Don't you love it when clang test suite fails because of pthread_create bad address or something and the stack trace points to glibc 
So it's like is clang breaking (scary) or it's glibc (scarier)
Though it only happens with clang-built-clang (and not with gcc-built-clang) so my guess is that it's clang that is broken but
Okay it seems to not happen when I set LIBCLANG_NOTHREADS so probably something happens during thread creation
Also, if I single-step through it with gdb then the problem's gone...
But it still happens if I just press continue through all the breakpoints
So, race condition?
So setting a breakpoint *just* before the instruction that traps to the kernel to do clone(2) seems to slow things down enough that the testcase succeeds, so now I strongly believe that this is indeed a race
Now the question is, how do I track this down 
So what I do know as of now is that the failing test `c-index-test -test-annotate-tokens` is structured like this:
main
-> spawn outerWorker
----> spawn innerWorker
----> join innerWorker
-> joinOuterWorker
The crash happens when it was about to start the inner worker, but that's pretty much all I know as of now
Oh and the test provides control over which threads are spawned; if I tune it so that it only spawn one level of workers, instead of two, it works, regardless of which spawning is disabled
@robryk
> either `pthread_create` or `pthread_attr_setstacksize` failed
This is the fun part: I said this because the error message and gdb disagrees about what function failed; the error message says something like "LIBCLANG FATAL ERROR: pthread_create failed: Bad address", but gdb's backtrace points to `pthread_attr_setstacksize` as the source of failure
@robryk Nothing died, it's more like either `pthread_create` or `pthread_attr_setstacksize` failed and then the entire program just noped out upon getting the return value
The chain is like:
`main` in c-index-test.c -> pthread_create `thread_runner` -> pthread_create `ParseTranslationUnitImpl`
It's the creation of ParseTranslationUnitImpl that fails
If I replaced the `pthread_create`s with a plain function call, then everything would work flawlessly