Ideas for handling hardcoded URLs on a CVS managed website?

[Georgi Kostov]

We just started developing a website and we want to use CVS to make our 
life easier.

There are several developers here and we want to give each one a 
"sandbox" (devel1.site.com, devel2.site.com etc) so we can take full 
advantage of CVS. There is also a site called prod.site.com to which 
we'll checkout a stable version of our work. The problem is that some of 
the static html files contain hardcoded urls (images, links, etc) 
peculiar to each of the development sandboxes, and always different from 
the correct urls for the production site.  Commiting them into the 
repository and later checking them out will have the undesirable effect 
of producing wrong links on the prod. site.

Can you give me some ideas how to manage with this problem? I indended 
to write a script that will replace each hardcoded link with a keyword 
(using a dictionary) before commiting the file to the repository but I 
haven't thought of how to deal with updates/checkouts.

P.S. Just to be on the safe side: Most of our content is generated by 
cgi scripts and it doesn't have this hardcoded links problem, but there 
are static pages and configuration files that do.

Georgi Kostov
SunFish Ltd.