Using root as a UID/GID for the superuser isn’t standard, the only standard for this is using UID/GID 0. I found https://bugs.gentoo.org/show_bug.cgi?id=315807 when I grepped eclasses and I wonder why using root as UID and/or GID isn’t an error/warning in QA checks. I modified few ebuild and eclasses in my overlay https://gitlab.com/lanodan/overlay/issues/3 Reproducible: Always Steps to Reproduce: 1. Have a system where UID 0 / GID 0 isn’t root 2. Try to merge packages which uses root:root Actual Results: Packages fails to build because root:root doesn’t exist. Expected Results: Packages manages to build. Use GID&UID 0 instead of root.
Please file separate bugs. At a minimum, we would need one bug for each unique maintainer.
Assigning this one to QA to decide if this is really a bug or not.
Well, I've been fighting this in the past but I guess I didn't open a tracker. The bug I just linked tells what you can expect of Gentoo developers, though.
Oh well thanks gentoo for being a choice distro… :P And I don’t think I’ll try discussing with people that doesn’t want to read standards and don’t seems like wanting to discuss. But well people can change.
Just to be clear, I agree with you. Do you happen to have a standards ref I could use to pursue this at hand?
(In reply to Michał Górny from comment #5) > Just to be clear, I agree with you. Do you happen to have a standards ref I > could use to pursue this at hand? I don't think there is a standard. POSIX only says that UIDs are non-negative integers, and that -1 shall not be valid (which seems redundant). Otherwise, it doesn't say anything about specific UIDs or user names: https://pubs.opengroup.org/onlinepubs/9699919799/ UID 0 is hardcoded in the Linux kernel and in coreutils, so I guess you'd have a hard time trying to use any other ID for the superuser. However, I don't see anything assuming that its name must be root. The FHS mentions a root user, but not sure if that's supposed to be read as "root user" or as "root" user: https://refspecs.linuxfoundation.org/FHS_3.0/fhs/ch03s14.html#ftn.idm236092633504
Looks like these eclasses rely on the root name: cron.eclass fcaps.eclass qmail.eclass toolchain.eclass webapp.eclass
For fcaps we have a patch, but it no longer applies: https://gitlab.com/lanodan/overlay/-/commit/ca5680a61871c420d1d147e206f093636b055993
(In reply to Haelwenn (lanodan) Monnier from comment #4) > Oh well thanks gentoo for being a choice distro… :P > And I don’t think I’ll try discussing with people that doesn’t want to read > standards and don’t seems like wanting to discuss. But well people can > change. Old but that tone was undeserved, sorry. I should have linked to the documentation, I think it was in the linux kernel docs rather than things like the FHS but it's been a lot of years since and I don't think I still have the notes I had. Might try doing this experiment again and would focus more on upstreaming my commits rather than dumping them in my overlay.
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=255c3f853ffa896fb9dc65294fc40cee2682b558 commit 255c3f853ffa896fb9dc65294fc40cee2682b558 Author: Ulrich Müller <ulm@gentoo.org> AuthorDate: 2022-03-26 09:29:52 +0000 Commit: Ulrich Müller <ulm@gentoo.org> CommitDate: 2022-03-28 19:47:18 +0000 webapp.eclass: Use UID 0 instead of root Bug: https://bugs.gentoo.org/595908 Signed-off-by: Ulrich Müller <ulm@gentoo.org> eclass/webapp.eclass | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=367ba7457991d008b011eba0b06455276933bf7c commit 367ba7457991d008b011eba0b06455276933bf7c Author: Ulrich Müller <ulm@gentoo.org> AuthorDate: 2022-03-26 09:28:30 +0000 Commit: Ulrich Müller <ulm@gentoo.org> CommitDate: 2022-03-28 19:47:17 +0000 toolchain.eclass: Use UID 0 instead of root Bug: https://bugs.gentoo.org/595908 Signed-off-by: Ulrich Müller <ulm@gentoo.org> eclass/toolchain.eclass | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=12a54e0eeefa9f2d80cf7eb57b1736134acad7e8 commit 12a54e0eeefa9f2d80cf7eb57b1736134acad7e8 Author: Ulrich Müller <ulm@gentoo.org> AuthorDate: 2022-03-26 09:25:07 +0000 Commit: Ulrich Müller <ulm@gentoo.org> CommitDate: 2022-03-28 19:47:16 +0000 qmail.eclass: Use UID 0 instead of root Bug: https://bugs.gentoo.org/595908 Signed-off-by: Ulrich Müller <ulm@gentoo.org> eclass/qmail.eclass | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=0295d66cf7cbc1b719c1325c1442f9404de2ba69 commit 0295d66cf7cbc1b719c1325c1442f9404de2ba69 Author: Ulrich Müller <ulm@gentoo.org> AuthorDate: 2022-03-26 09:22:52 +0000 Commit: Ulrich Müller <ulm@gentoo.org> CommitDate: 2022-03-28 19:47:16 +0000 cron.eclass: Use UID 0 instead of root Bug: https://bugs.gentoo.org/595908 Signed-off-by: Ulrich Müller <ulm@gentoo.org> eclass/cron.eclass | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=83e510378a6ae9906da52b0c276319f87d490593 commit 83e510378a6ae9906da52b0c276319f87d490593 Author: Haelwenn (lanodan) Monnier <contact@hacktivis.me> AuthorDate: 2016-10-02 10:32:23 +0000 Commit: Ulrich Müller <ulm@gentoo.org> CommitDate: 2022-03-28 19:47:11 +0000 fcaps.eclass: Use UID 0 instead of root Bug: https://bugs.gentoo.org/595908 Signed-off-by: Ulrich Müller <ulm@gentoo.org> eclass/fcaps.eclass | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
Eclasses done. For ebuilds, I think it makes sense to follow comment #1 and file individual bugs. This bug can be used as a tracker.
