Step 1 in fixing OpenMP detection in configure to work in cases where "-fopenmp" in required

[fgp]

There seems to be an oversight in the current code.

It doesn't reset R_NO_OPENMP back to 0 before trying to compile test-omp.c with -fopenmp. As a result, even if compilation succeeds, R_NO_OPENMP is still 1 afterwards, and configure thus assumes OpenMP support is not available.

[codecov]

Codecov Report

Merging #4707 (d3f4fff) into master (2791043) will decrease coverage by 0.03%.
The diff coverage is n/a.

❗ Current head d3f4fff differs from pull request most recent head d39b8fc. Consider uploading reports for the commit d39b8fc to get more accurate results

@@            Coverage Diff             @@
##           master    #4707      +/-   ##
==========================================
- Coverage   99.47%   99.43%   -0.04%     
==========================================
  Files          75       73       -2     
  Lines       14808    14512     -296     
==========================================
- Hits        14730    14430     -300     
- Misses         78       82       +4     

Impacted Files	Coverage Δ
src/fmelt.c	99.05% <0.00%> (-0.95%)	⬇️
src/gsumm.c	99.61% <0.00%> (-0.29%)	⬇️
src/fsort.c	95.75% <0.00%> (-0.18%)	⬇️
src/assign.c	99.56% <0.00%> (-0.15%)	⬇️
src/dogroups.c	99.62% <0.00%> (-0.05%)	⬇️
src/ijoin.c	95.45% <0.00%> (-0.02%)	⬇️
src/uniqlist.c	98.25% <0.00%> (-0.01%)	⬇️
R/like.R	100.00% <0.00%> (ø)
R/fmelt.R	100.00% <0.00%> (ø)
R/frank.R	100.00% <0.00%> (ø)
... and 31 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 86a383d...d39b8fc. Read the comment docs.

[mattdowle]

Thanks. Related to #4348 (comment).
This PR is a step in the right direction so merging, but note that the comment # TODO: and then nothing seems to be done with this outcome still stands.
I don't have a Mac so we need someone with a Mac to follow up and get -fopenmp passed through somehow.