[R] dev() error

(Ted Harding) Ted.Harding at nessie.mcc.ac.uk
Wed Jun 16 12:58:12 CEST 2004


On 16-Jun-04 Laura Quinn wrote:
> No nothing like that, I am running two parallel R sessions (something I
> often do) and the other one is working absolutely fine - I am working
> directly from my network, not via ssh, and certainly am not privvy to
> any super user passwords!

It looks as though you may be logged in to two different machines
for your two R sessions, in which case it looks as though one of
them has ascertained the X display for your machine (i.e. the one
you are sitting at), and the other has not. (Indeed, the one that
works might be your machine itself, in which case it knows anyway).

Test:

1. In the xterm logged in to the machine which works, before starting R,
   enter

     echo $DISPLAY

2. On the xterm logged in to the machine which does not work, before
   starting R, enter

     echo $DISPLAY

For example, I'm using machine "compo" here on which is an xterm
logged in to another machine called "brandy". So, in this xterm,

  ted at brandy:~ > echo $DISPLAY
  compo.fort.knox.uk:0.0

showing that "brandy" knows the X display destination for "compo".

So, when you do (1) above, you should see something similar. If I'm
right, when you do (2) above, you will probably just get a blank
response.

If that happens, then you have to see about getting the (2) machine
to set the DISPLAY variable when you log in to it. This can be tricky:
I've known cases where the cause was quite deeply hidden. Basically,
the information for the DISPLAY variable is ascertained by a script
which is run when you log in to the remote machine, which may be
a sub-script of some other script, ...

I hope this helps,
Ted.


--------------------------------------------------------------------
E-Mail: (Ted Harding) <Ted.Harding at nessie.mcc.ac.uk>
Fax-to-email: +44 (0)870 167 1972
Date: 16-Jun-04                                       Time: 11:58:12
------------------------------ XFMail ------------------------------




More information about the R-help mailing list