Peter,
On Tue, 24 Feb 2004, Peter W. Draper wrote:
> Clearly the presence of an f95 command on my system is causing some
> unintended havoc. I guess you need to prefer *77 compilers over *95 ones
> (but that "--mode=compile f95" looks odd as a parameter to me too).
I'm printing out the libtool manual as I type, as it's clear I'm going
to have to commune with it more deeply than I had hoped. I'm not sure
_exactly_ what's happening here, but I can sort-of see why it's happening.
Autoconf now has two sets of Fortran macros, the `F77' ones and the `FC'
ones; the former are about to be deprecated, I understand, and so I based
the automake support for fpp on the FC macros. There's little difference
at present, except that the FC ones are flexible enough that they can
cope with Fortran95 and will, I think, prefer Fortran95 compilers over
Fortran77 compilers if they find them.
My understanding was that this wouldn't matter, since Fortran77 is
a subset of Fortran95 so that a f95 compiler could grok f77 code.
It does look as if it's mattering here, though I have a recollection that
`unable to infer tagged configuration' means something more specifically
libtoolish. I'll need to get back to this, I'm afraid.
The --mode=compile is right. When you're building shared libraries,
you more-or-less have to let libtool oversee the compilation, too, so
that it can add the right PIC options.
> Tried setting the environment variable F77, but that doesn't seem to have
> gotten me any further. Any configuration options (--with-xxx) that I can
> also try to work around this?
F77 is right for the old F77 macros. The analogue for the FC macros
should be FC, so if you set that at ./configure time, either by
setting FC in the environment or, better, by giving it as a parameter
to configure (./configure FC=/path/to/g77), then it should propagate
through correctly. If you do it the latter way, then ./configure can
check that it's consistent between reconfigures.
See you,
Norman
--
---------------------------------------------------------------------------
Norman Gray http://www.astro.gla.ac.uk/users/norman/
Physics and Astronomy, University of Glasgow, UK [log in to unmask]
|