The test logic for running libc++ tests only looks to see if /usr/include/c++/v1 exists. This adds a fallback for including libc++ tests as long as $(CC) -stdlib=libc++ works.
This seems ok. The reason these checks are written this way is because we used to have the xcode build which bypassed cmake and lit, so this was the only place to implement that centrally. Nowadays, we could probably come up with a simpler way to do these things, but I don't have a clear idea on how to do that yet...
FYI, the check didn't work because this isn't exercising anything in libc++, it really only checks that -stdlib=libc++ is a valid compiler flag. 786b6db8e6fd87fb82f2ad3e94e20c5c9cf9c4e4 fixes this by adding a libc++ include, which fails the command if libc++ is not installed where clang can find it.