I thought a good approach would be something like the following:

Have a normal website that's pretty stable.

Have a dev site (cvs.euglug.org?) that contains the CVS web document
root that the web team does development/changes on.  Once we hit a
solid 'release', or have debugged a new feature/change, we transfer
that tree to the stable site (hopefully via a script).

Once we get the new box set up, we can get the current site on CVS,
alongside the current site in stable, and go from there.

(This would be a CVS learning experience for me.)

-Rob

> On 20011130.1227, Bob Miller said ...
>
> Rob, if you could put the whole web site into a single directory tree,
> including the Apache config files, then you could check that whole
> tree into CVS.
> 
> Then other people could create a clone of the site by doing something
> like,
> 
>       export CVSROOT=:pserver:[EMAIL PROTECTED]:/cvs
>       cvs get www
>       sudo httpd -d www
> 
> and test new content on the clone by doing
> 
>       lynx http://localhost/
>       (or netscape... (-: )
> 
> That way, everybody gets their own staging server.  We could have CGI,
> PHP, Python, etc. content on the site, and test it all before it goes
> live.
> 
> Bringing new content live on the web site would be a simple matter of
> doing "cvs update" in the real web server's tree.  You could do that
> manually or make it happen automatically after a checkin.*
> 
> The only bad bad part is that each of us would have to have an Apache
> server with the right modules installed.
> 
> Is this the direction you want to go?
> 
> * TiVo's engineering web server does something similar.  They have
>   a cron job that runs every five minutes updating the site from
>   the repository.
> 
> -- 
> Bob Miller                              K<bob>
> kbobsoft software consulting
> http://kbobsoft.com                     [EMAIL PROTECTED]
> 
> 

--
Rob <rob_at_euglug_dot_net>
my @euglugCode = qw(v+++ e--- eug+ bsd+++ gnu+ S+++);

Reply via email to