Thanks for your PR, @Shlufi, but it doesn't compile.

I was able to make the necessary changes with 331f7bc, which is available on the matrix-eigendecomposition branch. Could you please add some unit tests to ensure this is working as intended?

Sure I should be able to get to that at some point this week.

any updates?

@mattt I made a fork, merged master and resolved conflicts on the matrix-eigendecomposition branch.

Then, I added a commit to fix bugs and added tests. I am ready to PR, but would like some guidance on a few things (if you prefer to just comment on the PR, let me know and I'll make it)

1. _dgeev has 3 returns, the eigenValues and the left and right eigen vectors -- I represented that as a 3-tuple, but got a lint warning. I could make (a) a struct for just this return or (b) follow numpy and just return the right eigenvector (not the left) -- or (c) could suppress the warning for this function.

2. Any of these returns could have complex results (Real + Imaginary) -- so I can't use Matrix<> or [Double]-- instead I just used [(Double, Double)] for vectors and [[(Double, Double)]] for complex matrices. Does that sound ok? There isn't a Complex in the Surge and I didn't want to just add that.

3. I didn't see any standard way to handle errors in the underlying calls -- most things just assert on error (e.g. for when a matrix doesn't have a determinant). So, I did that too.

I made this PR for discussion: https://github.com/mattt/Surge/pull/95

It works, but has the issues outlined above

I replied to 1 in the PR. For 2 that is ok for now but we should add a Complex type. We should move it from Upsurge. For 3 let's throw errors where appropriate. Thanks.

I'm closing this due to inactivity. 💀

If you still feel that Surge should get this functionality feel free to open an updated PR for it. 🙂