You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
While I've certainly pitched this as a feature, I think
we really want something like --force-older-timestamp - basically
error out if the timestamps on one or more input repos were older.
Basically I think we need to do rpmmd-repo pinning by timestamps. I think things are mostly OK today if one fetches the rpmmd over https://and doesn't do rpm-ostree cleanup -m. (But that remains to be determined)
Fixing the second problem means that we should reject (or ignore) repos with older timestamps for layered packages? And actually implementing this means that we may need to change to "repo pin" layered packages by default. That has interesting UI implications.
The text was updated successfully, but these errors were encountered:
Or a different approach I was thinking here is to just do timestamp checking for the repos vs the final commit that we already write. That would be highly symmetric with the libostree side.
See #1079 (comment) which is for the server side of this:
Also related is ye olde is my system-up-to-date 🆕?.
Basically I think we need to do rpmmd-repo pinning by timestamps. I think things are mostly OK today if one fetches the rpmmd over
https://
and doesn't dorpm-ostree cleanup -m
. (But that remains to be determined)Fixing the second problem means that we should reject (or ignore) repos with older timestamps for layered packages? And actually implementing this means that we may need to change to "repo pin" layered packages by default. That has interesting UI implications.
The text was updated successfully, but these errors were encountered: