Re: Two new Git policy checks

2009/4/16 Owen Taylor <otaylor redhat com>:
> I added two checks that get on each "new" commit that wasn't previously
> in the repository. (Excluding the import phase.) Like all the other
> Git/Help pages the referenced pages haven't been written yet.
> Check 1: User hasn't configured their email address
> ===================================================
> If a commit has an Author: email that ends in 'localhost.localdomain'
> or '(none)', then the push errors out with:

I suppose that Committer is also tested.

> ---
> The commits you are trying to push contain the author email
> address '$email'. Please configure your
> username and email address. See:


> For instructions about how to do this and how to fix your
> existing commits.


> Check 2: git pull introduced stray merges
> =========================================

What is wrong with these merges? They are correct and normal if you do
distributed commits.

> If a commit has a subject that matches
>  Merge branch.*of.*\(git\|ssh\):
> Then the commit errors out with:
> ---
> The commit:
> [git log output for commit]
> Looks like it was produced by typing 'git pull' without the --rebase
> option when you had local changes. Running 'git  pull --rebase' now
> will fix the problem.

s/local changes/local commits/

Possible additional text:

If you really want to push a merge you can make it with:

$ git merge otherbranch

> Then please try, 'git push' again. Please see:
> ---
> If you look through the commit lists, we are getting a lot of these
> stray merges, even from people that should know better :-)

They are not bad per se. Are they *all* introducing extra unwanted
commits? Then they should better check with:

git push --dry-run

what would be pushed, maybe with the help of gitk.

Mistakes happen, but they should no harm the workflow for everybody.

Maybe it is OK to do it while the people learn the new system, but no


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]