Forum OpenACS Development: Re: Suggestions for release management after 5.1

Collapse
Posted by Don Baccus on
The branch-and-move-forward paradigm is followed by the Postgres, Mozilla and many other groups.  It is one that many people in the open source community are familiar with, and feel comfortable with.  It has worked well for us in the past and should for the future.  I don't see that CVS issues have much - indeed anything - to do with the long delay involved in releasing 5.0.

I'd favor running automated smoke tests against HEAD regularly, nightly if possible, with those who break things tasked with fixing them, with the understanding that the development process may introduce a time lag.  A big problem we have at the moment with development is not having a good understanding of the current state of the release under current development.  Changing the way we use CVS won't change that.

Likewise we don't need to change our use of CVS to switch to a calendar-driven rather than feature-driven release scheduling paradigm.

Nor do we need to do so to switch to an asynchronous core vs. packages paradigm (something we all agree on doing).