Hi,

we are using the following scenario, all our source is on a single repository but we split our code in main trunk into folders per technology (trunk/db, trunk/web, trunk/c, trunk/java):

• database is built on stage 1
• web, c++, java are built on stage 2 in parallel on a pool of compilation machines
• tests run on stage 3
• deployments on stages 4,5,6.

The problem is if we want to recreate the artifacts by running a build for a specific revision only the default repository definition uses the custom revision the rest are updated to head.

Due to the size of the code we cannot checkout full trunk on each build server and use default repository everywhere, we need a way to filter out database code and ansi c code from the web servers, or web and c from database.

Regards,
Iulius