Marc G. Fournier wrote:
> On Fri, 24 Mar 2006, Joel Miller wrote:
>
>> Joshua D. Drake wrote:
>>
>>> Is there any reason why we don't allow rsync access to the cvs repo?
>>>
>> The only reason I can possibly think of is server load, comparing the
>> timestamps for every file and directory in a repository for every
>> rsync session could be taxing if everyone under the sun and all their
>> cousins could do it.
>>
>> CVSync, on the other hand, can cache those timestamps into a file and
>> send it to the client, offloading the comparison workload from the
>> server to the client where it belongs.
>
>
> I'm going to look at installing this also ... what does the client
> side need?
>
> ----
> Marc G. Fournier Hub.Org Networking Services
> (http://www.hub.org)
> Email: scrappy@hub.org Yahoo!: yscrappy ICQ:
> 7615664
>
To build and install? Just the sources from www.cvsync.org and gcc. I'm
running OS X, and I needed to specify a useful value for HASH_TYPE in
mk/defaults.mk (doing 'make configure' creates that file).
Basically here's all you need to do: download sources, expand and cd into the folder make configure (tweak
HASH_TYPEin ./mk/defaults.mk if needed) make make install
I created an empty config file in /usr/local/etc/cvsync.conf to keep the
client app from complaining. You can create individual config files for
each collection if you want and store them wherever it makes sense for
you (I keep mine one folder up from the destination). The file format is
documented in the man pages for cvsync and cvsyncd. I have a commented
example config file for defining collections on the server side that I
can send if you want it. You can also specify most parameters on the
command line as well, both for the client and server.
If you want to take a look at a live deployment of cvsync, check out
http://openbsd.org/cvsync.html
Hope this helps!
Joel
--
// file: mice.h
#include "frickin_lasers.h"