Tobias Weingartner writes:
>
> On Monday, July 24, Larry Jones wrote:
> >
> > The client/server protocol was intended to address this need.
>
> In a rather poor way. If someone would document the latest incarnation
> of this "protocol", and also specify the semantics it currently employs,
> I believe a great service would have been done to the CVS community.
>
> (Yes, I know about cvsclient.texi, it is not particularly usefull, for doing
> an implementation, you still need to run tests against a 'cvs server' binary,
> using stdin/stdout to test/understand things...)
cvsclient.texi does document the latest incarnation of the protocol, but
I agree that it isn't sufficient for doing an implementation unless you
already have a pretty good understanding of how client/server works.
It's kind of like trying to learn C by reading the ISO standard instead
of a friendly tutorial.
-Larry Jones
I don't like these stories with morals. -- Calvin
