eliminating a cause of my recent rant and argument with Matt

[Kevin.Buckley at ecs.vuw.ac.nz]

>> Would you not consider using a perfectly decent autoconf-like
>> construct
>> such as, in the MPI case,
>>
>> --with-known-mpi-shared
>>
>> given that it is with the knowledge of the shared MPI that you are
>> building ?
>
>     We could do this. It is a bit wordy with the --with however, if it
> is always --with-known why not just eliminate the with part?
>
> Do people use --with-known-- with autoconf?
>

I dont; recall seeing such a usage but thinking about it a bit more
deeply, and possibly echoing another poster, "--enable-something" is
normally used for something optional as in the fuller form

--enable-something=yes

--enable-something=no

whereas options that configure the "something" are often the "--with"
prefix.

For example, from OpenMPI,

--enable-mpi-f90     enable f90 MPI bindings (default: enabled)

and then,

--with-mpi-f90-size=SIZE
                     specify the types of functions in the Fortran 90 MPI
                     module, where SIZE is one of: trivial (MPI-2 ...

--with-f90-max-array-dim=<DIM>
                     The maximum array dimension supported in the F90 MPI
                     bindings (default: $OMPI_FORTRAN_MAX_ARRAY_RANK).

Apologies for the noise,
Kevin

-- 
Kevin M. Buckley                                  Room:  CO327
School of Engineering and                         Phone: +64 4 463 5971
 Computer Science
Victoria University of Wellington
New Zealand