[ESS] ever see summary() hang Emacs/ESS? I did today, twice!

[Paul Johnson]

My computer lab has Centos 5.3 linux with R-2.8.1 and Emacs 22.2 with Ess 5.3.6.

I gave the students some regression exercises to work on.  They run lm
and then summary, nothing fancy.  One student came in and said "Emacs
crashed" and I went and looked and he was right.  The Gnome-Linux
session was still ok, but Emacs was just frozen, non responsive.  No
menus worked, it did not respond to Ctl-g.  I had to use xkill, which
is a depressing experience, except for the thrill of the little black
skull and bones and the chance to talk like a pirate and say "Aargh,
matie". Then I restarted Emacs, R and ran through his code step by
step and it was all fine.  I ran it over and over, no trouble.  So I
accused him of doing something evil to make the freeze happen and I
left.

About 2 hours later, another student came in.  She was using a
different PC in the lab, same thing happened.  Emacs was stuck after
running summary on a linear model. In the terminal from which she ran
Emacs, there was only the word "KILLED".  I xkilled that one, and
started running her code step by step, and I was able to reproduce the
freeze up twice. R would get half way through showing the summary of a
linear model and stop.  But it did not happen on the exact  same use
of the summary command in both tests, and after xkill, I restarted
Emacs and guess what?  Everything worked fine. I ran it through
several times.

I wondered if perhaps there could be a corrupted saved R session that
makes this blow up, but in the working directory there are no hidden
files at all.

If you have any ideas, I would be very glad to hear them.  I've not
had this "Windows System Administrator" feeling in the 10 years since
I started using Linux.  It is a bummer :)

-- 
Paul E. Johnson
Professor, Political Science
1541 Lilac Lane, Room 504
University of Kansas