[Rd] libR.dylib on OS X

[Stefano Iacus]

Jan,
If you inted to use RAqua as a CLI app you don't need to do "make  
install" but only "make install-aqua" and create a couple of symlinks.  
Otherwise you have conflicts. If you install the RAqua from CRAN and  
apply the symlinks as explained int eh ReadME file you'll have an  
equivalent version R as of installing from sources with make install.

I have discussed a bit with Simon on the relocatable possibility for  
RAqua.
I think we will have something ready for 1.9 and /Library/Frameworks is  
probably a good place.
Having not a monolithic app means that you should also provide a  
un-installer, otherwise dragging RAqua icon into the trashcan only  
removes a couple of bytes but the rest of 50mb still resides in your  
system (the same for user installed packages).

The other problem we are faced is the startup script R, if you move it  
you need to take care of $R_HOME and other similar things. Simon seems  
to have a solution for this.

We will probably be able to use one bumping app only (i.e. only RAqua  
and not the couple of StartR and RAqua).

To summarize: it's in the todo list of R 1.9.0 (as well as the  
eventloop problem which should be a cross platform solution tough).

stefano


On Sabato, ott 18, 2003, at 19:27 Europe/Rome, Jan de Leeuw wrote:

> If one installs the CLI and RAqua versions, by saying both "make  
> install"
> and "make install-aqua", then packages that use libR.dylib get their
> symbols from  
> /Applications/StartR.app/RAqua.app/Contents/Frameworks/libR.dylib
> and not from /usr/local/lib/R/bin/libR.dylib.
>
> That's unfortunate, because the first is likely to change, and the  
> second is not.
> For instance, what happens if Stefano succeeds in constructing an  
> StartR.app
> which can sit anywhere on the system (as a Mac application should) ?
>
> Again, these types of problems can be circumvented by putting all of R,
> except the CLI and Aqua front ends, in /Library/Frameworks. As in  
> MacPython
> or in TclTklAqua. If you distribute an installer package, you don't  
> need
> a monolithic application, and you avoid useless duplications and  
> problems
> with hard-coded paths.
> ===
> Jan de Leeuw; Professor and Chair, UCLA Department of Statistics;
> Editor: Journal of Multivariate Analysis, Journal of Statistical  
> Software
> US mail: 8130 Math Sciences Bldg, Box 951554, Los Angeles, CA  
> 90095-1554
> phone (310)-825-9550;  fax (310)-206-5658;  email:  
> deleeuw at stat.ucla.edu
> homepage: http://gifi.stat.ucla.edu
>   
> ----------------------------------------------------------------------- 
> --------------------------
>           No matter where you go, there you are. --- Buckaroo Banzai
>                    http://gifi.stat.ucla.edu/sounds/nomatter.au
>
> ______________________________________________
> R-devel at stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-devel
>