Dylan Copeland

@IdoAkkerman and @JacobLotz please feel free to push commits to this branch. Your contributions and corrections would be appreciated.

> Very cool! It might be nice to have some comments in `spacing.hpp` for each spacing type describing what the input parameters are, and what the final spacing will be....

One of the new unit tests is failing, only for `macos-latest-make-opt-seq-int32`. Does anyone have suggestions on how to debug the optimized build on macos?

> @dylan-copeland -- This is super cool, but there seem to be a testing error [here](https://github.com/mfem/mfem/actions/runs/6554885708/job/17802536130?pr=3648#step:20:78). Do we understand what is going on? @tzanio Unfortunately I do not understand this...

All the CI tests were passing, but now the AppVeyor build fails. I ran this unit test on linux and mac, even with valgrind, with no errors. Does anyone know...

@tzanio do you have any advice on diagnosing the Windows AppVeyor failure?

@IdoAkkerman I think I addressed your comments so far in this PR. Are there any remaining issues from your review?

Unfortunately, I had to revert to the sign trick, removing the "spacing" keyword, in order to get the Windows CI (appveyor) to pass. The idea to use `seekg` and `tellg`...

I implemented a new NURBS mesh format with a section devoted to spacing formulas. If no knotvectors have a formula, the number of spacing formulas is simply 0. That is,...

Ping @hughcars