This might be quite a nasty hack, so I'd like to solicit some feedback
on what the best course to take with it is. I've symlinked the benchmark lib
and XRay microbenchmark sources into LNTBased/XRay. When the test is run, cmake
is invoked and the retref-bench binary is built. When it is run, its output is
parsed and the appropriate TestSamples objects are returned.
This is to accomplish a TODO item in https://reviews.llvm.org/D32272#794759 :
"In the future teach LNT to understand the Google Benchmark output."