mhostetter

Thanks for the pull request. > I needed a solver for linear systems over GF2 for simple codes decoding, but the library only has a solver that just inverts the...

> I changed solve to lstsq for overloading, but saved galois.GF2.solve. Anyway, there is no such solver for other fields in the library. Which algorithm is implemented here? It's likely...

What issues are you having with the JIT code? Do you wish to remove the `numba` dependency completely? Or simply not use `numba` for the arithmetic? See the [Compilation Modes](https://mhostetter.github.io/galois/latest/basic-usage/compilation-modes/#explicit-calculation)...

Thanks for the report. While we should look into this, here's a quick workaround. Since your field is very large, the `ufunc_mode` will be `"python-calculate"`, meaning no LUTs/etc will be...

I can reproduce the bug. Pickling/unpickling works inside the same Python session. ```python In [7]: with open('out.pkl', 'wb') as f: ...: pickle.dump(GF, f) ...: In [8]: with open('out.pkl', 'rb') as...

I wasn't aware `np.linalg.matrix_rank()` accepted an arbitrary number of dimensions. In theory, yes. All of [these JIT'd functions](https://github.com/mhostetter/galois/blob/main/src/galois/_domains/_linalg.py) would need a loop that iterates over the extra dimensions. Perhaps the...

I'm open to adding support for this. My time is a bit limited at the moment. If you'd like to submit a PR, I can work with you to get...

I agree this would be nice. However, currently SciPy doesn't support subclasses of `ndarray`. Also, there are no hooks in `np.ndarray` that can catch and override SciPy function calls. If...

```python import numpy as np import galois x = galois.GF2([1, 0, 1]) print(x) x = np.bitwise_or(x, 0b10) print(x) ``` ``` $ python3 test_bitwise_or.py [1 0 1] [3 2 3] #...

Maybe use `_view()` instead of `_view_without_verification()` for operations that could expand the bit depth.