Tom Clune

Not sure about "compiler can optimize". The algorithm is baked into the implementation of STL vectors. The vector interface does not really give the compiler any hint of how the...

No, or at least not as written. There is an inherent conflict between privacy/encapsulation and testability. A test must be able to invoke the logic to be tested. So here...

Hmm. The main obstacle with using pFUnit on co-array code is that you must ensure that such tests span all of MPI comm world. So if you are running on...

Can you redo the make step with `VERBOSE=1`? And does the build succeed if you have `FC=gfortran` instead? I still cannot envision any error other than the fact that `loader.mod`...

OK - the pure gfortran case got past the earlier problem, which is good. My best guess is that gfortran somehow changes the format of `.mod` files when compiling with...

OK - so first, you will want to have `use pfunit` instead of `use funit` in any parallel tests. And then each test procedure needs to have a "this" or...

OK - I suspect this problem is because CAF is launching MPI in the background and the command line has already launched MPI. You'll need to ask someone more familiar...

The approach you are advocating cannot be done with traditional MPI nor OpenMP which don't really support the notion of recovery after a crash. In theory fault tolerant MPI (or...

@kurtsansom I think there may be 2 or maybe 3 independent issues here, but I don't know much about FRUIT beyond its rough equivalence to pFUnit. The issue I opened...

Side comment: I've been a bit negligent in pushing the snake_case throughout the source code. I'd best do that soon, or it will be hard to argue against retaining CamelCase...