Petar Mlinarić

It looks like tests fail now because `hypothesis` is producing rank-1 matrices, e.g., ```python from pymor.algorithms.chol_qr import shifted_chol_qr from pymor.vectorarrays.numpy import NumpyVectorSpace A = NumpyVectorSpace(5).ones(2) print(A) Q, R = shifted_chol_qr(A)...

> > Fixing this doesn't seem simple to me. > > Indeed, I looked into it but am not sure how to deal with this. The question is what we...

As I recall, hypothesis runs tests with the zero array. So, if the tests pass, I'm fine with merging.

> > As I recall, hypothesis runs tests with the zero array. So, if the tests pass, I'm fine with merging. > > There's > > ```python > assume(len(U) >...

I guess it might be expected that `to_*` and `from_*` are inverses of each other, but if it's convenient for `to_*` to work with parametric systems, I guess that's fine.

> Maybe it would make sense to add this feature also in the PHLTI case for consistency reasons, what do you think? I think it makes sense to also have...

Indeed. It would also be nice for discretizers to know how to define multiple right-hand sides, such that it would be easy to define Models for testing.

Sounds good to me.

The demo and tests are failing, probably because the `msd` model was changed in the meantime.

This requires some thought on time-stepping, so I postponed it to the next milestone.