Fixed gauss_fir to work with numpy-2

[yarda]

It's related to #1149

[andynoack]

I have fixed the CI pipeline (macos-12) in the master branch now.

[yarda]

I have fixed the CI pipeline (macos-12) in the master branch now.

Could you restart the CI? IMHO the failure is unrelated to the PR.

[yarda]

Context of the proposed PR:
The problem observed is related to the numpy-2 change in data type promotion, which is now according to the NEP 50, i.e. the float_32 was promoted to float_64 which maps to the C type double which causes the gauss_fir to traceback. The proposed change is compatible with numpy-1.

[andynoack]

I have fixed the CI pipeline (macos-12) in the master branch now.

Could you restart the CI? IMHO the failure is unrelated to the PR.

You have to pull the CI fix first, at the moment the CI is based on the old config.

[yarda]

You have to pull the CI fix first, at the moment the CI is based on the old config.

OK, NP, thanks for the info, I will rebase it.