[plt-scheme] Broken pipe?

I recently had a similar bug.  Both the #<eof> and broken pipe  
suggest the server side has been closed, which would happen if the  
procedure passed to `run-server' returned or raised an exception.    
Have you tried something like this, to see if that is happening:

(run-server
    ...
    (lambda (in out)
      (dynamic-wind
         (lambda () #f)
         ...
         (lambda () (printf "Unexpected return!~n"))


Rob


On Jul 19, 2007, at 1:16 PM, Erich Rast wrote:

> Hi,
>
> I've further tracked down the problem with my tcp code. run-server  
> seems to work and tcp-connect seems to work, too. My protocol  
> works, too, at least with pipes instead of tcp ports. However, for  
> some odd reason my client continually receives #<eof> from the  
> server when I use tcp. Now if I ignore this and write something to  
> the client's output port, I get the following cryptic error message:
>
> tcp-write: error writing (Broken pipe; errno=32)
>
> Does anybody have an idea what this means and by what it may be  
> caused? How did I manage to break the pipe?
>
> OS X 10.4.10, PLT Scheme version 370.6-svn17jul2007 [3m]
>
> Regards,
>
> Erich
> _________________________________________________
>  For list-related administrative tasks:
>  http://list.cs.brown.edu/mailman/listinfo/plt-scheme