Bump ManifoldsBase to 1.0

[kellertuer]

I am not sure we can keep the 0.15, but for now I also do not understand the error messages of the tests failing. Something with Duals, and I never found the time to read up on all the dual variables magic.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 95.69%. Comparing base (b8c3464) to head (6a21119).
Report is 12 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #55      +/-   ##
==========================================
+ Coverage   95.46%   95.69%   +0.22%     
==========================================
  Files          21       13       -8     
  Lines         375      395      +20     
==========================================
+ Hits          358      378      +20     
  Misses         17       17              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[kellertuer]

Nice! Locally this failed with conversion errors between Float and Duals :D

[mateuszbaran]

Are you sure this ran with ManifoldsBase.jl 1.0? I can see "⌅ [3362f125] ManifoldsBase v0.15.24" in the CI logs.

[mateuszbaran]

This couldn't run with 1.0 because Manifolds.jl is a test dependency and there is no registered Manifolds.jl that supports ManifoldsBase v1

[kellertuer]

Oh, I did not see that. Can you than take a look and see what is going wrong?

I am already quite confused what would be the right order to update. For me it seems we would have to register Manifolds.jl and ManifoldDiff.jl at the same time which is not possible?

[kellertuer]

Yes...and Manifolds.jl tests currently can not run because they require ManifoldDiff to be already bumped to a next version.

That is exactly the deadlock I had locally yesterday, So we have to resolve this circular dependency.

[mateuszbaran]

IMO ManifoldDiff.jl should be first and we have to accept CI failures until Manifolds.jl is updated. But we need to be 100% sure ManifoldDiff.jl actually works with ManifoldsBase.jl 1.0.

[kellertuer]

Could you check that then? There are two errors appearing with conversion from duals to floats or so, I do not understand them.

[mateuszbaran]

OK, I will check that.

[kellertuer]

In the long run we should try to avoid the circular dependency. I would prefer this here not to depend on Manifods.jl

[mateuszbaran]

I agree. This package should not depend on Manifolds.jl for tests.

[kellertuer]

Since tests run locally fine for 1.0 and here (missing Manifolds version) fine for the previous one as well – can we just register and then it is fine?

[mateuszbaran]

Yes, it should be fine now to register a new version of ManifoldDiff.jl