[R] R hangs. Uninstall/reinstall seems insufficient.

[Eliot Miller]

Hello,

I'm running a Mac PowerBook G4 PowerPC with OSX 10.5.8

I was using R 2.12.xx with no problems.

I tried to install R 2.13. Have never had any trouble with simply
downloading the pkg and installing over previous versions.

I did that, selected the packages I use regularly, set them to
download with "Install Dependencies" checked and went off to sleep.

When I came back in the morning, I had a message saying something to
the effect of a large number of packages (dependencies) were not able
to be installed. Other than that, all seemed fine.

I can run codes and generally use R normally.

However, if I misspell anything in the console, e.g. type
"lirbary(xx)"...anything that would tell me such and such doesn't
exist, it hangs indefinitely. I have to force quit.

I tried uninstalling it and installing it again. I tried uninstalling
it and installing a more recent build from http://r.research.att.com/

I tried uninstalling it and re-installing R 2.12 from the same site.

I followed many of the instructions here,

** OS X 10.5 and higher you have to use
  ** For 32/64 Bit Binary pkg
     pkgutil --forget org.r-project.R.Leopard.fw.pkg


  ** For the 32 bit Tiger binary dmg:
     pkgutil --forget org.r-project.R.framework
     pkgutil --forget org.r-project.R

** Etc:
  ** Check for pkgs:
     pkgutil --pkgs | grep r-project


** UNinstall R
  rm -rf /Library/Frameworks/R.framework
  rm -rf /Applications/R.app

until I couldn't find anymore R files on the computer, then
re-installed, trying both 2.12 and 2.13 after first doing the above
steps, and still no luck.

I have also used pkgutil --unlink for the two files --
org.r-project.R.Leopard.fw.pkg, org.r-project.R.Leopard.GUI.pkg  --
then restarted, installed again, all to no avail. In other words, I
repeatedly deleted both the GUI and the Framework and--I believe--all
associated files (e.g. Library/Receipts/bom), installed, and it did
not solve the problem. No other programs on the computer seem to be
running improperly.

I have also created a new user on the computer and tried to install R
with that user. Installation worked, but left with the same problem.
There were some errors in the installation log from the installation
on this new user:

Jul  3 20:58:12 Macintosh-102 Installer[474]: JS: Package Authoring
Error: Exception thrown while running volume check. TypeError: Result
of expression 'my.target.systemVersion' [undefined] is not an object.

and also a series of:

Jul  3 20:59:10 Macintosh-102 Software Update[440]: Package Authoring
Error: installation-check results requires a message

Throughout all of this (except when I created a new user), whenever I
re-installed R and started it up, it would start with the same
background color I had chosen years ago, and was saying it was
re-creating the history from .Rhistory (this latter part only until I
went ahead and set in preferences to stop loading the history
automatically). Obviously, there were still some files/settings
somewhere.

Finally, I downloaded an old version of appZapper capable of running
on OSX 10.5 and deleted Rhistory and did the full uninstall with
pkgutil...I was able to make R not load with any previous colors, etc,
but the problem still exists.

I also updated Java today to the newest version of Java 5, but I went
into preferences and switched back to the old version, which had no
effect on the problem with R.

Apologies if you think this isn't an R question, and it's actually a
problem of my computer, but I can't figure how that would be. Any
suggestions?

Any help would be greatly appreciated. Thanks!

Eliot