H5Part development and discussion

[Achim Gsell <achim AT cybercity.ch>]

On Friday 01 December 2006 23:23, Kurt Stockinger wrote:
> Hi,
>
> I just noticed that H5PartGetNumParticles() seems to return
> the wrong number of particles. It's always off by one, i.e.
> n-1 rather than n. I've tested the program H5PartTest with 4
> processes and 5 particles each.

The bug seems to be in the function H5PartSetView(), not in 
H5PartGetNumParticles()! The view has

end - start

particles. If we have 5 particles:

start = 0
end = 4
end - start = 4

The "memshape" is created with 

dims = end - start 

i.e. 4 particles.

I found this odd behavior last summer, but forgot to ask John 
whether this is intentional (which I cannot imagine) or I miss 
something and to fix it if not.  May be it is time to start 
using a bugtracker (we can use Bugzilla @svn.psi.ch).

cheers 

Achim