>>>>> "J" == Joseph McDonald <joe / vpop.net> writes:

J> Interesting. So are you saying that if an interrupt occurs, ensure
J> blocks are not *supposed* to be executed when rb_sys_fail() is called?

 No, I'm trying to say (with diplomacy :-)) that there is a problem in
 sysvipc.c 

 The call must be protected against control-c, and perhaps try to be
 thread-safe 


Guy Decoux