[MOAB-dev] Make check failure (5143)

[Jason Kraftcheck]

On 09/01/2011 01:30 PM, Xiabing Xu wrote:
> Yes, this is on fusion. 
> 
> By delving into the problem, I found that the code in iBase_f.h is looking
> for the user predefined macro PTRSIZE. If this macros is not found, then for
> different compilers, define the different size for IMESH_HANDLE_T. In the
> case for searching gfortran compiler, it  is looking for predefined
> macro __SIZEOF_SIZE_T__. However, this macro is not defined by gfortran
> compiler on fusion. 
> 
> The default mpi used on fusion is compiled with intel compiler, in this case
> a different code path is chosen, intel's fortran compiler is used. That
> explains why it works for Rajeev.
> 
> The __SIZEOF_SIZE_T__ is defined by gfortran compiler on mcs machines, I
> guess that's why the nightly tests are not complaining.
> 
> For greater portability, I think the assumption for the predefined macro
>  __SIZEOF_SIZE_T__ in gfortran compiler should be removed.
> 

The code does not assume __SIZEOF_SIZE_T__ nor gfortran, but that isn't
important here.  If you are using the intel compiler and the header is not
working, then presumably something is wrong with the section of that header
corresponding to the intel compiler.  That would be the part beginning with
the line:

  #elif defined(__INTEL_COMPILER)

- jason