Reorganized organization of checkNvector routines

[drreynolds]

This PR cleans up how ARKODE and its steppers check for required N_Vector routines. Previously, ARKODE required routines that were only actually needed when certain options were configured, and each stepper performed redundant checks on the supplied N_Vector, when in reality, no stepper utilized routines that were not already required by the main ARKODE infrastructure.

In this PR:

1. Stepper-specific N_Vector operation checks have been removed.
2. ARKODE performs an initial check on the supplied N_Vector for operations that are always required.
3. Individual ARKODE utilities perform their own N_Vector checks when they are configured.
4. ARKODE then performs a final check after user configuration is complete to check any remaining N_Vector operation requirements.

[gardner48]

Other than the comment here about checking for clone this looks good

[gardner48]

Minor change I should have noticed before

[drreynolds]

I don't understand why the "spelling" check just failed. The patch file that it produces "fixes" numerous variable names in completely unrelated files (examples/cvodes/parallel/cvsAdvDiff_ASAp_non_p.c, examples/kinsol/CUDA_mpi/kin_em_mpicuda.cu, src/cvodes/cvodea.c, and src/idas/idaa.c).

Did something change in the CI that could have caused this?

[Steven-Roberts]

I don't understand why the "spelling" check just failed.

codespell had an update today which caught a few more things. I have a fix in #636 so you can just wait for that to be merged. It might be good to select a fixed version of codespell and manually update it as need to avoid issues like this.

[drreynolds]

I don't understand why the "spelling" check just failed.

codespell had an update today which caught a few more things. I have a fix in #636 so you can just wait for that to be merged. It might be good to select a fixed version of codespell and manually update it as need to avoid issues like this.

Actually, the PR had no typos to fix, it's just failing the CI due to the codespell issue. I don't see how this PR can be merged before #636, given that the "spelling" check is required before any PR can be merged, so it can just wait.

[Steven-Roberts]

Yes, codespell recognizes lins and indx as typos now which are used in code totally unrelated to this PR.