Re: Parallel and MPI with default branch

[kingcrimson]

> On 22 Jan 2018, at 17:26, Sebastian Schöps <address@hidden> wrote:
> 
> Hello Carlo, 
> 
>> If you still get the same error with a recent snapshot of the package you 
>> will have to provide more details
>> in order to help me debug. but, at first sight, this appears to be a problem 
>> related to your MPI installation
>> rather than Octave.
> 
> 
> I recompiled open-mpi (via homebrew) and tried again. Still no luck. I attach 
> the complete error message below. 

From the complete log the errors are more clear:

MPI_Send.cc:98:26: error: no viable overloaded '='
                 retval = info;
                 ~~~~~~ ^ ~~~~
/usr/local/Cellar/octave/HEAD-c241918d4c92/include/octave-4.3.0+/octave/ov.h:358:17:
 note: candidate function not viable: no known conversion from 'Array<int>' to 
'const octave_value' for 1st argument
 octave_value& operator = (const octave_value& a)


MPI_Recv.cc:79:15: error: no matching function for call to 'MPI_Get_count'
             MPI_Get_count (&status, MPI_CHAR, &num);
             ^~~~~~~~~~~~~
/usr/local/Cellar/open-mpi/3.0.0_2/include/mpi.h:1477:20: note: candidate 
function not viable: no known conversion from 'octave_idx_type *' (aka 'long 
long *') to 'int *' for
     3rd argument
OMPI_DECLSPEC  int MPI_Get_count(const MPI_Status *status, MPI_Datatype 
datatype, int *count);


MPI_Recv.cc:91:19: error: no matching function for call to 'install_types'
                 install_types ();    
                 ^~~~~~~~~~~~~
/usr/local/Cellar/octave/HEAD-c241918d4c92/include/octave-4.3.0+/octave/ov.h:1633:27:
 note: candidate function not viable: requires 1 argument, but 0 were provided
extern OCTINTERP_API void install_types (octave::type_info&);


this looks like it is due to some changes that occurred in the interpreter, 
after the last time I updated my build.
I'll see what I can do about this next week when the semester is over ...

> Shall I open a bug report at savannah or bitbucket? 

Not needed, thanks.

> Thanks,
> Sebastian

c.


> P.S.: I am trying to finish a nice parareal implementation :) 

I immagined that ;)