Maintaining patchset with GIT (was Re: Hot standby, RestoreBkpBlocks and cleanup locks)

Simon Riggs wrote:
> Do you want me to start using the GIT repo to make it easier to
extract parts?

It would be useful. Even more so for your own sanity, I think :-). I
find maintaining multiple interdependent patches much easier with GIT,
though it's still easy to get confused.

Feel free to continue with CVS, of course.

> You'd need to show me the setup you use first.

Well, the first thing to do is to clone the repository, see wiki for
that. And get an account at git.postgresql.org so that you can publish
your stuff as a git repository. (I should get one myself..)

For reviewing hot standby, I created one "hotstandbyv6a" branch, and
applied and committed all the patches in right order. But if you want to
keep the patches separate, you should create a separate branch for each.

If patch B depends on patch A, create branch A first, and then branch
branch B from that. That's what I did for the relation forks and FSM
work. Just remember to always commit to the right branch. Whenever you
commit changes to branch A, also merge those changes to branch B with
"git checkout B; git merge A". To sync with PostgreSQL CVS HEAD: "git
merge origin/master"

To generate diffs, you can do for example "git diff A..B" to create a
diff between A and B, or "git diff origin/master..A" to create a diff
between PostgreSQL CVS HEAD and A. "git log" is also very helpful.

There's a learning curve, but don't hesitate to ask.

--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com