Philip indicated yesterday on IRC that all local move features
he wanted to get to implemented for 1.8 now have code written.
So now we're all waiting for the "local moves" yellow light on
roadmap.html to turn green.
I think we're at the point where we need testing by real users.
We've got some test coverage in our test suite that indicates things
are all well. But the test suite is somewhat removed from the actual use
cases which local moves are supposed to improve. For instance, we
haven't yet seen a single off the mill Java developer hit the 'refactor'
button in an IDE and then update the working copy.
I expect that there are bugs lurking in the implementation of local
moves, and I also expect that we can still improve the user interface
a bit. But to get there we need some feedback from the outside world
about what does and what doesn't work.
I think now is the time where we should either start cutting 1.8 alpha
releases from trunk, or branch 1.8 and then start issuing alpha releases
from the branch. In either case we should ask the community to build
binaries and ask our user base to help test them. (I don't expect alpha
builds to be used in IDEs, however many users could try to update
refactored working copies with the command line client and report back.
This would also be a good test for the 'svn upgrade' code.)
What do you all think?
If we start cutting alpha releases from trunk there will be less overhead
with committing fixes, which means we might stabilise the code a bit faster.
However, it might be worthwhile to force fixes through the backport
review process to make sure we get more eyes on them, which means we
might stabilise the code even more at the expense of potentially delaying
the final dot zero release a bit further.
I'd be happy of course to drive the initial alpha releases as soon as
we've decided on whether to make them and how we are going to manage them.
Received on 2013-02-14 18:22:32 CET