Thanks very much Dan -- and Steve.
I perfectly understand that setting _R_CHECK_FORCE_SUGGESTS_=false globally
is not really an option, and that checks must be as strict as possible. I
am fine with using the Enhances field, since it keeps the dependency. Will
fix the package and re-submit.
An idea for a future bioconductor check feature, it would be nice if
authors could specify which of the Suggests packages are to be ignored,
this would give the desired flexibility, while giving control on those
check holes, as well as enabling possible programmatic tests on these
skipped dependencies.
Bests,
Renaud
On 17 May 2013 23:41, Dan Tenenbaum <dtene...@fhcrc.org> wrote:
> On Thu, May 16, 2013 at 10:53 PM, Renaud Gaujoux
> <ren...@mancala.cbio.uct.ac.za> wrote:
> > Hi,
> >
> > I am trying to submit a package to Bioconductor with a Suggests
> dependency
> > to RcppOctave, which is not available on Windows or Mac, and available on
> > Linux only if octave is installed (RcppOctave is mine as well but getting
> > working on other platforms than Linux is not yet on the agenda). All
> builds
> > go fine but checks are not performed due to this missing dependency. This
> > can be solved using _R_CHECK_FORCE_SUGGESTS_ = false, which is what is
> used
> > on CRAN, at least for Windows and Mac.
> >
> > Is there a recommended way around this?
> > I know I could "hide" the dependency using require('RcppOctave',
> > character.only = TRUE), but I am wondering if there is a
> > Bioconducotr-chekc-specific solution to still keep the dependency listed
> in
> > the DESCRIPTION file.
> >
>
> What we have done in situations like this is to (mis-)use the
> Enhances: field. R CMD check will not fail if a package in Enhances is
> not present.
>
> We do not want to set _R_CHECK_FORCE_SUGGESTS_=false globally, because
> on our build system, in 99% of cases, we want to make sure that all
> Suggested packages really are present.
>
> We could conceivably set _R_CHECK_FORCE_SUGGESTS_=false when specific
> packages are checked, but that functionality does not exist yet.
>
> BTW, it looks like RcppOctave is not building on Mac because Octave is
> not installed on the Mac build machine a CRAN (though there may be
> other problems):
>
>
> http://www.r-project.org/nosvn/R.check/r-release-macosx-x86_64/RcppOctave-00install.html
>
> If you do use the Enhances approach, you'll of course have to make
> sure that any code in your vignette or unit tests that use RcppOctave
> will need to be run only if require(RcppOctave).
>
> Thanks,
> Dan
>
>
>
> > Thank you.
> >
> > Bests,
> > Renaud
> >
> > [[alternative HTML version deleted]]
> >
> > _______________________________________________
> > Bioc-devel@r-project.org mailing list
> > https://stat.ethz.ch/mailman/listinfo/bioc-devel
>
>
[[alternative HTML version deleted]]
_______________________________________________
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel
