[Rd] alpha, portable use

[Douglas Bates]

On 9/25/06, Paul Gilbert <pgilbert at bank-banque-canada.ca> wrote:
> I am still confused about this (and it is still happening with R-beta).
> Writing R Extensions suggests I need a Makefile or Makevars in my
> package,  but that has not been the case previously.  Is there now a
> requirement that all packages need Makefiles or Makevars if there is
> fortran to be compiled?  This is only happening on one of my systems.
> Building R and make check work fine on that system, but it seems that
> not all the information gets passed along to package compiles.
>
> (BTW, this is just a warning, but Kurt suggested we try to eliminate
> warnings.)

>From the warning it seems that you have a Makevars file in the src
directory for your package.  The change in R-2.4.0 is that packages
that did have a Makevars file of the form

PKG_LIBS=$(LAPACK_LIBS) $(BLAS_LIBS)

should change that to

PKG_LIBS=$(LAPACK_LIBS) $(BLAS_LIBS) $(FLIBS)

>
> Paul Gilbert
>
> Prof Brian Ripley wrote:
> > On Wed, 20 Sep 2006, Paul Gilbert wrote:
> >
> >> When I build one of my packages with alpha from yesterday I am getting
> >>
> >> * checking for portable use of $BLAS_LIBS ... WARNING
> >> apparently missing $(FLIBS) in 'PKG_LIBS=$(LAPACK_LIBS) $(BLAS_LIBS)'
> >>
> >> Is this something I should worry about?  (Possibly I got this before and
> >> didn't notice.)
> >
> > Yes, please do check Writing R Extensions ....
> >
> ====================================================================================
>
> La version française suit le texte anglais.
>
> ------------------------------------------------------------------------------------
>
> This email may contain privileged and/or confidential inform...{{dropped}}
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>