Damian Ivanov wrote:

I guess we're at a breakpoint here in the discussion. You say people are wrong. I say you are.

No, I said his claiming that systemd wasn't monolithic because it was 60 pieces of interlocking software that nothing else could tie into was incorrect. You could disagree with that, but I could add a further comparison with MS's OS being monolithic except where other pieces of software can be inserted as replacements. The parts that work together and exclude any others are 1 monolithic mass. That is nothing about people. As for the statement for the writer, I can back them up with the authors quotes. Can you show me anything that shows him working with other software groups? Do you disagree with the sentiment of antipathy that is held by his followers to anyone who disagrees with his methodology of not supporting alternatives and actively cutting them off? Again, none of this was about people, but specific actions. You can argue against their truthfulness, or not. But by being unable to argue against their truthfulness and only degenerating to calling me "wrong" only lends credence to that my statements are not readily or easily refutable as you know that I can backup my assessment of his abilities to work with others by his own actions and opinions which he readily share. Trying to make this personal is a tactic of his supporters who want to ignore decades of what has passed for good software design. As an example of a hot headed personality that has managed to put together a team to build their project, I might suggest linus. Others might be Larry Wall, or Jeremy Allison. There are many other examples of people who don't operate software projects the way Poettering does, who seem to be intent on creating a his walled garden. I do note, though, that he claims it is SuSE's* fault for not supporting boot without initrd and separate /usr. That those issues continue to be misrepresented and supported as "the only way" in this forum don't reflect well on his supporters. *-not systemd's but any given distro's "choice" to exclude alternatives. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Sun, Aug 10, 2014 at 10:52 AM, Linda Walsh wrote:

Damian Ivanov wrote:

...

I guess we're at a breakpoint here in the discussion. You say people are wrong. I say you are.

--- No, I said his claiming that systemd wasn't monolithic because it was 60 pieces of interlocking software that nothing else could tie into was incorrect. You could disagree with that, but I could add a further comparison with MS's OS being monolithic except where other pieces of software can be inserted as replacements.

The parts that work together and exclude any others are 1 monolithic mass.

That is nothing about people.

I have to agree with Linda (and the archive shows I don't always) There's a lot of stupid denial around systemd. It's not a good idea to adopt software that is in such a sorry state of development (being in denial of concerns raised by knowledgeable peers is a sorry state). That said, it was adopted, rolling back is no easy trip, nor a desired one it seems. The state is fixable, the bad design isn't as bad as to not be fixable, and I hope it gets fixed. Pragmatically, I've had no issues with systemd, except it's complexity. It is complex. People say its .system files are simple, and they are, but their interactions and even the file structure is not. It's not easy to grasp how it all fits together, I honestly feel sysvinit, with all the horrible hacks it had, was simpler to understand. That is fixable. If not with documentation, with experience: as low hanging fruits in transparency are found, and fixed, the system will get easier to handle. The support I give systemd is, thus: I don't like it, but I'm not intrinsically opposed to the principle, just the terrible implementation, so I hope it gets to the point where I can like it soon. If I could help, I'd help. Maybe when I have the time and find something I can fix, I'll help. On Mon, Aug 25, 2014 at 7:55 PM, Kyrill Detinov wrote:

I've just found that the systemd package includes 369 files (381 already, while writing this message). I'm shocked. openSUSE applied 300+ patches??? Is systemd good and stable for openSUSE? It seems, NO.

I see, every systemd commit to Factory is ~ +10 patches. What's the hell systemd is good for openSUSE? Looking at Fedora? No Fedora way, please. We are openSUSE.

That's openSUSE supporting systemd through patches. Since systemd is in such a horrible state, openSUSE has to fix tons of stuff, and it has. I've seen the improvement in stability at least in my setups. Those patches will eventually be part of systemd, because openSUSE has folks that persevere in trying to convince their developers of the need for them. I think that must be applauded. I hope this trend moves also into design and interoperability issues, which is an area where systemd sucks big time. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Mon, Aug 25, 2014 at 08:28:12PM -0300, Claudio Freire wrote:

On Sun, Aug 10, 2014 at 10:52 AM, Linda Walsh wrote:

...

Damian Ivanov wrote:

...

I guess we're at a breakpoint here in the discussion. You say people are wrong. I say you are.

--- No, I said his claiming that systemd wasn't monolithic because it was 60 pieces of interlocking software that nothing else could tie into was incorrect. You could disagree with that, but I could add a further comparison with MS's OS being monolithic except where other pieces of software can be inserted as replacements.

The parts that work together and exclude any others are 1 monolithic mass.

That is nothing about people.

I have to agree with Linda (and the archive shows I don't always)

There's a lot of stupid denial around systemd. It's not a good idea to adopt software that is in such a sorry state of development (being in denial of concerns raised by knowledgeable peers is a sorry state).

That said, it was adopted, rolling back is no easy trip, nor a desired one it seems. The state is fixable, the bad design isn't as bad as to not be fixable, and I hope it gets fixed.

Pragmatically, I've had no issues with systemd, except it's complexity. It is complex. People say its .system files are simple, and they are, but their interactions and even the file structure is not. It's not easy to grasp how it all fits together, I honestly feel sysvinit, with all the horrible hacks it had, was simpler to understand. That is fixable. If not with documentation, with experience: as low hanging fruits in transparency are found, and fixed, the system will get easier to handle.

The support I give systemd is, thus: I don't like it, but I'm not intrinsically opposed to the principle, just the terrible implementation, so I hope it gets to the point where I can like it soon. If I could help, I'd help. Maybe when I have the time and find something I can fix, I'll help.

On Mon, Aug 25, 2014 at 7:55 PM, Kyrill Detinov wrote:

...

I've just found that the systemd package includes 369 files (381 already, while writing this message). I'm shocked. openSUSE applied 300+ patches??? Is systemd good and stable for openSUSE? It seems, NO.

I see, every systemd commit to Factory is ~ +10 patches. What's the hell systemd is good for openSUSE? Looking at Fedora? No Fedora way, please. We are openSUSE.

That's openSUSE supporting systemd through patches. Since systemd is in such a horrible state, openSUSE has to fix tons of stuff, and it has. I've seen the improvement in stability at least in my setups. Those patches will eventually be part of systemd, because openSUSE has folks that persevere in trying to convince their developers of the need for them. I think that must be applauded.

Sorry but this is simply not true. Most of the patches *are* part of the current upstream version! You may have a look into the patch collection and read the git headers. Werner -- Dr. Werner Fink -- Software Engineer Consultant SuSE LINUX Products GmbH, Maxfeldstrasse 5, Nuernberg, Germany GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 21284 (AG Nuernberg) phone: +49-911-740-53-0, fax: +49-911-3206727, www.opensuse.org ------------------------------------------------------------------ "Having a smoking section in a restaurant is like having a peeing section in a swimming pool." -- Edward Burr

On Tue, Aug 26, 2014 at 05:31:37AM -0300, Claudio Freire wrote:

On Tue, Aug 26, 2014 at 5:28 AM, Dr. Werner Fink wrote:

...

...

On Mon, Aug 25, 2014 at 7:55 PM, Kyrill Detinov wrote:

...

I've just found that the systemd package includes 369 files (381 already, while writing this message). I'm shocked. openSUSE applied 300+ patches??? Is systemd good and stable for openSUSE? It seems, NO.

I see, every systemd commit to Factory is ~ +10 patches. What's the hell systemd is good for openSUSE? Looking at Fedora? No Fedora way, please. We are openSUSE.

That's openSUSE supporting systemd through patches. Since systemd is in such a horrible state, openSUSE has to fix tons of stuff, and it has. I've seen the improvement in stability at least in my setups. Those patches will eventually be part of systemd, because openSUSE has folks that persevere in trying to convince their developers of the need for them. I think that must be applauded.

Sorry but this is simply not true. Most of the patches *are* part of the current upstream version! You may have a look into the patch collection and read the git headers.

What is not true? (there are many assertions in the quoted paragraph)

Simply to count patches without looking into the patches systemd> grep -E '^(From|Based on) ' *.patch | wc -l 286 which are 99.9% upstream, including our own upstream submits which afterwards are included as upstream backports. Some of our own patches are adopted but not backported as those would require an version update.

I never quantified how many were backports and how many were fixes, and will accept the previous post numbers, but I know from following the list and using the distribution that lots of thing have been fixed since the first systemd-enabled release, and lots of those fixes were pushed by openSUSE contributions.

So, to which are you referring there as false? Is openSUSE not contributing as much as I say? Does everything come from upstream? Or does upstream refuse everything perhaps?

They do not refuse everything. They are very carefully and as described in an other mail in the same thhread they accept or adopt patches as well as sometime refuse patches. And sometimes they are a bit scary about changes like here https://bugs.freedesktop.org/show_bug.cgi?id=82004 Werner -- Dr. Werner Fink -- Software Engineer Consultant SuSE LINUX Products GmbH, Maxfeldstrasse 5, Nuernberg, Germany GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 21284 (AG Nuernberg) phone: +49-911-740-53-0, fax: +49-911-3206727, www.opensuse.org ------------------------------------------------------------------ "Having a smoking section in a restaurant is like having a peeing section in a swimming pool." -- Edward Burr

On Sat, 09 Aug 2014 13:25:09 -0700 Linda Walsh wrote:

Damian Ivanov wrote:

...

Have a look at: http://0pointer.de/blog/projects/the-biggest-myths.html\

---

I'll just address the 1st.

Systemd isn't open to any current standards. It is monolithic - because the 69 parts can't work with anything else. They work as 1 piece. They can't be dropped into other environments and be useful and other pieces of software can't be plugged in to replace any of those 69 parts. Monolithic means it isn't open to any standards outside itself. It's a self-enclosed ecosystem -- and that is what makes it monolithic. The designer has consistently refused to allow for alternatives. That inherently makes it monolithic.

That it's writer cannot understand basic ideas of interoperability and thinks that his way is good for everyone reminds me of any dictator that has preceded a totalitarian regime. I don't really expect much better in this case. I'd love to be surprised, but the antipathy for anyone who doesn't step into line and follow -- is reminiscent of various lovely times in history -- the inquisition, the crusades, the rise of stalin, mussolini and their period peer, to name a few. The mindset is simply antithetical to a healthy ecosystem.

I've just found that the systemd package includes 369 files (381 already, while writing this message). I'm shocked. openSUSE applied 300+ patches??? Is systemd good and stable for openSUSE? It seems, NO. I see, every systemd commit to Factory is ~ +10 patches. What's the hell systemd is good for openSUSE? Looking at Fedora? No Fedora way, please. We are openSUSE. -- WBR Kyrill

On Mon, 25 Aug 2014 16:26:32 -0700 Greg KH wrote:

On Tue, Aug 26, 2014 at 02:55:15AM +0400, Kyrill Detinov wrote:

...

I've just found that the systemd package includes 369 files (381 already, while writing this message). I'm shocked. openSUSE applied 300+ patches??? Is systemd good and stable for openSUSE? It seems, NO.

Does the number of patches in a package relate to how "good and stable" it is for a distro?

I think so. If a package is programmed well for any distro, why do we need to apply 300+ patches to make it work?

Have you looked at the kernel package?

I believe, the kernel package is well maintained.

What you should be looking at is how many of those patches are not upstream already (hint, most are), so how is this relevant?

...

I see, every systemd commit to Factory is ~ +10 patches. What's the hell systemd is good for openSUSE? Looking at Fedora? No Fedora way, please. We are openSUSE.

What does that mean? Are we not all both good solid community based Linux distributions?

I see, the patches are not upstreamed. So, I understand that we aren't happy in the current state (who are happy?). I mean, Fedora is very experimental for end-users. For example: https://plus.google.com/u/0/+AlanCoxLinux/posts/aCiB7kTLXTh -- WBR Kyrill

On Monday 25 of August 2014 16:26:32 Greg KH wrote:

On Tue, Aug 26, 2014 at 02:55:15AM +0400, Kyrill Detinov wrote:

...

I've just found that the systemd package includes 369 files (381 already, while writing this message). I'm shocked. openSUSE applied 300+ patches??? Is systemd good and stable for openSUSE? It seems, NO.

Does the number of patches in a package relate to how "good and stable" it is for a distro?

Given that (well maintained) Factory packages tend to follow the upstream releases, they should only contain the patches that are (a) in the process of getting into an upstream release (b) rejected by upstream but we still consider them useful (c) intentionally deviating from upstream behaviour So yes, I believe that the number of patches we keep in a Factory package (relative to the size of the original sources) tells us something about how usable the upstream version is for our purposes. It's not the only criterion but it cannot be disregarded entirely.

Have you looked at the kernel package?

Well, I did, out of curiosity. At the moment, the Factory kernel source package has 230 patches for a 77MB (565 MB uncompressed) tarball. The Factory systemd source package has 347 patches for a 2.5MB (28 MB uncompressed) tarball. The ratio of ~30 ( = (347 / 28) / (230 / 565)) between those is IMHO kind of disturbing. Michal Kubeček -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Tuesday 26 of August 2014 06:46:19 Dominique Leuenberger a.k.a. Dimstar wrote:

Quoting Michal Kubecek :

...

So yes, I believe that the number of patches we keep in a Factory package (relative to the size of the original sources) tells us something about how usable the upstream version is for our purposes. It's not the only criterion but it cannot be disregarded entirely.

if THIS is the metric to improve the quality of our packages, then that's easy:

I hope you see the difference between your (mis)interpretation and what I actually wrote.

I'll just submit a new package that has all diffs merged into one patch, making the package entirely unmaintainable.. but hey: your metric is much better, as it's only one patch left.

Absurd examples are absurd, that's the point of them. Real life packages shouldn't be. I would also like to remind that it was Greg who suggested to compare the number of patches in systemd package with kernel, I only did what he suggested to show him it wasn't a good example to show his point. Michal Kubeček -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Tue, Aug 26, 2014 at 05:17:39AM -0300, Claudio Freire wrote:

On Tue, Aug 26, 2014 at 3:46 AM, Dominique Leuenberger a.k.a. Dimstar wrote:

...

AsS for systemd with patches: the maintainer prefers to backport fixes to the version of systemd we currently have. Almost all patches come from upstream. The maintainer for now decided to stay on systemd version 210 (from February 2014) wheras upstream has produced releases up to version 216 (latest from 19 Aug).

The beauty of this? It's the maintainers free choice to provide what he feels is maintainable and gives the distribution the highest gain. If you feel the version number means a lot: offer assistance in updating.

You do notice, however, how that puts in evidence the awful design of systemd right?

I think I've heard that the reason it's not upgraded to the last version is that they depend on a specific version of the kernel to work, so you've got to update both kernel and systemd in order to get a functional system.

That's just ridiculous design. Why is changing that not a first priority for packagers? (I'd imagine it's in their best interest to remove kernel version lock-in).

No, the Factory systemd is just currently developed parallel with the SLE12 systemd. Ciao, Marcus -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org