What would you say about using a system similar to Linux, i.e. a
permanent stable branch and a permanent development branch? Stable
versions would have even numbers and the experimental versions
would use odd numbers.

When we decide it is time for a new stable release, the development 
branch goes into a feature freeze and when everything has been 
debugged we release the development branch as a new stable branch 
and immediately start a new development branch.

We can continue to make minor changes to the stable branch and
release new versions of it during the development of a new stable
version, though.

We could start using this as soon as we release 1.2. The code in the
main trunk + plucker_1_2 branch would be the stable branch, while the
code in the hires and i18n branches would be merged and form the
unstable branch.

Then we could have two different snapshots on the web site, one for
stable and one for unstable. If you want safety, you use the stable
version, and if you want to test out new features, you get the 
development version.

What do you think?

/Mike

Reply via email to