Sean,
if type is set to mac.binary, then the contrib url is what reported
by Seth (and 2.3 only applies to R-devel, no binaries available at
the moment unless you point to a repository that builds packages
against R-devel. I think Seth is using R-devel for his test).
If you install from source, then it should not happen that you get
"macosx" in the url, as we only have one "path" for the sources (no
matter the architecture).
On R-2.2.0 I get
> contrib.url('foo')
[1] "foo/bin/macosx/powerpc/contrib/2.2"
which is correct. Cran binary of CRAN reports
> sessionInfo()
R version 2.2.0, 2005-10-06, powerpc-apple-darwin7.9.0
so it seems to me that Sean is using a R-devel version of R-2.1.0 and
not the final one.
The contrib.url function has been changed indeed just before official
final release of R-2.2.0
stefano
On Nov 30, 2005, at 6:47 PM, Sean Davis wrote:
On 11/30/05 12:44 PM, "Seth Falcon" <sfalcon at fhcrc.org> wrote:
In a recent post to the Bioconductor list, Sean reported trouble with the affy package's automagic package installation. The affy package is wrapping install.packages. The issue seems to be that on his system he sees:
sessionInfo()
R version 2.2.0, 2005-08-11, powerpc-apple-darwin7.9.0
contrib.url('foo')
[1] "foo/bin/macosx/2.2" Shouldn't an R 2.2.x see the following (that's what I get):
contrib.url("foo", type="mac.binary")
[1] "foo/bin/macosx/powerpc/contrib/2.3" Sean, can you try reinstalling R? Perhaps you somehow have an older version that doesn't have the right contrib.url function? I'm confused, but this seems to be the crux of your problem.
I'll try reinstalling at some point here. I tend to install from source, so that may be an issue. Sean
_______________________________________________ R-SIG-Mac mailing list R-SIG-Mac at stat.math.ethz.ch https://stat.ethz.ch/mailman/listinfo/r-sig-mac