R help - Dec 2009 - Difficulty with terminal properly displaying help function in an ESS remote session

Hi all,

I'm logging into a Debian server and running R remotely using ESS. The
steps I use to do this are below (pasted from my webpage). However,
we're having a problem whenever we want to use the help function,
e.g.,
?hist

The remote buffer gives a warning:
 "WARNING: terminal is not fully functional
-  (press RETURN)"

At this point we can't get back to our normal R session. When I use
ESS locally on my MacPro, help screens open up in a split buffer below
the one I'm working on, which is fine.

So there are two issues:
1) How do we switch from the help screen back to the R session?
2) Is it possible to have help screens open up in separate buffers or
in split buffers when using ESS remotely?

Any help is appreciated!

Matt


Steps we take to use ESS remotely:
1) Open up your *.R script you?d like to use

2) Open a shell inside Emacs by typing ?M-x shell?

3) From within this shell, ssh to the server you want to use. When
doing this, you need to make sure to specify two important ssh
options: compression (which compress data coming to you, making the
connection seem *much* faster) and X11 forwarding (which allows you to
use interactive graphing features via X11). E.g.:
ssh -XC username at servername.colorado.edu

4) You should now be logged into the server, just as you wold be if
you?d used terminal rather than emacs. Now open up R as you usually
would on that server. E.g.:
R --arch=x86_64

5) You should be in R now. To allow this R session to be linked to
your *.R script, use this command in the remote R session:
M-x ess-remote
In the Emacs mini-buffer prompt, type:
r

6) Now you should be able to send code from your *.R script to the
remote R session as you normally would (e.g., C-c C-j).

7) Last, you might need to change the options in your remote R session
to graph using X11 rather than whatever default driver is being used.
To do this in R, type:
options(device=?x11?)

8) That?s it. Make sure it all works by typing something like:
hist(rnorm(50)) #which should return a histogram of rnorm to your screen!


-- 
Matthew C Keller
Asst. Professor of Psychology
University of Colorado at Boulder
www.matthewckeller.com