Hi Larry,
thatnks for reinforcing my observations.
It seems that the cvs system could do with a review from the protocol
level. It seems to have evolved from an rcs local file tool to a client
server system. Thsi seems to have lead to the server doing things that the
clinet shoudl be doing and visa versa.
For instance I was surprised to see that the server changes the format of
the data sent if the co request is sent with a pipe argument. It seems that
this should be a client process only and not have effect on the format of
the data being sent by the server. This would be the same as an apache web
server send a web page as postscript because the client is going to print
it.
I have noticed a review of the server process in train resulting in the
nserver version with seperated out authentication. This seems like a step
in the right direction but perhaps version 2 of cvs should break with the
problems of version 1.
Cheers,
Gerrit Thomson.
Larry Jones wrote:
> Gerrit Thomson writes:
> >
> > >Synopsis: atempting to use global option -n to stop transmission
> > of data from server results in "E cvs [server aborted]:there is no
> > version here; run 'cvs checkout' first"
>
> This doesn't really have anything to do with client/server, the same
> thing happens in local mode. As I recall, -n has never been useful with
> checkout and there's no easy way to fix it.
>
> -Larry Jones
>
> I take it there's no qualifying exam to be a Dad. -- Calvin
