I would _love_ to do an alpha release before we get into further
refactoring efforts. I use 2.1 as my default now, and while I don't
venture too far afield into EAR, EJB, or other project types, I believe
it's at least stable enough for an alpha-1. Obviously, there is an
endless wishlist we could put into 2.1, but even if we do intend to hold
2.1-final off for the end of that list, we should let the existing,
large changes soak into the community for awhile to see what's broken
that we missed.
Also, keep in mind that NetBeans, Eclipse, and IDEA are all using
snapshots of 2.1 already. I know there are plenty of issues in those
projects - some relating to their respective codebases, and many
relating to Maven itself - but the fact that we have multiple releases
each making more progress toward stability, and built on 2.1, means it's
time to get an alpha out.
Now, if we were to move in this direction, do we need to think about a
codefreeze branch, to help chase out the lingering problems, so as to
avoid slowing down parallel progress on trunk? Considering the diversity
of motivations in this group, I don't think it's wise to try to put
everyone on the exact same task of cleaning up 2.1-alpha-1 for
release...nor is it strictly necessary. If we have a codefreeze branch,
we can do the minimal updates to get alpha-1 out, porting these back to
trunk as well, do the release from that branch, then leave it dormant or
remove it until the next alpha release.
Thoughts?
-john
Brett Porter wrote:
I went through the "unscheduled" bunch of issues looking for reported
2.1 regressions this morning, and in what probably makes a clear
statement about my sanity, cleaned up a number of the 180 issues and
scheduled some for 2.0.x / 2.1 depending on their type (for those that
couldn't be closed out). I expect I missed a few duplicates, etc.
What remains for 2.1-alpha-1 is just the list of regressions, minor
updates, and backwards compatibility documentation (39 issues in
total, though some may be related):
* http://tinyurl.com/5fuwcm
* http://tinyurl.com/6k49gt
I believe we should start to knock these off, and prepare for an alpha
release as is, and wanted to see what others thought.
To cover the inevitable questions:
- Why release now?
163 fixes, 32 months since 2.0. 'Nuff said.
The only thing I care about right now is getting a release out that -
to the best of our knowledge - doesn't have any regressions from
2.0.x. This is really a necessary platform before making further,
bigger changes.
The sooner it gets out, the more time the current changes will have to
be exposed before a final 2.1 release.
- Why not add new feature X first?
There are a bunch of good things going on (I count 8 active feature
branches). As good as each is, none are ready to be merged on to trunk
today, and the time until they are ready is not predictable. Releasing
first doesn't harm the acceptance of these, or ability to make those
changes.
We just need more exposure to the changes that have already been made,
and to narrow down the potential sources of regressions.
So, thoughts?
Cheers,
Brett
--
Brett Porter
[EMAIL PROTECTED]
http://blogs.exist.com/bporter/
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
--
John Casey
Developer, PMC Member - Apache Maven (http://maven.apache.org)
Blog: http://www.ejlife.net/blogs/buildchimp/
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
