The package was introduced to the tree, but it was forgotten to add a developer with write access. https://gitweb.gentoo.org/repo/gentoo.git/commit/dev-util/gyp/metadata.xml?id=03107fd1ed966a6eb33fbc93c422c4c38df0dabf
Every package needs at least one developer is responsible and has write access. If no developer monitors and cares for the package, it is maintainer-needed. There are many standard users, who want to introduce their packages in the same way, but we suggest to go with an overlay until a developer is willing to be responsible. It would be a sign of bad quality, if we introduce packages to the official tree which are not maintained by a dev but not marked as maintainer needed. I asked the two involved but could not find an easy solution: Arfrever (user, who is not dev and not proxied maintainer) asked me not to change the metadata. floppym (dev, who merged) does not want to be in the metadata, because he does not want to get mails from bugzilla. QA Team, what is the best solution? Please do the necessary things.
(In reply to Jonas Stein from comment #1) > Every package needs at least one developer is responsible and > has write access. There is no such policy. When a developer pushes a commit authored by somebody else, he/she is co-responsible for this specific commit, but is not responsible for future maintenance of given package. > If no developer monitors and cares for the package, it is maintainer-needed. The term "maintainer-needed package" means a package which has no maintainer of any kind (any person or project) at all. > It would be a sign of bad quality, if we introduce > packages to the official tree which are not maintained by > a dev but not marked as maintainer needed. This is continuation of above fallacy. > Arfrever (user, who is not dev and not proxied maintainer) No idea what you mean, but I am maintainer of dev-util/gyp. > QA Team, what is the best solution? There is no technical problem in this metadata.xml. This metadata.xml accurately reflects reality. I could copy project from reverse dependency of dev-util/gyp (i.e. app-i18n/mozc), but this would be unnecessary overcomplication. Status quo does not trigger any problems.
The way I see it, this is a problem. While so far Arfrever's track record as proxied maint wasn't apparently bad, this sets a bad precedent. When a package is maintained by a developer and the developer goes MIA, Undertakers eventually take care of the reassignment and the package becomes m-needed. When a package is maintained by a proxied maintainer and a proxy, and the maintainer goes MIA, the proxy (or Undertakers, in the scenario indicated above) eventually takes care of reassignment and the package becomes m-needed. Now, if we have a package that's maintainer purely via proxied maintainer without a single proxy, there is no fallback for when maintainer goes MIA (or simply roguely ignores bugs). The bugs can be ignored for years, and nobody may even notice.
Doesn't the proxy-maintainers project exist for precisely that purpose? Why can't they be added to metadata?
To address this bug specifically: I do not believe there is any actionable QA violation here. No policy is being violated, and no packages are currently broken. On a personal level, if I notice that Arfrever goes inactive for some long time period, I would certainly reach out to him. Regarding the possibility of my own activity: I have been a developer for several years, and I do not have any intention of retiring for the foreseeable future. To address the more general concern: Maybe write a script to identify inactive maintainers in general? Build a list of email addresses based on metadata.xml entries, and then bump that list up against the git commit log. If you find people that haven't authored anything in some time period, file a bug and ask them for a status update, similar to the typical undertakers process.
(In reply to Ulrich Müller from comment #4) > Doesn't the proxy-maintainers project exist for precisely that purpose? Why > can't they be added to metadata? Proxy-maint exists to serve proxied maintainers who choose to work with us. FWICS Arfrever is not interested in working with us, and I won't force anyone to.
Personally I do not mind if proxy-maint@gentoo.org is listed in metadata.xml IFF it does not restrict my ability to ask any Gentoo developer to review and push patches for given package (including initial addition of given package). However suggesting to add proxy-maint@gentoo.org to metadata.xml is inconsistent with their actual past actions, as seen in bug #687594. I would not want to endanger other Gentoo developers with receiving such threats from proxy-maint@gentoo.org project if they add a new package with proxy-maint@gentoo.org in metadata.xml. I would like to request clarification from proxy-maint@gentoo.org project about their policies / recommendations, since apparently opposite suggestions have been made by their members (mgorny, jstein) in bug #687594 and bug #687660.
https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=5f1bcb1eb47f99243ebbbc72366ebf24fa8fe180 commit 5f1bcb1eb47f99243ebbbc72366ebf24fa8fe180 Author: Mike Gilbert <floppym@gentoo.org> Date: Tue Sep 10 10:56:14 2019 -0400 Add myself as a proxy on packages maintained by Arfrever Signed-off-by: Mike Gilbert <floppym@gentoo.org> dev-db/sqlite/metadata.xml | 5 +++++ dev-libs/protobuf/metadata.xml | 5 +++++ dev-libs/utfcpp/metadata.xml | 5 +++++ dev-util/gyp/metadata.xml | 5 +++++ net-irc/kvirc/metadata.xml | 5 +++++ net-libs/neon/metadata.xml | 5 +++++ net-libs/serf/metadata.xml | 5 +++++ net-misc/cadaver/metadata.xml | 5 +++++ 8 files changed, 40 insertions(+)
