Using CVS the right way

[m0llbuz_]

Hi!

I work in a company where speed is an essential ingredient. Releases occur 
frequently and irregularly (from one release per month up to two per week) and 
mostly only a part of the code base is affected in each release.

We code in perl and our code base structure looks like this:

lib -- apple
    -- banana

cgi-bin -- apple
        -- banana

site1 -- apple  -- templates
      -- banana -- templates

site2 -- apple  -- templates
      -- banana -- templates

Directory "lib" contains perl modules, "cgi-bin" contains CGI scripts, and the 
site directories contain templates and configuration files. The perl modules 
and the CGI scripts are shared by all sites. We are using mod_perl so can only 
use _one_ instance of each module for all sites.

We are currently doing all development on the main trunk (except medium to 
long-time projects that affect other projects - these we put in branches) and 
we also put bug fixes in the trunk. It works quite well, but sometimes it's 
hard to know what files are to be released to the testing environment (patch or 
something else?).

What is the best way to use CVS with this scenario? We want to be fast and not 
get slowed down with a lot of merging, branch creation, etc. We have talked 
about creating a patch branch, but it will get tedious to create a new patch 
branch efter each release. Another problem is that several projects can be 
going on at the same time with different start dates and release dates.

Thanks in advance! :-)
Milo

-- 

Powered by Outblaze