Alright. I've done a build test in unstable-swh and buster-swh and both builds have succeeded (at least until the point of tests, which will need a new tag). Thanks!

I'll start with a general reasoning about origin visit vs. origin visit state objects in our "conceptual" data model, as it was sprinkled throughout my comment initially.

Reuse anlambert's changes

Possessive forms bar's foo are harder to read than their foo of bar counterparts, and should generally be avoided, when bar is an inanimate object.

In T2346#43057, @vlorentz wrote:

I'd say, let's keep the metadata field for now, just to avoid migrating back and forth.

And if we want to pack it with lots of data, we can switch from semantic 1 to semantic 2 later, which shouldn't be too much trouble.

Thanks for recording this.

octocatalog-diff output:

Apply my own comments: refactor config to use a hiera dict

I just thought of a possible compromise. #ShowerThoughts

Having the full commit history displayed when a patch is applied is very important to me. I agree that the display could be improved. However here's my rationale for showing it in full:

Thanks!

Rebase

Silly newlines

In D2907#70414, @ardumont wrote:

Build is green

I triggered a build back to check whether the great new improvments that @olasd installed would have worked on the diff.
That did not unfortunately ;)

I'm not too sure about making the branch name be an opaque sha256. It's not great from a "end user browsing the archive" point of view.

I'm happy now :)

Actually rebase

Noop change to trigger build