Make benchmarks compiling in clang-cl

[AlexGuteniev]

Towards #5479

Most of the time it is warning about deprecated benchmark::DoNotOptimize for constants.

🏗️ Build steps

Build STL as usual:

cmake --preset x64
cmake --build --preset x64

Then add -DCMAKE_CXX_COMPILER=clang-cl to benchmark build:

cmake -B out\bench -S benchmarks -G Ninja -DSTL_BINARY_DIR=out\x64 -DCMAKE_CXX_COMPILER=clang-cl
cmake --build out\bench

[frederick-vs-ja]

But I wonder why changes elsewhere are relevant.

[AlexGuteniev]

@frederick-vs-ja the rest are just const that can potentially be optimized, so DoNotOptimize is deprecated

[StephanTLavavej]

Thanks! I cleaned up pre-existing weirdness in priority_queue_push_range.cpp (instead of trying to make Clang happy with it), and more significantly I added these builds to Azure Pipelines. This is very important because otherwise it will immediately bit-rot.

[StephanTLavavej]

I'm mirroring this to the MSVC-internal repo - please notify me if any further changes are pushed.

[StephanTLavavej]

Thanks for making the benchmark directory more portable and conformant! 😻 🚢 🚀