-
Notifications
You must be signed in to change notification settings - Fork 165
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
cmdlib: use repo-packages
for overrides/rpm
#2954
Conversation
In addition to a lockfile, we can be even stronger here by using repo-packages to tell rpm-ostree that the overridden packages must come from the generated overrides repo. This also allows one to override RPMs which themselves are specified as `repo-packages` in the manifests without in most cases having to edit them. Related: coreos/rpm-ostree#3789
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Seems sane to me, though it's worth noting that (AFAIK) CI doesn't cover any of the "local dev" flows like adding to overrides/rpm
.
} | ||
] | ||
} | ||
for line in sys.stdin: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I know this isn't a new issue, but hopefully in the future we can drop this python-in-shell-script in favor of Go
Indeed re. CI (though rpm-ostree's CI at least uses
|
Yep I assumed you did, I personally am totally fine with local-only testing for now for this type of stuff. |
The recent changes to coreos-assembler in coreos/coreos-assembler#2954 broke this, and I think what we were trying to do here should be fixed.
The recent changes to coreos-assembler in coreos/coreos-assembler#2954 broke this, and I think what we were trying to do here should be fixed.
As mentioned in coreos/rpm-ostree#3802, this actually changes the semantics of So this breaks workflows where people just dump the output of I think I actually prefer the new semantic because it's conceptually cleaner and easier to understand, but I'm cool with reverting it if it's too confusing. |
Ehh. I'm not a big fan of the new semantic. For example I'm trying to test a development SELinux build. I download the rpms.zip file and extract it to the
But it's not very obvious to me what's going on. I need to manually inspect the downloaded rpms and filter through them to find the ones I don't want. What problem was this solving? Could we solve it a different way or add a new "mode" for the new behavior? |
That's in the commit message. :)
Yup, follow-up in #2969! |
In addition to a lockfile, we can be even stronger here by using
repo-packages to tell rpm-ostree that the overridden packages must come
from the generated overrides repo.
This also allows one to override RPMs which themselves are specified as
repo-packages
in the manifests without in most cases having to editthem.
Related: coreos/rpm-ostree#3789