Jan> Marco Antoniotti <marcoxa(a)copernico.parades.rm.cnr.it&gt; writes: Jan> Well. I am not sure whose problem this is exactly, but for me the Jan> problem seems to be with something (i.e. either XEmacs or Clisp[1]) not Jan> flushing pipe buffers correctly. The patch below "fixes" the problem Jan> for me. I can run Clisp now and evaluate "(+ 1 2)", define a function Jan> and have it complete the function name in a call. One other thing: I think you're supposed to use the -I option when running clisp with ilisp. I'm vaguely remember the -I option being added exactly for this purpose. I causes clisp to flush it's buffers at the right time, or something like that. Clisp works fine for me, without Jan's patch. (Well, except it can't autoload clisp-hs from ilisp-chs.el, but that's a different issue.) Ray