![](https://seccdn.libravatar.org/avatar/c4d991702dcb0afa2b2afd8464be7f66.jpg?s=120&d=mm&r=g)
On 12 January 2016 at 12:05, Bernhard M. Wiedemann <bernhardout@lsmod.de> wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 29/10/14 17:29, Bernhard M. Wiedemann wrote:
Hi,
during our Hackweek at SUSE I spent a day or two to improve my tool to help finding the maintainer for software in Factory, which comes in handy for bugzilla-screening (getting bugs to the right people).
You can use the tool at http://maintainer.zq1.de/
one thing I forgot to mention since I parsed all those package databases, I noticed that openSUSE is rather on the low end of number of packages in the list of those active distributions that I found
slackware 1328 entries voidlinux 6451 entries centos 7009 entries archlinux 8259 entries opensuse 8655 entries mageia 12839 entries altlinux 17621 entries fedora 17698 entries gentoo 18519 entries debian 23914 entries ubuntu 25417 entries
While I wholeheartedly share the critique of these numbers provided already by Jan and Michael, I do have some comments about your conclusions which I still think are useful discussion points
There are some reasons I know of and probably some more
1. We dont always split off sub-packages and the count is for source-packages (at least for openSUSE - for some others it counts binary packages)
I do not see this a problem - If we find this the best way of maintaining our packages, we should keep doing it. Changing our development methods just to artificially increase our package count numbers would be stupid.
2. We are quick to drop packages that dont build. And sometimes that even breaks packages in devel projects because those usually just link to Factory
I also believe this is not a problem - if packages don't build, they should be fixed quickly. We could possibly do with better notifications/dashboards in this area to make it easier for maintainers to realise when their stuff is broken.
3. We dont bother submitting everything from devel projects. I see that a lot with the games repo, but I guess there are others containing useful software that is not in Factory for various reasons - - e.g. because we always want the latest version to be used everywhere and submitting maintenance-updates is not as easy as just telling people to use the extra repo.
I believe this to be a big problem. I support and want to strongly encourage any effort which decreases the diff between [Sum of Packages in Devel Projects] and [Sum of Packages in Tumbleweed]
Maybe we should do something similar to the old Greg-K-H Tumbleweed where we create an overlay repo that only contains leaf packages in their latest stable version but that builds for all released openSUSE+SLE versions. That does not overlap with https://build.opensuse.org/project/show/openSUSE:Backports:SLE-12 because it only contains packages that are in Factory.
I don't see how that fixes the issue. People need to submit their packages to Tumbleweed, submitting to some overlay repo is just going to be another step which people may or may not do, and I'd much rather see them take the steps to get their packages out of Devel hell and into our distributions properly. I'm surprised I have to remind the grandfather of openQA that just because something builds, doesn't mean it works; Our history with GregKH-style Tumbleweed showed that overlays are not inherently more maintainable or reliable than Modern Tumbleweed style rolling releases. Assured Quality comes from testing, and right now Tumbleweed is tested, no Devel Repo (and I assume no overlay repo) is. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org