Closed simonbaird closed 4 months ago
Draft because I don't know if this is something we want or not. Happy to abandon unless we can decide there's some value in using ubi9/go-toolset
instead of plain ubi9/ubi-minimal
.
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 87.39%. Comparing base (
3ce6dd4
) to head (903a911
). Report is 2 commits behind head on main.
I've seen RPM repositories fail with 500/404, perhaps just for Fedora mirrors, not sure. So the benefit of this could be that we depend on only registry.access.redhat.com rather than that and the RPM repository as well...
I've seen RPM repositories fail with 500/404, perhaps just for Fedora mirrors, not sure. So the benefit of this could be that we depend on only registry.access.redhat.com rather than that and the RPM repository as well...
Hmm yeah, it also means we don't need to figure out the cachi2 prefetch for those rpms.
Hmm yeah, it also means we don't need to figure out the cachi2 prefetch for those rpms.
Not true, since we also microdnf install jq later on.
Why don't we give this a try, I don't see any downsides really...
I think #1608 is slightly preferable, wdyt?
Merged #1608 and abandoned this one.
The --chown=1001:0 was required to avoid an error like this:
fatal: detected dubious ownership in repository at '/build' To add an exception for this directory, call: git config --global --add safe.directory /build
I'm not exactly sure the reason why, but I believe it's because the new base image comes with a user called 'default' with uid 1001. When we copy the source code into the build container it comes in owned by root, and that causes git to be suspicious when it runs as user default.
Ref: https://issues.redhat.com/browse/EC-625