[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: problems with the concept of unstable -> testing




On 15/12/2008, at 21.25, Cyril Brulebois wrote:

Or you might use experimental, and keep unstable for lenny.

Another model that I think has not been discussed is never freezing stable.

Say "stable" is redefined as "bug-free" in the sense that there are no RC bugs in that repo. If a serious bug is found in a package, it is removed from stable, until the bug has been fixed in "testing".

When it's time for a release, that is simply forked off "stable", and the release branch is frozen (with the usual development work going on towards the final release.)

After the branching, "stable" will continue to receive updated/fixed packages from "testing", which again will continue to receive packages from "unstable".

This would create a situation where there are three development branches, graduated according to degree of "bug-freeness" and one release branch. The advantage is that work does not need to halt in any of the development branches, while keeping a "stable" release that at any point is ready for branching into a release.

Cheers,
Morten


Reply to: