Path: utzoo!utgpu!utstat!jarvis.csri.toronto.edu!mailrus!cornell!rochester!udel!mmdf
From: jnall%FSU.BITNET@cunyvm.cuny.edu (John Nall 904-644-5241)
Newsgroups: comp.os.minix
Subject: Re: Re: /usr/minix/amoeba/examples/client3.c
Message-ID: <11449@louie.udel.EDU>
Date: 23 Mar 89 13:49:25 GMT
Sender: mmdf@udel.EDU
Lines: 17

Responding to an earlier message of mine which pointed out that the
client3.c program has an error, Thomas V. Frauenhofer states that
  >  The code is correct.....

Well.....actually, it is not correct.  You see, the "n" that the read
returns is assumed by the client3 program to be the number of bytes
which were remotely read.  The program then proceeds on that basis.
It is quite easy to verify: Andy's networking works very well with both
the client and server on the same machine.  Just do a make of server3
and client3, start server3 running, and type "client3 client3.c".  It
will not list anything out (because the count of returned bytes is not
correct). Then go in and change the offending line to
have the correct nesting, and it does work.  The proof of the pudding is
in the eating!

I agree it is a subtle error, though.  That is why I thought it was so
neat that my beginning student caught it.