[1.7] Plans for bzr 1.7
John Arbash Meinel
john at arbash-meinel.com
Thu Aug 14 17:50:01 BST 2008
-----BEGIN PGP SIGNED MESSAGE-----
It seems Martin has put me in charge of the 1.7 release, and I'd really
like to get back into the habit of having our metronome emails.
So this email is serving four purposes:
1) Describe the planned release dates, and get everyone to agree to get
back into a time-based release schedule. Martin's idea was to have:
freeze 28 August
1.7rc1 4 September
1.7 11 September
This gives us 2 weeks of open development for the 1.7 release cycle.
Which should be plenty.
2) Solicit feedback about what people want to get into the 1.7 release.
I don't necessarily know the details about what everyone is working on,
so if I've missed anything, please respond to this thread.
This also includes high priority bugs etc. So even if you can't fix it,
if you feel it is important please start a sub-thread.
3) Describe what I think should land in 1.7, and try to convince people
to work towards these goals.
Looking at the BB pending page, we have quite a few approved patches
that haven't been merged yet. I'll try to take care of some of this.
Remember, though, that if you are the last reviewer to approve
something, you're the one who is supposed to merge it.
Things I know about:
a) Robert has done a lot of great work on btree indexes and
groupcompress. The btree index stuff is pretty production ready,
and shows good improvement in lots of cases (and I don't believe
has any regressions.)
I'm planning on working to bring in at least btree into a --dev
format in bzr core. I think groupcompress might need a bit more
baking as a plugin. But if Robert is motivated, I would certainly
be interested in reviewing it.
b) I have a tweak to the Merge code, to handle path resolution better
after a criss-cross merge. It just needs someone to review it.
c) I'd really like to get Mark Hammond's win32 installer polished, and
get his win32 test fixes merged into core. It would be wonderful if
we could get to 0 test failures in this release, though I don't
think that is going to specifically happen.
d) Speaking of which, would anyone want to champion getting the test
suite to pass cleanly on Mac OSX? I don't use it as my primary
platform anymore, so I don't run into it as often.
e) New support for things like '.bzrrules' that we want to propagate
between branches, merge, etc, but aren't really user files that
belong with the tree.
Robert has offered to do some of this to help Ian with getting
content filtering, etc, merged.
f) Andrew Bennetts is working on doing effort tests, and some
iterative improvements to the HPSS protocol to make things like
'bzr push' faster, also to make it cheaper to open a remote object.
I think he wanted to do it under the effort framework, so I'd like
to at least get the basic effort testing into 1.7.
4) Poking at all the core devs that have approved patches that they
haven't merged yet.
Robert: (some of these I will probably merge as part of finishing up 1.6
and getting it merged back into bzr.dev, but some are just patches that
have been sitting for a while.)
Andrew: (old approved patch)
though with removing support for .bzrrules this may no longer be relevant.
I'll try to do Martin's patch, since he is on vacation. And I might do
some of Andrew's if he doesn't get to them before he leaves.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
-----END PGP SIGNATURE-----
More information about the bazaar