Support nvc++ and msvc as nvcc host compilers

[robertmaynard]

Continues the work of #1884

Extends the nvcc compiler to understand the unique requirements of using nvc++ or cl.exe ( msvc ) as the host compiler.

Due to the fact that nvc++ pre-processing always has force includes the compiler detection can't presume that the first line of the file will hold the compiler id string. So we instead search for lines starting with a unique starting text.

[sylvestre]

plenty of new tests, nice :)

[robertmaynard]

I am currently working on adding nvc and nvc++ as supported compilers ( for C and C++ ) which also requires the changes to compiler type detection.

If needed I can separate the commit that changes those parts of compiler.rs to a new PR if needs a more in-depth review

[codecov-commenter]

Codecov Report

Attention: 142 lines in your changes are missing coverage. Please review.

Comparison is base (ac8c225) 29.88% compared to head (0fb1922) 29.77%.

❗ Current head 0fb1922 differs from pull request most recent head 957bfe8. Consider uploading reports for the commit 957bfe8 to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1889      +/-   ##
==========================================
- Coverage   29.88%   29.77%   -0.11%     
==========================================
  Files          49       49              
  Lines       17538    17699     +161     
  Branches     8443     8518      +75     
==========================================
+ Hits         5241     5270      +29     
- Misses       7214     7311      +97     
- Partials     5083     5118      +35     

Files	Coverage Δ
src/test/tests.rs	33.33% <0.00%> (ø)
src/compiler/compiler.rs	36.73% <52.63%> (+0.57%)	⬆️
src/compiler/nvcc.rs	41.17% <34.00%> (-1.37%)	⬇️
tests/system.rs	26.16% <10.00%> (-4.84%)	⬇️

... and 11 files with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[miscco]

@sylvestre it seems those CI failures are from infrastructure and not this PR.

Can we rerun the failing ones?

[miscco]

for what its worth, looks good to me

[sylvestre]

@robertmaynard @miscco I guess you are colleagues working on this together, right?

Maybe we should discuss (no agenda but to know each others)
if you are interested, my email is s@mozilla.com

cheers

[miscco]

Super happy to talk, sccache is essential for us to be able to run a CI on github PRs for https://github.com/NVIDIA/cccl