Re: Git history author/committer fields (was Re: msdtc with 32-bit app fails to resolve in-doubt or not-notifed transactions)

On 06/26/2014 03:49 PM, Heikki Linnakangas wrote:
> On 06/25/2014 05:26 PM, Craig Ringer wrote:
>> Please merge branch fix-syswow64-msdtc from my repo at
>> https://github.com/ringerc/psqlODBC.git .
>
> Oh, and while you do that, please avoid creating a merge commit. I hate
> those merge commits when I browse the git history, they're just useless
> noise. Please use "git rebase" to clean up the history in your local
> repository first. You can use "git push --dry-run" before the actual
> push to see what will be done. It will print the commit IDs that it
> would push, something like this:
>
> a87a7dc..de42ed4  master     -> origin/master
>
> You can then do "git log a87a7dc..de42ed4" to see those commits.

I usually do a rebase to current master before pushing a feature branch
and requesting a merge. I'm surprised a merge commit was created -
someone must've committed other things between when I pushed and when it
was merged. There was a lot happening around then.

The committer may just:

    git merge --ff-only

to reject commits that aren't simply fast-forwards.


Or, as rebase merges are usually trivial, they can:

git checkout theremote/thebranch
git checkout -b thebranch
git rebase master
git checkout master
git merge --ff-only -

which will just rebase on top and merge.

> Also, it would be good to reset the author/committer information in the
> commit before pushing. Looking at the git history the other day, I was
> quite surprised to see commits from Craig, as I didn't know he's a
> psqlodbc committer.

Heh. Soon I shall have commit to all the drivers, then my evil plans
will finally be put in motion!

> That's what we do with the PostgreSQL main repository, anyway. We could
> also decide to use the git's Author field to track the original author,
> but at least the Committer field should reflect the actual psqlodbc
> committer. I'm not sure how to change that without changing the Author
> field, though.

I had a quick look without seeing anything. The two fields must exist
separately for some reason, though.

--
 Craig Ringer                   http://www.2ndQuadrant.com/
 PostgreSQL Development, 24x7 Support, Training & Services