Fix METIS compilation error on MinGW (gettimeofday issue)

[Foadsf]

This commit resolves the compilation error in the METIS library when building Elmer FEM on Windows using MinGW. The error was caused by the use of the POSIX function gettimeofday, which is not directly available in Windows.

Changes made:

1. Modified elmergrid/src/metis-5.1.0/GKlib/gk_arch.h to define gettimeofday as mingw_gettimeofday when compiling with MinGW.
2. Updated elmergrid/src/metis-5.1.0/GKlib/timers.c to include gk_arch.h.

These changes allow the METIS library to compile successfully on Windows with MinGW, resolving the "implicit declaration of function 'gettimeofday'" error.

Fixes #489

[mmuetzel]

Thank you for looking into this and proposing a fix.

Adding #include <sys/time.h> in timers.c should be enough (and that would follow the include-what-you-use principle). It should be possible to do that without platform-specific guards. Matching how gettimeofday is used in this file, you could guard the header inclusion like the following:

#ifdef __GNUC__
#  include <sys/time.h>
#endif

I'm not sure if you should call mingw_gettimeofday in your code directly. No changes to gk_arch.h should be necessary.

[mmuetzel]

As I noted in #489, this is likely not MinGW-specific. Instead, it is probably because the latest versions of GCC and LLVM default to C17 where implicit function declarations are an error. (These compilers emitted a warning in earlier versions for that.)

The issue should indeed be fixed by including sys/time.h as proposed in the previous comment.

[tzwinger]

I added include of sys/time.h in timers.c - @Foadsf : can you please try if @mmuetzel 's hint worked

[Foadsf]

Dear @mmuetzel , Thanks for the kind support. Checking out the latest commits on the repo 77bbe42cd96e213af77e4e83453db905e2232f8f, the #489 is now resolved and can be closed. Any chance that I can also communicate with you via email? OR have you on the Discord channel?

Dear @tzwinger , please feel free to reject this PR.

[mmuetzel]

@Foadsf: No worries.
I'm not sure what you are trying to discuss that others aren't allowed to read along. Writing here in issues or PRs is fine with me.
I'll also try to connect to the Discord channel from time to time.

[Foadsf]

@mmuetzel thanks for the response. It is not about secrecy 😅 but that there are others on Discord who, for whatever reason, do not use GitHub. Also to avoid bloating the discussions here on GitHub. 🖖

[tzwinger]

OK, I then hereby close this PR, as it has been resolved with another code change