Proposition for fixing CMake procedure

[jvdp1]

Here is a proposition for fixing the compilation with cmake 3.31.1 introduced by #913 and mentioned in #922

Fixes #922

@jalvesz @perazz

[perazz]

LGTM, thank you @jvdp1. I would wait for @jalvesz's review, as he is the author of that refactor.

[jalvesz]

@jvdp1 I installed CMake 3.31.5, and managed to rebuild without problems. I'm wondering if the problem is the generator under your OS Linux Fedora. Do you have Ninja installed? Would you mind double checking by adding -G Ninja to the cmake build?

The PR looks good to me, would just like to know where the issue comes from as indeed I did not see that problem on Windows or Ubuntu.

[perazz]

On macOS I get (before this PR):

federico@Federicos-MBP stdlib % mkdir build
federico@Federicos-MBP stdlib % cd build 
federico@Federicos-MBP build % cmake ..
-- The Fortran compiler identification is GNU 14.2.0
[...]
-- Build files have been written to: /Users/federico/Documents/GitHub/stdlib/build
federico@Federicos-MBP build % make
[  0%] Generating stdlib_version.f90
[  0%] Generating blas/stdlib_blas.F90
error: Failed to open file '/Users/federico/Documents/GitHub/stdlib/build/src/blas/stdlib_blas.F90' for write [FyppFatalError]
error: [Errno 2] No such file or directory: '/Users/federico/Documents/GitHub/stdlib/build/src/blas/stdlib_blas.F90' [FileNotFoundError]
make[2]: *** [src/blas/stdlib_blas.F90] Error 1
make[1]: *** [src/CMakeFiles/fortran_stdlib.dir/all] Error 2
make: *** [all] Error 2
federico@Federicos-MBP build % cmake --version
cmake version 3.29.2

[jalvesz]

Ok, so it seems the generator added some extra layers. I think we can merge this PR, it is much more robust this way! Thank you @jvdp1

[perazz]

Will merge this as the Intel issue is unrelated to the PR. Thank you @jvdp1 @jalvesz