[R-SIG-Mac] Installation failed to install GUI

Bob Harris bob at bumblebeagle.org
Sat Oct 28 04:05:16 CEST 2006


Howdy,

Short description:

I tried to install R 2.2.1, the Power PC version.  I can now run 2.2.1 
from the terminal app but the GUI did not install.

--- Longer description ---

I have been using R 2.0.1, via the GUI, on my PowerPC OS10.3.9 Mac, but 
today attempted to upgrade this to 2.2.1.  I logged in as an 
administrator, downloaded from the link powerpc/R-2.2.1.dmg at 
cran.r-project.org/bin/macosx/old.  I validated the md5 code, opened 
the .dmg file, the doubled clicked the top-level .mpkg file.

During install, it claimed it was going to install R.app in my 
applications folder.  However, no such file or folder is to be found.  
I realize that it will appear in the finder as just plain R.  I have 
manually hunted for it, and have also used the finder's find function 
to search for it.  I note that the read-me-type info shown during 
installation describes a couple of the *other* packages as "optional" 
but provides no means to choose or not choose them.

The command-line r runs from the terminal window, which I presume means 
that the framework version installed OK.

The .dmg file has a subfolder called packages so I looked in there and 
saw four .pkg files, one of which was R-GUI.pkg.  Grasping at straws, I 
double-clicked on this and it brought up the installer.  Unfortunately, 
when it came to the panel which allows me to elect the destination 
folder, it marked my hard drive with a red exclamation point at grayed 
out the buttons that would allow me to choose a destination folder.

Please help.  I use R on a regular basis for my research (I'm a PhD 
student).  I regret having attempted to upgrade, because I have now 
lost the ability to use R at all.

--- Even longer description ---

The reason for attempting to upgrade was a difference I was seeing in 
the results from ks.test in R version 2.0.1 on my machine, vs R version 
2.2.1 accessed from RPy, on a different machine (the latter 
occasionally reports negative probabilities which is not a good sign).  
In attempting to chase down that problem, I thought it would be a good 
idea to try 2.2.1 without RPy so I could isolate whether the problem 
was in RPy or R.

The reason I chose 2.2.1 instead of a newer version is twofold.  First, 
I wanted to compare results to the version 2.2.1 running on another 
machine.  Second, I am running OSX 10.3.9, and am not interested in 
attempting to upgrade to 10.4.X at the moment.  The download page 
advises me that I can build the latest version from the source code, 
but I've been down that path before on three Macs and have never been 
anywhere close to successful.

Thanks for any help,
Bob H



More information about the R-SIG-Mac mailing list