|
From: | Mike Solomon |
Subject: | Re: survey on multiple development versions |
Date: | Tue, 10 Dec 2013 15:52:24 +0200 |
On Dec 10, 2013, at 3:41 PM, Carl Peterson <address@hidden> wrote:
The idea is that the canonical version will pick up all inclusion-worthy features a few weeks / months after the experimental versions. It is true that it’s annoying to wait, but the current state of things is for these features either to not exist at all or exist at a much later date. I think that the scenario you describe above, while frustrating, is preferential to the current state of things.
The versions are always “branched" off of canonical development, like so: canonical development |— A |— B |— C Think of canonical development as a foundation for A, B and C. This means that any bugs fixed in the canonical version will be fixed in the experimental one unless the experiments are so experimental that they break everything, which we’d obviously try to avoid. Conversely, a problem in branch A will never be fixed in just branch B - it will first be fixed in A and then applied to all of the branches (or it will persist in all the branches if no one catches it). Let’s assume that canonical devel is at 2.19.23. You are using 2.19.3.A. It would go as follows: A: "I have this problem. I am using version 2.19.3.A” an appropriate response would be either: “This was fixed with version 2.19.23.” or “This was fixed with version 2.19.23.A” but never “This was fixed with version 2.19.23.B" It is true, though, that attentiveness to versioning would become important for testers and responders alike. Cheers, MS |
[Prev in Thread] | Current Thread | [Next in Thread] |