Excessive network round trips on new files?
I have a 2 machine network here at home, with the CVS repository on a
linux box. The main client is my Windows 2000 machine, and I connect with
ssh to the repository. The machines are both running 100mbit full duplex
and get really fast transfers as there is no other traffic or users :)
Ok, on my Windows box, I added several .lib and .dll files I had built
as part of a redist (btw I did remember the -kb flag :). However, if i run
any kind of command (status, update, etc) before commiting these large
files, the client and/or server send just huge amounts of traffic to each
other every time. Seems like its sending a full diff to the server on what
is a nonexistant file over there? Seems like a design flaw.
