@vlorentz I should have started with this... from the nixguix manifest, the integrity is for now only sha256... [1]
So not sure we need to touch the model after all [2], especially since that diff got a tad bigger since you reviewed it...

[2] D8582

$ jq . /var/tmp/sources.json | grep -c sha256
13629
$ jq . /var/tmp/sources.json | grep -c sha384
0
$ jq . /var/tmp/sources.json | grep -c sha512
0

@vlorentz I should have started with this... from the nixguix manifest, the integrity is for now only sha256... [1]
So not sure we need to touch the model after all [2], especially since that diff got a tad bigger since you reviewed it...

Adding tests revealed it missed stuff...

Fix docstring which failed the build!

Address review

Could you use a smaller test file? That one is really big...

Well, that's the sole one i found.

any file would do, you don't need to get one from Guix/Nix

please document which terminal of the grammar you are aiming for. (the current implementation is hash-expression, but I don't know if that's intentional)

ok then ;)

This cannot work as explained time and again.

Add tests that explicitly fail as it's missing the required version in the current
package loader implementation.

til

Fine, i've one comment i'd like others to have a look at though [1] regarding where
that new discovery (interface) code should go. It feels currently a bit off to me that this code
is in loader-core. Loaders are not the sole archive consumers (scanner, webapp, cli, indexer, cooker, ...).

Still a couple of remarks/questions.

1. Isn't this missing an asyncio dependency requirements in /requirements.txt

That's been worked around in the lister to skip those 500 which were breaking the pagination listing.
There is no reply to the upstream bug yet.

one question where i got confused by the old code inline ¯\_(ツ)_/¯ .

"As usual", i don't know how well that module is covered, it'd be good to
have a test around the scan function to check it behaves as we want.

This lgtm overall.

Ingestion ok

Some bower origins visited and ok:

I'm coming late to the battle ¯\_(ツ)_/¯.

one suggestion inline.

one non-blocking suggestion inline.

Use the right number of commits

Rebase and adapt according to recent lister refactoring

Although, i only skimmed through the code. I was mostly sold by the screenshots ¯\_(ツ)_/¯.

I think we can open the linked tasks. They are tagged staff but there is nothing inside preventing it from going public.
@vsellier do you concur about the task ^ ?