Maintaining my own branch with git

[Steve O'Hara-Smith]

On Wed, 28 Apr 2021 13:49:23 +0200
Andrea Venturoli <ml at netfence.it> wrote:

> Hello.
> 
> I know tons of messages have been written on the subject: I read them 
> all, along with the Git Primer docs, but somehow I think I'm not doing 
> it right and I cannot understand what I'm doing wrong.
> 
> I need to have some patches applied to both src and ports and distribute 
> them to several machines.
> So I have my own git server and I want to keep my own FreeBSD branches.
> 
> I cloned FreeBSD's git repository and pushed into my server; then I 
> created my own branches.

	OK let's stop there. I would not do that.

	I would do this:

1: Clone the FreeBSD repository into my work area
2: Create my branch(es) off freebsd/main like this:
   git checkout main
   git checkout -b myworkingbranch
3: Apply my patches in myworkingbranch and commit them to that branch
   NB: git commit --amend and git rebase -i are your dear friends for
       keeping a clean history in your branch.
4: Push myworkingbranch to your server whenever you're done with a commit 
   Don't push the freebsd branches ever
5: Update main with git pull (while freebsd/main is checked out)
   NB: This will *always* be fast forward since you *never* commit to it
6: Update myworkingbranch with:
   git rebase main
   NB: This will stash your changes, bring in the updates and then reapply
       your changes, requiring you to sort out any conflicts that come up.
7: After this rebase you will need to force push myworkingbranch to
   your server

	Branches in git are cheap, easy and throwaway things make branches
at the drop of a hat to keep separate threads of work separate, make
integration branches to merge them into and test branches to load with
throwaway debug and never think twice about doing it.

-- 
Steve O'Hara-Smith <steve at sohara.org>