Le 16/04/2016 18:55, Scott Kostyshak a écrit :
On Sat, Apr 16, 2016 at 07:45:29PM +0200, Jean-Marc Lasgouttes wrote:
Le 16/04/16 19:31, Scott Kostyshak a écrit :
On Sat, Apr 16, 2016 at 08:34:47AM -0400, Richard Heck wrote:
As usual, fixes for bugs committed to 2.2.1-staging (=2.2.x) should first be
committed to 2.3-staging (=master). These should be marked "fixedinmaster"
in trac but NOT "fixedinstable" and tagged with milestone 2.2.1.
So now "fixedinmaster" does not necessarily mean "fixed in the next
major release" (2.2.0). This is confusing I think. But on the other hand
creating another category (e.g. fixedinstaging) would also be confusing.
I am not sure what the point of 2.2.1-staging is. What are the urgent fixes
for 2.2.1 that one would not want to have in 2.2.0? A 2.2.2-staging branch
would make more sense to me.
This is a good question I've been trying to understand. Georg used the
phrase "really critical" and Richard "absolutely critical" and although
I at first disagreed slightly (wanting to remove the adverbs), the more
I think about it the more I agree.
From what I understand, the idea is that we will hopefully release 2.2.0
soon, and so we have less time to catch something if it is wrong before
release. We will release 2.2.1 later than 2.2.0, so we have more time to
catch a problematic commit, and thus we can allow (slightly) riskier
commits to 2.2.1-staging.
I have not fully convinced myself of the above, but it is making more
and more sense to me. I'd be curious about the opinions of others.
So not critical enough to be in 2.2.0, but critical enough that it
cannot wait 2.2.2? I would need more than these adverbs to make sense of
this.
In addition there is nowhere to backport commits from 2.3.0 intended for
2.2.2. I agree with Jean-Marc here, I think there is more use for a
2.2.2-staging branch, possibly later merged with 2.2.1 if it is decided
that no emergency 2.2.1 release is necessary.
