[petsc-dev] should MPICH do this?

[Åsmund Ervik]

> Date: Wed, 3 Sep 2014 14:58:58 -0500
> From: Barry Smith <bsmith at mcs.anl.gov>
> To: For users of the development version of PETSc <petsc-dev at mcs.anl.gov>
> Subject: [petsc-dev] should MPICH do this?
>
> Configuring MPICH version 3.1 with
'--prefix=/Users/barrysmith/Src/petsc/arch-mpich'
> 'CC=/Users/barrysmith/Src/petsc/arch-mpich/bin/mpicc' 'CFLAGS= -fPIC
-Wall -Wwrite-strings
> -Wno-strict-aliasing -Wno-unknown-pragmas -g3 -O0 ' 'AR=/usr/bin/ar'
'ARFLAGS=cr'
> '--disable-cxx' '--disable-fortran' '--disable-fc' '--disable-f77'
'--disable-f90' '--enable-shared'
> '--with-device=ch3:sock' '--without-mpe' '--with-pm=hydra'
'--enable-g=meminit' '--enable-fast'
> Running on system: Darwin anlextwls002-182.wl.anl-external.org 13.3.0
Darwin Kernel Version
> 13.3.0: Tue Jun  3 21:27:35 PDT 2014;
root:xnu-2422.110.17~1/RELEASE_X86_64
> x86_64configure: error:
> --enable-f90 and --disable-f90 are NO longer valid configure options,
> use --enable-fc or --disable-fc instead
>
*******************************************************************************
>
> Note that --disable-fc is also given along with '--disable-f77? and
'--disable-f90?  for backward
> capability. If configure is intended to be scriptable (as opposed to
always run by hand by
> someone) should it really error out in this situation. Now my script
that runs configure on
> MPICH needs to be super complicated and try different options based on
the version of MPICH
> etc etc.
>
> We should make sure we don?t error PETSc stuff out for silly reasons
like this.

I had a similar experience with the PETSc configure flag
"--download-f-blas-lapack" which was renamed to
"--download-fblaslapack". Then trying to configure 3.5.1 with the former
version gave an error message suggesting that I probably wanted the
latter version.
When you've already decided to have this semi-helpful logic in the code,
how about making it fully helpful and just understanding both flags as
the same? We also have a script that configures/makes/installs PETSc,
since we install both optim and debug versions for three different
compilers.

Åsmund

>    Barry

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: OpenPGP digital signature
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20140904/78a4fb6c/attachment.sig>