On 1/21/2012 2:43 AM, Justin Sherrill wrote:
On Fri, Jan 20, 2012 at 2:35 PM, John Marino<[email protected]> wrote:
For the 2.12 release there was maintained a "must-fix" buglist. Did we have
this for the 3.0 release?
There isn't a separate one for 3.0, but that's in part because 2.12
sorta morphed into 3.0, and the 2.12 one was cleared.
A lot of new material was committed since the 2.12 branch. It's not
like we were in a freeze. So 3.0 is not a relabeled 2.12, and a new set
of bugs could have been created.
Plus I think a formal review of the buglist for each release should be
part of release planning. It forces us at least once every 6 months to
really review it, and give the users/developers a platform to opine on
what really needs to be fixed and what can wait.
I don't think bug 2279 is fixed yet. I would think that would hold up the
release? There might be more. We should review the bug list before tagging,
IMO.
YONETANI Tomokazu has a fix for it. I think Matt did some cleanup of
the build files to make it go faster - this may be related? I'm
guessing.
My goal was to tag so that anyone who wanted to put stuff in could,
without risk of destabilizing the release, and then bug fixes could be
applied back to that tag. That way we have something to work on while
the packages build.
Applying back to the releas branch is kind of a pain, and at least two
of us have made "git" mistakes doing it. Our goal (and I think it's
Matt goal too) is that the actual release is pretty similar to what's
tagged.
And yes, I'm also guessing that Matt's parallel build work is the cause
of bug 2279. But that bug report was just one example, there could
easily be more.
John
