Instead of trying to download artifacts and then break on ingestion. Try to
prevent spurious downloads which cost too many wasted resources (both upstream
servers and our own infra) + I don't know how many noisy issues in
sentry ¯\_(ツ)_/¯.
Ideally, we should improve our archive support. In the mean time, this approach
allows more sensible treatment than the current approach (see previous point).
This will also allows to incrementally add support, and demonstrating it by
removing the new supported archive pattern in the current regexp in next diffs ;)
Note: Expects current tests to fail as there is noisy and unsupported
extensions in our test dataset (which I did not yet take care of, this
impacts most like both load status and/or hash computations depending
on the answer to the following question).
Needs now to determine what do we want to do with those:
- Do we keep those references in the snapshot with None branch?
- Do we remove them altogether? [1]
[1] ~> that's the current approach started here, open for discussion as I don't
really what's more sensible here
Related to T2510