I've searched a lot about this subject in the past. And we're currently using what Anders described. For completeness sake: you could possibly abuse classifiers for this as well.
Sander Verhagen [ [email protected] ] NOTICE: my e-mail address has changed. You may still e-mail me at [email protected] but you will see me using [email protected] from now on. Feel free to update your address book. -------- Original message -------- From: Anders Hammar <[email protected]> Date:18/04/2015 12:35 (GMT-08:00) To: Maven Users List <[email protected]> Cc: Subject: Re: what are some best practices for maven artifact naming when multiple code branches exist? I would add a suffix to the version for each branch. For example, if trunk is 1.0.0-SNAPSHOT, feature branch A would get version 1.0.0-featureA-SNAPSHOT. /Anders On Sat, Apr 18, 2015 at 9:10 PM, Nikita Tovstoles < [email protected]> wrote: > Suppose we have an app that's being developed on trunk and a couple of > branches simultaneously - and we'd like to publish artifacts from each dev > branch into our maven repo (so that we can run a cluster for each code > branch). Currently the 'V' in GAV equals our CI's build plan execution > number. > > What are some approaches for differentiating GAVs between each code branch? > I can think of: > > 1) using different artifact names (ie myapp for trunk, myapp-b1 for branch1 > version, etc) > > 2) reset version numbers so that trunk's starts at 100, branch 1 starts at > 10,000, etc (unlikely to collide in our case but seems cheesy) > > 3) Something else? > > thanks > > -nikita >
