Jonathan Lebon

> ``` > === RUN fcos.upgrade.basic > === RUN fcos.upgrade.basic/setup > === RUN fcos.upgrade.basic/upgrade-from-previous > === RUN fcos.upgrade.basic/upgrade-from-current > --- FAIL: fcos.upgrade.basic (50.43s) > --- PASS: fcos.upgrade.basic/setup (2.71s) > ---...

Hmm, and you're sure you're using new enough cosa that has the newer bootc as well?

Can you dig into this more? E.g. compare the output of `ostree show --raw $merge_commit` between the two imports to show the raw commit objects (which will also include the...

I think https://github.com/bootc-dev/bootc/issues/1637 could also be a factor here that breaks reproducibility.

I think that could allow us to drop [the `ostree-cmdline.sh` hack](https://github.com/coreos/fedora-coreos-config/blob/75a2ffd8745f8194c754435f63d6b29b2c9b0d1f/overlay.d/05core/usr/lib/dracut/modules.d/35coreos-live/ostree-cmdline.sh) in our CoreOS live environment. Well... maybe not. I guess we'd still need to inject at least `ostree=current`.

bootc-1.6.0-1.fc44 is in rawhide now. I've tagged bootc-1.6.0-1.fc42 into the continuous tag, so it should be in the next cosa build.

> Some comments. > > I think there are a few things we need to iron out before we can really move forward with this: > > 1. supporting both...

Would be good to sanity-check el9-based SCOS before merging this.

For the longer term RFE, one idea I had in https://github.com/coreos/rpm-ostree/issues/4688#issuecomment-1800250818 was to add the concept of "builder images". This is not novel of course; it's similar to OpenShift S2I...

> > For the longer term RFE, one idea I had in [coreos/rpm-ostree#4688 (comment)](https://github.com/coreos/rpm-ostree/issues/4688#issuecomment-1800250818) was to add the concept of "builder images". > > We should probably split off longer...