a very emotional reaction to bits that manipulate transistors on a chip made from inorganic materials. Later, Robert - -- Robert Schweikert MAY THE SOURCE BE WITH YOU SUSE-IBM Software Integration Center LINUX Tech Lead Public Cloud Architect rjschwei@suse.com rschweik@ca.ibm.com 781-464-8147 -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iQEcBAEBAgAGBQJT6gr3AAoJEE4FgL32d2UkuD8H/RCWKGMzgvQ/22b2fs1f5xGn 7imFF1UX8UDer8QHwd2g9OZNOelyrxePFfNziahQRj0mj0K7F9auxjILrejZcsNy 4qCbswSJcnDp3jCwKP05neYxLZuP6anItKbEq7darhe9fl3fSy89I2UjrJOOLHMA Cr7GM7IXqnjNpjDv693QL8mCVihuuniGK0LOPQxt+X2wBVcbpzj/n1Q7pAQeMn8E 3+HSepcsiAmPyrECLRCAo1yRvDsAj9HhC92jcuI5m6AeL14KnxpYNC5isi9Uktlh +L+trDkicP91SxH2AznBBXruDRMmKTb2bZHeYxtV6J42tB2b8yPcZD9PQXXsiyU= =DpmH -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 08/12/2014 09:03 AM, Michal Kubecek wrote:

On Tuesday 12 of August 2014 08:39:19 Robert Schweikert wrote:

...

Well something obviously makes you hate systemd, a very emotional reaction to bits that manipulate transistors on a chip made from inorganic materials.

You know well it's not about the bits. It's about the people behind it and their behaviour. David said it clearly: if it was just another project available as an option, he wouldn't mind. But systemd people make an enormous effort to make their project the only possibility, to make it as hostile and impossible to coexist with as possible. That's what makes people hate it. Unfortunately those people are also very skilled in the PR targeted on the narrow groups of decision makers. That's why systemd people were allowed to throw everyone else out of the distributions rather then being told that they are not allowed in until they are able and willing to live with others.

Or maybe the maintenance of multiple init systems in a distribution is just very expensive and possibly more than doubles the required testing. And maybe systemd showed enough promise for improving the status quo that the so called decision makers thought it was worth the trouble of switching. Certainly people have their misgivings about Lennard and Kay. However, on the one hand apparently people like to paint their picture as evil incarnate and then on the other hand they are accused of being very good at PR to convince all these "decision makers" that systemd is the way to go. Hmmmm, makes me wonder...... Later, Robert - -- Robert Schweikert MAY THE SOURCE BE WITH YOU SUSE-IBM Software Integration Center LINUX Tech Lead Public Cloud Architect rjschwei@suse.com rschweik@ca.ibm.com 781-464-8147 -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iQEcBAEBAgAGBQJT6hePAAoJEE4FgL32d2UkR08H/iKt6i+bXVKkRS7x32b0ACM1 nWZfuT9taHlQe36KsxNastfxrlGLXVoRE9llDF+RgBALYkj6epQ77Ji7Z3ljWAfj dZzokIxsZS0HLMg02wgYdoEcpK20WS5YpY7kVrlf4WQUzT4JmYcXE9wBaSshvMh2 Dl8ILtrRhnghJjjunmAv6LibsHgR+uvbDwO8Ed4epZXzRL5vGJZi9OFIHl5RC0Ad i8gBR+Z4bEYG8S+2sRJAXg/5nN/J0yEfkggULLBSG9Sx1ju6EQtXsgq+S1RT+D0b MZhg7laURjFGXTLELyT4cpcikNcRgfe6QwMdNQ7LxRo4pjFLV5hb4yOrgzblJCI= =QWgC -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Hey, alright I heard enough. Guess I have to moderate this thread for constructiveness... Henne -- Henne Vogelsang, Mailing List Admin http://www.opensuse.org -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 08/09/2014 10:37 AM, Cristian Rodríguez wrote:

El 09/08/14 a las #4, David Haller escribió:

...

PS: I seriously do consider jumping ship (to gentoo or FreeBSD ATM). And I've been using SUSE for >15 years now and contributing quite a bit.

That would be fantastic, so you will stop posting stupid, uninformed, incorrect pieces of misinformation.

So here is the other side of the coin. Comments like these are just as inflammatory and do not help to resolve any misgivings people might have about systemd. Be CONSTRUCTIVE Later, Robert -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQEcBAEBAgAGBQJT50rHAAoJEE4FgL32d2Uk47MIALlrSWQ27lf18A380DPaPGlK hRNA9Y2oOifam9ocCJ5auQRYHtlmMRZnf/waQ3d5NWunxgAAefPO9hRp0uELSYiz aD9AuHZXZCvPn2yK6UOsWGnLTi2/gV+83OqEd7GtDrssy5tH2jSMYqA+mgX/5Ryb xOvMBvVVwBWwuT2wmqKuN1uay2ZIrOjvKEB24YNddyhlcMDT8ZKYFSLbeeUL79Hp eU+K/6C3Tr52XUzp8FUx57BTBA6MyUh7PfAvNDHKa42R6rmhMi0jyaq2hRuXq8ws b4ymBxoqTpjIfKdbGT2xTrnYileMpgZEAWvu4d+366RSjG0KS6ccrh8Vo2jfL7Q= =Gqdl -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Am 10.08.2014 um 19:57 schrieb Cristian Rodríguez:

You can't. Instead of getting actual usable feedback, you get vague complains, ideological crap, conspiracy theories, arguments from tradition and ridiculous posts comparing free software with tyranny.

+1. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Sun, 2014-08-10 at 13:57 -0400, Cristian Rodríguez wrote:

El 10/08/14 a las #4, Robert Schweikert escribió:

...

So here is the other side of the coin. Comments like these are just as inflammatory and do not help to resolve any misgivings people might have about systemd.

Nothing will help with that, these people are impervious to reason, facts, have never bothered to understand why, how things work or what's the rationale behind changes.

With this statement you admit that _you're_ unable to help and _you_ have resigned to try to reason with or understand "these people". So why do you get involved in such threads? What do you want to achieve?

...

Be CONSTRUCTIVE

You can't.

So you resort to being unconstructive? For what end?

Instead of getting actual usable feedback, you get vague complains, ideological crap, conspiracy theories, arguments from tradition and ridiculous posts comparing free software with tyranny.

You mean being unconstructive? ;) -Scott -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

El 11/08/14 a las #4, David Haller escribió:

So, what is incorrect on the linked page?

Everything. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

В Sat, 09 Aug 2014 09:55:55 -0700 Lew Wolfgang пишет:

Skepticism is a critical part of the scientific method

It has nothing to do with science - it is natural selection. I was able to use Mandriva with systemd version around 15 (give or take). At least it was long before version 20 which included Mandriva support. That's almost 4 years ago and about 1 year after the first version of systemd appeared. That's how you make software to be adopted - you create it and let users use it. Systemd was useable from the day one. If those "scientific skeptics" were not able to come with working alternative in 4 years - what makes you believe they will be able to do it in next 40 years? If they want to help fixing bugs in systemd - they are welcome to do it. If they want to offer alternative - they are welcome to do it. But personally I'm tired of listening to the same "We don't need no change" over and over again. It is far from being scientific and has nothing to do with skepticism. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Saturday 09 of August 2014 21:49:13 Andrey Borzenkov wrote:

If those "scientific skeptics" were not able to come with working alternative in 4 years - what makes you believe they will be able to do it in next 40 years?

There _was_ a working alternative. But then systemd people came and rather than providing an alternative, they created a _replacement_ and put a lot of effort to making sure a systemd-compatible system would be as incompatible with anything else as possible. In the old days when users decided what alternative they prefer, such project couldn't prevail unless its authors changed their approach and made it able to coexist with others. Unfortunately, we now live in era when even in "community" distribution, key decisions are made by a small group of managers.

If they want to help fixing bugs in systemd - they are welcome to do it.

There are well known examples of systemd bugs that were ignored even if patches were submitted, so that in the end, workarounds in other projects had to be created because the systemd maintainers simply refused to fix their own bugs. Michal Kubeček -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Cristian Rodr�������������������������������� wrote:

El 11/08/14 a las #4, Michal Kubecek escribi�:

...

There are well known examples of systemd bugs that were ignored even if patches were submitted, so that in the end, workarounds in other projects had to be created because the systemd maintainers simply refused to fix their own bugs.

That's usually because of disagreement on where and whose bug is. nobody refuses to fix bugs, that is non-sense. Bull doo doo!

It's called "we no longer support you way of doing things" and the unfixed bug gets marked rejected. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

After all this emotional discussions, this suggestion sounds very constructive. Stefan Am 13.08.2014 15:49, schrieb Tomas Cech:

Sorry to enter this thread this late. I'll try to be constructive as much as possible. I don't want to continue in discussion about features or bugs in systemd but rather talk about possible alternative.

As I read opinions here, people here agree that the way switch to systemd was done and the fact that we don't have possible alternative to that - it was unfortunate and something we should learn from.

1] I see no reason, why not to do it properly now. Can we somehow create poll with question

'Do you want support for systemd in openSUSE?'

That could quantify opinions of community and it wouldn't depend on how strong voice Christian or Linda have (in number of e-mails or emotional outbursts :).

In case that openSUSE community would feel that having working alternative to systemd is important for them, we could

2] set the rule, that package should be able to compile and work both with and without systemd.

We could create system wide RPM macro like '%with_systemd' and use it in spec files to make the tests easier to perform and bring back old init scripts. (It could also help people to derivate systemd-less system).

Bringing back old init (or adapting another one like dmd :b) is huge task for small group, but can be reasonable task for each package maintainer.

So shortly,

check if there is will of people and if there is, change this political problem to technical one which can solve together.

We are the driving community, we can do the change _iff_ we want.

Best regards,

Tomas Cech Sleep_Walker openSUSE member, package maintainer and occasional factory reviewer Best effort is

PS: no pathos meant

-- www.invis-server.org Stefan Schäfer Ludwigstr. 1-3 63679 Schotten -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Tomas Cech wrote:

I created poll in our connect tool:

https://connect.opensuse.org/pg/polls/read/sleep_walker/45294/do-we-want-to-...

Every logged user can vote, use the same credentials as in OBS or bugzilla. I tried to be as objective in formulation as I could.

--- For the 'yes', I'd rephrase that... "Yes, I want it to be be possible to allow users a choice of which parts of systemd they want to use with the option to use other alternatives for parts where systemd doesn't work for them." and "No" should read: No, I don't think suse should support any alternatives to systemd. As it is now, it sounds like a yes vote is one of desperation, -- wanting support for ANY alternative to systemd, which isn't very true even for me. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Wed, Aug 13, 2014 at 10:44:19PM +0200, Tomas Cech wrote:

On Wed, Aug 13, 2014 at 01:36:03PM -0700, Linda Walsh wrote:

...

Tomas Cech wrote:

...

I created poll in our connect tool:

https://connect.opensuse.org/pg/polls/read/sleep_walker/45294/do-we-want-to-...

Every logged user can vote, use the same credentials as in OBS or bugzilla. I tried to be as objective in formulation as I could.

--- For the 'yes', I'd rephrase that...

"Yes, I want it to be be possible to allow users a choice of which parts of systemd they want to use with the option to use other alternatives for parts where systemd doesn't work for them."

and "No" should read:

No, I don't think suse should support any alternatives to systemd.

As it is now, it sounds like a yes vote is one of desperation, -- wanting support for ANY alternative to systemd, which isn't very true even for me.

Right, the use of ANY wasn't appropriate, my english teacher would scold me if she could :b

Rephrased to: Yes, I want make possible support some alternative to systemd.

No, I don't want make possible to support any alternative to systemd

I'm really trying not to provide unballanced, leading choices.

OMFG, current status was reseted when the answers were rephrased. Nice feature to not influence the result but was somehow unexpected. I'm sorry about that. S_W

S_W

On Wed, Aug 13, 2014 at 10:48:40PM +0200, Tomas Cech wrote:

On Wed, Aug 13, 2014 at 10:44:19PM +0200, Tomas Cech wrote:

...

On Wed, Aug 13, 2014 at 01:36:03PM -0700, Linda Walsh wrote:

...

Tomas Cech wrote:

...

I created poll in our connect tool:

https://connect.opensuse.org/pg/polls/read/sleep_walker/45294/do-we-want-to-...

Every logged user can vote, use the same credentials as in OBS or bugzilla. I tried to be as objective in formulation as I could.

--- For the 'yes', I'd rephrase that...

"Yes, I want it to be be possible to allow users a choice of which parts of systemd they want to use with the option to use other alternatives for parts where systemd doesn't work for them."

and "No" should read:

No, I don't think suse should support any alternatives to systemd.

As it is now, it sounds like a yes vote is one of desperation, -- wanting support for ANY alternative to systemd, which isn't very true even for me.

Right, the use of ANY wasn't appropriate, my english teacher would scold me if she could :b

Rephrased to: Yes, I want make possible support some alternative to systemd.

No, I don't want make possible to support any alternative to systemd

I'm really trying not to provide unballanced, leading choices.

OMFG, current status was reseted when the answers were rephrased. Nice feature to not influence the result but was somehow unexpected.

Miska fixed that, thanks. S_W

Tomas Cech wrote... Considering the bias nature of the question, I presume that the polling represents a representative solution -- and that any solution will proportionately represent the numbers. I.e. for those not wanting any alternative -- no compatibility packages would be installed. Versus, representing the majority who who didn't vote for no choice, a compatibility solution would available for installation. First question should be "Do you want to allow alternatives to systemd". Saying "any" makes it sound like there could be only 1. The world 'any' brings up a binary totality that only systemd claims. Ex. Can you think of any solution? (Makes it sound like it is some big thing to allow ANY other possible solution, while the desire is to allow not require a choice of "a" or "b", but have a framework that supports multiple choices. Currently, choice "C" is the only possibility for those who realize that having a choice of only Democrats or Republicans is only being allowed to vote for 2 sides of the same coin. Is that really a choice? Note: those in political science will note that having a 2 party system where the winner takes all is one of the worst forms of "representative" government. Unless you are considering that the answer represents a proportionate representation of what the final solution should be -- which leads to the initially proposed interpretation. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Hi people, On Wed, Aug 13, 2014 at 04:39:19PM +0200, Tomas Cech wrote:

On Wed, Aug 13, 2014 at 03:49:28PM +0200, Tomas Cech wrote:

...

Sorry to enter this thread this late. I'll try to be constructive as much as possible. I don't want to continue in discussion about features or bugs in systemd but rather talk about possible alternative.

As I read opinions here, people here agree that the way switch to systemd was done and the fact that we don't have possible alternative to that - it was unfortunate and something we should learn from.

1] I see no reason, why not to do it properly now. Can we somehow create poll with question

'Do you want support for systemd in openSUSE?'

I created poll in our connect tool:

https://connect.opensuse.org/pg/polls/read/sleep_walker/45294/do-we-want-to-...

Every logged user can vote, use the same credentials as in OBS or bugzilla. I tried to be as objective in formulation as I could.

results of the poll are: Yes, I want make possible support some alternative to systemd - 65 votes No, I don't want make possible to support any alternative to systemd - 83 votes I don't care - 24 votes I take it that majority doesn't want to allow alternative available. Fortunatelly there were also voices who said that if we want alternative and will do an alternative, it will be accepted in packages. So the situation is somehow same before the poll but some opinion proportion was shown. I hope that this poll will end all the discussion about systemd and we'll do something productive once again. Thanks for your votes. S_W

Tomas Cech wrote:

That could quantify opinions of community and it wouldn't depend on how strong voice Christian or Linda have (in number of e-mails or emotional outbursts :).

---- Number of emails, yes**, but I don't think I've had any recent emotional outbursts. Christian seems to be almost avoiding them by filtering out my messages. which is his choice, but it again lends it self to the perception that no matter what concerns are voiced about systemd, they won't be heard or acted on. ** but in my defense, I go through long periods on not posting much, as well, but when the topic comes up, I try to

2] set the rule, that package should be able to compile and work both with and without systemd.

---- I'd love to see that -- AND I'd love to see systemd go foward, as long as there is a choice -- a fallback position should should something come up that would prevent or someone from using systemd. From what I've seen so far, it doesn't seem like has auditing support and that makes me wonder how far it is going to go forward in security sensitive areas. If it cannot not be proven secure and validated as being secure, then there are no assurances to what is being promised. This should be a consideration for the intended use of this on a "secure boot" system. Of course if OS audit functionality was enhanced to include systemd, then its authors might have realized that some of their requirements (like needing to run as process 1 were not justifiable with what auditing already provides.

PS: no pathos meant None taken, and ditto here.

Besides, at this point, contrary to what some thing about rehashing these things not being productive. I have seen progress and ideas come out of this discussion. At the very least, I think it is important to point out that for those who would like to not hear or see any more complaints of those who miss features or have problems with systemd - those people are just as empowered to fix those rough edges and provide solutions to those complains as those making complains are to provide alternatives to systemd. I.e. it's a two way street. If they are not willing to put in as much effort fixing systemd rough edges then they have no right to whine about those who bring up new problems, or what is likely to be more people bringing up old problems (just that the are "only now" hitting them). Not everyone lives out of factory or on the bleeding edge... -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org