Dear Akim Demaille, you wrote on Today:
> See the patch: f77 is preferred. It is more discriminant. But I
> might be wrong, I don't really have a strong Fortran experience :)
All I can say is neither g77 nor any of the other compilers I know
(among which the native ones on IRIX, SOlaris and AIX) supports the
extension "f77". That would mean indirect compilation, slower compilaton,
and more disk usage on basically all platforms.
> | - maintainers need to put their source with some extension.
> | The standard (e.g. ".f" for F77) can't be so bad a choice.
>
> This explicitly means no project can have 77 and 90 sources.
Not quite: People can use the f90 compiler to compile both. Most
f90 compilers are backwards-compatible, even most F95 compilers are.
However, that's not the ideal solution, I know.
> This is what I was thinking about. This script is `compile', shipped
> with Automake.
OK, I see, didn't notice it's existence yet. Perhaps an improved "compile"
would be the solution then.
> I didn't think about the debugging issue.
That's a minor point after all. We could insist that people who want to
debug use a proper compiler.
> Can't we always use this? I mean, how do Make's behave on
>
> me: me
> echo foo
>
> Or just
>
> @F77_INEXT_IS_NOT_f77@.77.@(F77_IN_EXT):
> @F77_INEXT_IS_NOT_f77@ cp $< $@
>
> .$(F77_IN_EXT).o:
> $(F77) $(FFLAGS) ... $<
>
> where @F77_INEXT_IS_NOT_f77@ would evaluate to `#' or nothing.
That's a beautiful solution. But some doubts (maintainers have to rename
all their files, performance and resource usage) remain.
Martin
--
Martin Wilck <[EMAIL PROTECTED]>
Institute for Tropospheric Research, Permoserstr. 15, D-04318 Leipzig, Germany
Tel. +49-341-2352151 / Fax +49-341-2352361
