best practices for private ports?

[Vasil Dimov]

On Thu, Mar 02, 2006 at 12:34:22PM -0600, Jeremy Messenger wrote:
> On Thu, 02 Mar 2006 09:24:48 -0600, Vivek Khera <vivek at khera.org> wrote:
> 
> >I have a few meta-ports I use to configure my base systems.  eg, I have  
> >a "kci-dbserver" port that just has runtime requirements for postgres,  
> >slony, mysql, and nullmailer.
> >
> >Right now I made a subdirectory in ports called local to put them in,  
> >but portsnap likes to delete it regularly :-(  These are very specific  
> >to my server configurations so don't really make sense for inclusion in  
> >the main ports tree.
> >
> >What is the preferred location/mechanism for having local ports?   
> >Ideally I'd like it to integrate with portupgrade which looks only in  
> >/usr/ports/* for the port files.
> 
> Use CVSup instead portsnap or copy marcusmerge.sh[1] with some tweak or  
> write your own script something similar to marcusmerge.sh that fit your  
> need. That's all I can think of.
> 
> [1] http://www.marcuscom.com:8080/cgi-bin/cvsweb.cgi/portstools/
> 
> Cheers,
> Mezz
> 
> 
Or create a script that calls portsnap and then restores your "local"
ports automatically, something like
cp -r /usr/ports-local/category/port /usr/ports/category/

Btw, have you tried playing with /usr/ports/.cvsignore?

Good luck!

-- 
Vasil Dimov
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 155 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-ports/attachments/20060306/ca0b2d49/attachment.bin