[R-package] fix MM_PREFETCH and MM_MALLOC checks in configure.ac

[jameslamb]

This PR fixes two bugs in the CRAN package

• uses a C++ compiler, not a C compiler, for MM_PREFETCH and MM_MALLOC checks
• removes extra int main(...) scaffolding in MM_PREFETCH and MM_MALLOC checks, which was causing them to fail

More explanation

Look at the diff for R-package/configure and the autoconf documentation on AC_LANG_CONFTEST...autoconf adds that scaffolding for you! So this check was always failing for the CRAN package on Mac and Linux failing for the CRAN on Mac, because the conftest code has an extra int main(...)!!

NOTE: Once CI runs, I'm going to update this and a table here comparing the MM_MALLOC and MM_PREFETCH check results for CRAN vs. CMake builds...I think they'll match now.

Comparison of R 4.0 builds

✅ below means

checking whether <thing> works... yes

❌ means

checking whether <thing> works...no

Latest commit on master: https://github.com/microsoft/LightGBM/actions/runs/338789296

	CRAN	CMake (gcc)	CMake (clang)
Mac	MM_PREFETCH ❌ MM_PREFETCH ❌	MM_PREFETCH ✅ MM_PREFETCH ✅	MM_PREFETCH ✅ MM_PREFETCH ✅
Linux	MM_PREFETCH ✅ MM_PREFETCH ✅	MM_PREFETCH ✅ MM_PREFETCH ✅	MM_PREFETCH ✅ MM_PREFETCH ✅

this PR: https://github.com/microsoft/LightGBM/actions/runs/338808198

	CRAN	CMake (gcc)	CMake (clang)
Mac	MM_PREFETCH ✅ MM_PREFETCH ✅	MM_PREFETCH ✅ MM_PREFETCH ✅	MM_PREFETCH ✅ MM_PREFETCH ✅
Linux	MM_PREFETCH ✅ MM_PREFETCH ✅	MM_PREFETCH ✅ MM_PREFETCH ✅	MM_PREFETCH ✅ MM_PREFETCH ✅

Notes for Reviewers

@StrikerRUS I am really glad you asked me about the compiler used in R-package/configure.ac in #3509 (review). I think I misunderstood your question the first time I read it, but that misunderstanding led me to an investigation that resulted in fixing a bug 😀

@guolinke I think we should try to get this into 3.1.0 (#3484)

[jameslamb]

/gha run r-valgrind

[jameslamb]

valgrind is happy!

https://github.com/microsoft/LightGBM/actions/runs/338827520

[StrikerRUS]

@jameslamb Happy to see this bugfix! Awesome job, thanks a lot! I remember I was very upset seeing this check failing for CRAN.

[github-actions]

This pull request has been automatically locked since there has not been any recent activity since it was closed. To start a new related discussion, open a new issue at https://github.com/microsoft/LightGBM/issues including a reference to this.