Document difference between Comm.Barrier() and Comm.barrier()

There are two methods Comm.Barrier() and Comm.barrier(), which seem to be implemented differently:

https://github.com/mpi4py/mpi4py/blob/ddefbb790d14777cb322ea9007d61345f81f2099/src/mpi4py/MPI/Comm.pyx#L668-L675

https://github.com/mpi4py/mpi4py/blob/ddefbb790d14777cb322ea9007d61345f81f2099/src/mpi4py/MPI/Comm.pyx#L1557-L1560

I guess one releases the GIL and the other doesn’t?

The upper vs. lower case convention seems to imply an analogy with the generic Python object interface vs. buffer-like object interface (e. g. Comm.Bcast() vs Comm.bcast()), but since MPI_Barrier() doesn’t communicate any data, it’s not obvious what the difference is here (if any). The docstrings unfortunately don’t have any more information either (“Barrier synchronization” vs. “Barrier”).

[Socob]

Ah, right, I didn’t consider that subclassing Comm by the user could be a relevant/supported use case. I was just trying to say that the different method bodies surprised me, since they’re functionally equivalent (a better example would have been to say that I would’ve expected both methods to have the same body, e.g.

    cdef MPI_Comm comm = self.ob_mpi 
    return PyMPI_barrier(comm)

).

[leofang]

to say that I would’ve expected both methods to have the same body

During the review of #126 I actually have the same question in mind 😄