Duncan Murdoch wrote: > Andrew Piskorski wrote: > >>I'm currently using R CMD INSTALL to build and install some of my own >>custom R packages. Basically, I use a script which first builds a >>tarball of my R source code, and then calls R CMD INSTALL, which >>builds and installs that source package from the tarball, including >>re-compiling all my C code from scratch every single time, which is >>both totally unneccessary and tediously slow. >> >>What I would like to do instead (and what I do in fact do with S-Plus) >>is simply cd to the directoy where the production copy of my R package >>lives, do a CVS update, and then rebuild my R package right there in >>that directory. I do not want to generate any intermediary tarballs >>nor copy the package files to any other locations. >> >>Is there any current way to build R packages in place like that? Or >>if not, how would you suggest I go about accomplishing this? > > > You can use > > make pkg-foo > > to build and install a package whose source is in R_HOME/src/library/foo, or > > R CMD INSTALL foo > > to build and install one which is in the foo directory. No tarballs. > > Duncan Murdoch
I think the question was how to avoid to pollute the source tree. And the answer to this question is to write a makefile that provides a mechanism to clean up after installation - or at least a "make clean". For myself: For packages I am using the mechanism Andrew descibed at first (building at first); for building R, I am copying the source (svn) tree and "make"ing from the copy. Uwe Ligges ______________________________________________ [email protected] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
