Remove BLAS.h, Linpack.h and Lapack.h from bindings

[CGMossa]

Inspired from a news item

Headers ‘R_ext/Applic.h’ and ‘R-ext/Linpack.h’ used to include ‘R_ext/BLAS.h’ although this was undocumented and unneeded by their documented entry points. They no longer do so.

We actually don't use the wrappers to BLAS.h, Lapack.h or even Linpack.h, so we could remove them and improve clarity of bindings. If we were to include them, we'd actually need to write wrappers on top of them, as they aren't directly applicable.

Diff on removing:

• BLAS.h 7a0b7ed (-6K+10 LOC)
• Linpack.h 20d4f45 (no change)
• Lapack.h dddf24e (+4-88)

[CGMossa]

/bindings

[CGMossa]

/bindings

[CGMossa]

/bindings

[yutannihilation]

Just in case you don't notice.

• I believe libR-sys doesn't use R-ext/Linpack.h, so you don't need to exclude it.
• libR-sys does use R_ext/Applic.h, and the BLAS removal was already reflected by Update bindings #182. But, this happens only on R-devel, and I have no strong opinion on removing them from the prior versions as well.

Also, I'm feeling this reasoning is not very true. These BLAS functions are independent from SEXP, so I don't think it's extendr's responsibility to provide such wrappers anyway.

We actually don't use the wrappers to BLAS.h, Lapack.h or even Linpack.h, so we could remove them and improve clarity of bindings. If we were to include them, we'd actually need to write wrappers on top of them, as they aren't directly applicable.

Personally, I'd vote for removing R_ext/Applic.h.

[CGMossa]

Thanks! As always, very good feedback! So I wanted to make a "Remove Applic.h" next, but I remembered why (apparently I added it), it is because it contains some routines that aren't Fortran ones. Maybe #199 is a better solution to this?

[CGMossa]

Closed in favour of #199