Skip to content

Q: Segmentation fault at all commands ?

6 messages · mhv@dmu.dk, Brian Ripley, Peter Dalgaard +1 more

Original
mhv@dmu.dk
# 
Dear List

Sorry if this is trivial, but I can't find it in the manual, FAQ or mail
list (2002).

I just installed R 1.6.0 (2002-10-01) on Linux R.H. 7.2 (2.4.9-32.5) on an
DIGITAL AlphaServer 800 5/500

The program starts up nicely but exits immediately with a "Segmentation
fault" at the first command. It seems that it doesn't matter what command, I
tried q(), help(), version, licence(), x<-1, etc. 

All comments are welcomed.

:-? Martin Hvidberg
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Brian Ripley
# 
What compiler?

RedHat's used of `gcc 2.96' is notorious around here.  Could you try
building under an actual released version of gcc if you used `2.96'?

BTW, R 1.6.1 is the current version, so why not use it?
On Mon, 25 Nov 2002 mhv at dmu.dk wrote:

            

      
      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Peter Dalgaard
    

    

      
      #
    

  


  

      
<ripley at stats.ox.ac.uk> writes:

            

      
      
      
    

        
I believe we fixed a bug with the readline stack for 1.6.1, which
only affected 64-bit systems. That could be the one sticking its head
up here.

(Well, only and only --- we found that fixing that bug uncovered
another one on all systems where the stack isn't dropped on Ctrl-C.)

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Brian Ripley
    

    

      
      #
    

  


  

      
On 25 Nov 2002, Peter Dalgaard BSA wrote:

        

            

      
      
      
    

        
I run R on two different 64-bit systems (using readline), and that bug did
not occur at startup on those.  Starting R with --no-readline would
check if that is the cause.

R is regularly built and tested on several 64-bit systems, which is why I
suspected the compiler no being 64-bit.

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Peter Dalgaard
    

    

      
      #
    

  


  

      
ripley at stats.ox.ac.uk writes:

            

      
      
      
    

        
It was PR#2165 I was thinking of. That was on Tru64, but isn't that the
same Alpha hardware? Symptoms are certainly similar.

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Michael Mader
    

    

      
      #
    

  


  

      
Hi

        
Peter Dalgaard BSA wrote:

            

      
      
      
    

        
Just for Curiosity:
R-1.3.0 up to R-1.5.1 compiled well on both Tru64 OSF4.x and OSF5.1 with
the respective CC/CXX/FXX and Compaq's libs.

Starting with R-1.6.0, OSF4.x has troubles compiling/linking R with
Compaq's compilers/libs (and readline). On OSF5.x R compiles still fine
out of the box. 

Since most of our boxes are now (far too late) on OSF5.x, I did not find
it worth anything to go into details and search for incompabilities.

I have no idea on which Tru64 OSF  version the mentioned AlphaServer is
running but perhaps this workaround might be worth a shot.

Cheers

Michael

P.S.: My gut feeling is that this is merely an incompability with
Compaq/HP's libs (which aren't best conforming standards).