[Rd] fortune() in .Rprofile conflicts with R CMD INSTALL

Kurt Hornik Kurt.Hornik at wu-wien.ac.at
Mon Mar 5 22:29:06 CET 2007

>>>>> Kurt Hornik writes:

>>>>> Gregor Gorjanc writes:
>> Seth Falcon <sfalcon <at> fhcrc.org> writes:
>>> The quick fix for you is, I believe,
>>> if (interactive()) {
>>> library(fortunes)
>>> fortune()
>>> detach("package:fortunes")
>>> }
>>> Whether something could be changed so that R's output doesn't confuse
>>> itself and gcc when doing R CMD INSTALL, seems like it should be
>>> possible.

>> Yes, this solves my! problem. I do not know what is causing this
>> behaviour.

> Well, r-devel's src/scripts/INSTALL.in now has

> if test -z "${lib}"; then
>   lib=`echo "cat(.libPaths()[1])" | \
>     R_DEFAULT_PACKAGES=NULL "${R_EXE}" --no-save --slave`
>   message "Installing to library '$lib'"

> so we need to find a way to "just get" the result of cat(.libPaths()[1])
> into $lib as intended.

> One idea might be ensuring that this gets into the last line on its own,
> and then taking $lib as the last line of what we got ...

I think I found (and committed) a solution for this.


More information about the R-devel mailing list