[R] save.image() error reappears

[Marc R. Feldesman]

A week or so ago, I reported a problem on my new Windows XP box getting R 
1.4.1 and R 1.5.0 (development from BDR's site) to save the .RData file 
upon program exit.  Specifically, either clicking on the "X" in the GUI or 
invoking q() from the command line produced the following error:

Error in save.image(name) : image could not be renamed and is left in 
.RDataTmp1

After some fiddling, I was able to "cure" this problem by repartitioning 
the hard drive to add a small amount of space to my data drive.

This worked until yesterday when the problem reappeared.  As far as I can 
see, there are no permission problems, the drive is not in any way 
write-protected, and every diagnostic program I can throw at the machine 
finds absolutely no problem with either the physical surface of the drive 
or the FAT.  The entire (large) hard drive is partitioned and formatted as 
NTFS5.

I have no difficulty writing any other files to the "g" partition, and 
every other program that makes use of temporary copies of large working 
files on "g" has no difficulty.

If I move the R data directories to the same partition as the binary file 
("f" in this case), I don't have this problem.  If I save the .RData via an 
explicit call to save.image("file name"), I don't have any problem, and I 
can save the .RData file by using the pull-down menu in the GUI (File|Save 
Workspace) without difficulty.

I can, of course, modify the defaults to save.image (...safe=FALSE) and get 
around this problem, but I am plain out of ideas of why this should be 
happening.

Brian Ripley has suggested "permission" problems, but there are none that I 
can see.  Brian and Peter have both mentioned possible RACE problems (but 
why only on "g" and not anywhere else?).  Brian also confirms that having 
the binary on one logical partition and the data on another should not be 
the problem as his wife's system is set up this way.

I'm left with a couple of questions:  1)  does R have any way of knowing 
how large the hard drive is (unpartitioned).  The "g" partition lies out in 
the region beyond 80 GB, which is the upper limit for some programs and 
BIOS revisions - obviously my BIOS supports this.  2)  does anyone have any 
experience with running R under Windows XP on a 2.2 GHz P4?  (This should 
hardly be an issue, but I'm grasping at straws right now.  My machine has 
1024 MB RAM and a 120 GB Hard drive).

Any suggestions welcomed.  I'm tearing out what little hair I have left.








Dr. Marc R. Feldesman
Professor and Chairman
Anthropology Department - Portland State University
email:  feldesmanm at pdx.edu
email:  feldesman at attglobal.net
fax:    503-725-3905

"I've proved who I am so many times,
the magnetic strip's worn thin"   Bruce Cockburn

Powered by Tyrannochoerus - the 2.2 GHz WinXPP Box

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._