Move test-suite rpm configuration away from /root/.rpmmacros #3521
Labels
test
Testsuite-related
Comments
pmatilai
added a commit
to pmatilai/rpm
that referenced
this issue
Jan 16, 2025
My reading of https://docs.docker.com/build/building/variables/#scoping suggests that we could now remove the second "ARG PREFIX" from the make stage because it should be inherited in the FROM base thing, but s*** breaks if I remove it, so we'll just *shrug* and let it be I guess. This is the first step to fixing rpm-software-management#3521.
pmatilai
added a commit
to pmatilai/rpm
that referenced
this issue
Jan 16, 2025
The second step to fixing rpm-software-management#3521
pmatilai
added a commit
to pmatilai/rpm
that referenced
this issue
Jan 16, 2025
The third step to fixing rpm-software-management#3521
pmatilai
added a commit
to pmatilai/rpm
that referenced
this issue
Jan 16, 2025
Signing key configuration is obviously something that belongs to the user's macros, but lets get on with the times and use the new XDG config directory for that. The remaining references to ~/.rpmmacros in rpmmacro.at are related to testing behavior with the old-style configuration and need to be preserved. This is the fourth and final step to rpm-software-management#3521. Fixes: rpm-software-management#3521
pmatilai
added a commit
to pmatilai/rpm
that referenced
this issue
Jan 16, 2025
Signing key configuration is obviously something that belongs to the user's macros, but lets get on with the times and use the new XDG config directory for that. The remaining references to ~/.rpmmacros in rpmmacro.at are related to testing behavior with the old-style configuration and need to be preserved. This is the fourth and final step to rpm-software-management#3521. Fixes: rpm-software-management#3521
pmatilai
added a commit
to pmatilai/rpm
that referenced
this issue
Jan 17, 2025
This override exists to make system rpmdb accessible in eg "make shell" but it causes complications wrt the prefix (see below + next commits) and also masks test bugs (see previous commit). We can resurrect this functionality later on if we find solutions that avoid the issues. This is the first step to fixing rpm-software-management#3521.
pmatilai
added a commit
to pmatilai/rpm
that referenced
this issue
Jan 17, 2025
The second step to fixing rpm-software-management#3521
pmatilai
added a commit
to pmatilai/rpm
that referenced
this issue
Jan 17, 2025
The third step to fixing rpm-software-management#3521
pmatilai
added a commit
to pmatilai/rpm
that referenced
this issue
Jan 17, 2025
Signing key configuration is obviously something that belongs to the user's macros, but lets get on with the times and use the new XDG config directory for that. The remaining references to ~/.rpmmacros in rpmmacro.at are related to testing behavior with the old-style configuration and need to be preserved. This is the fourth and final step to rpm-software-management#3521. Fixes: rpm-software-management#3521
pmatilai
added a commit
to pmatilai/rpm
that referenced
this issue
Jan 17, 2025
This override exists to make system rpmdb accessible in eg "make shell" but it causes complications wrt the prefix (see below + next commits) and also masks test bugs (see previous commits). We can resurrect this functionality later on if we find solutions that avoid the issues. This is the first step to fixing rpm-software-management#3521.
pmatilai
added a commit
to pmatilai/rpm
that referenced
this issue
Jan 17, 2025
…iner This override exists to make system rpmdb accessible in eg "make shell" but it causes complications wrt the prefix (see below + next commits) and also masks test bugs (see previous commits). We can resurrect this functionality later on if we find solutions that avoid the issues. Instead of trying to inject the system rpmdb path into the test-image in a prefix-independent manner, set + create an empty, test-suite specific rpmdb path for all the tests to use by default. This dodges the pesky prefix question and accidental test-dependencies on system rpmdb. It should also allow us to optimize away a bunch of RPMDB_INIT's in the tests later on, but that's an exercise for another day. This is the first step to fixing rpm-software-management#3521.
pmatilai
added a commit
to pmatilai/rpm
that referenced
this issue
Jan 17, 2025
…iner This override exists to make system rpmdb accessible in eg "make shell" but it causes complications wrt the prefix (see below + next commits) and also masks test bugs (see previous commits). We can resurrect this functionality later on if we find solutions that avoid the issues. Instead of trying to inject the system rpmdb path into the test-image in a prefix-independent manner, set + create an empty, test-suite specific rpmdb path for all the tests to use by default. This dodges the pesky prefix question and accidental test-dependencies on system rpmdb. It should also allow us to optimize away a bunch of RPMDB_INIT's in the tests later on, but that's an exercise for another day. This is the first step to fixing rpm-software-management#3521.
pmatilai
added a commit
to pmatilai/rpm
that referenced
this issue
Jan 17, 2025
…iner This override exists to make system rpmdb accessible in eg "make shell" but it causes complications wrt the prefix (see below + next commits) and also masks test bugs (see previous commits). We can resurrect this functionality later on if we find solutions that avoid the issues. Instead of trying to inject the system rpmdb path into the test-image in a prefix-independent manner, set + create an empty, test-suite specific rpmdb path for all the tests to use by default. This dodges the pesky prefix question and accidental test-dependencies on system rpmdb. It should also allow us to optimize away a bunch of RPMDB_INIT's in the tests later on, but that's an exercise for another day. This is the first step to fixing rpm-software-management#3521.
dmnks
pushed a commit
that referenced
this issue
Jan 17, 2025
…iner This override exists to make system rpmdb accessible in eg "make shell" but it causes complications wrt the prefix (see below + next commits) and also masks test bugs (see previous commits). We can resurrect this functionality later on if we find solutions that avoid the issues. Instead of trying to inject the system rpmdb path into the test-image in a prefix-independent manner, set + create an empty, test-suite specific rpmdb path for all the tests to use by default. This dodges the pesky prefix question and accidental test-dependencies on system rpmdb. It should also allow us to optimize away a bunch of RPMDB_INIT's in the tests later on, but that's an exercise for another day. This is the first step to fixing #3521.
pmatilai
added a commit
to pmatilai/rpm
that referenced
this issue
Jan 20, 2025
The second step to fixing rpm-software-management#3521
pmatilai
added a commit
to pmatilai/rpm
that referenced
this issue
Jan 20, 2025
Keyring type is not something that would be user-specific in a real world scenario anyhow. The third step to fixing rpm-software-management#3521
pmatilai
added a commit
to pmatilai/rpm
that referenced
this issue
Jan 20, 2025
Signing key configuration is obviously something that belongs to the user's macros, but lets get on with the times and use the new XDG config directory for that. The remaining references to ~/.rpmmacros in rpmmacro.at are related to testing behavior with the old-style configuration and need to be preserved. This is the fourth and final step to fixing rpm-software-management#3521.
ffesti
pushed a commit
that referenced
this issue
Jan 20, 2025
Keyring type is not something that would be user-specific in a real world scenario anyhow. The third step to fixing #3521
ffesti
pushed a commit
that referenced
this issue
Jan 20, 2025
Signing key configuration is obviously something that belongs to the user's macros, but lets get on with the times and use the new XDG config directory for that. The remaining references to ~/.rpmmacros in rpmmacro.at are related to testing behavior with the old-style configuration and need to be preserved. This is the fourth and final step to fixing #3521.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
We're currently using /root/.rpmmacros for some rpm configuration things in the test-suite because that avoids dealing with changing prefix, but this is now blocking various other things such as #3467.
The text was updated successfully, but these errors were encountered: