Re: [gsoc] Merging completed tasks to develop branch in upstream repo

From: Mateusz Loskot (mateusz_at_[hidden])
Date: 2019-06-07 07:26:40

On Fri, 7 Jun 2019 at 02:44, Olzhas Zhumabek
<anonymous.from.applecity_at_[hidden]> wrote:
>
> I believe I didn't miss anything and read from start to end. I would like
> to work from my original fork.

Fine. As Stefan said, you are free to choose.

> At the end of GSoC, I could push the latest version to the repo in gsoc org,
> for exposition or whatever Boost might consider it for.

OK

> I would still like to have one branch before develop for me though,
> the gsoc2019 one.

As Stefan said, you are free to decide on your workflow.

As I mentioned earlier in [1] , the only question that is unclear to me is about
flow and direction of PRs for partial submissions (of completed features,
of course).

For example, if you will do all your work in gsoc2019 branch, then how to
approach the iterations of PRs for completed features for review? e.g.

1. Update gsoc2019 branch from upstream develop branch
2. Branch off gsoc2019 to create short-lived feature branch
3. Submit PR from feature branch to upstream develop branch for review
4. Once PR is merged into upstream develop, go to 1. to update your gsoc2019
5. Repeat steps 2 to 4

Although Stefan expressed he is not enthusiastic about formal guidelines,
I still believe it is important to outline some basic steps of workflow to avoid
any Git mess and try to maintain clean history, and to make your, the students,
life easier if a procedure is given as no-brainer instruction.
That is the question I asked Stefan in [1] to clarify.

[1] https://lists.boost.org/boost-gil/2019/06/0227.php

Best regards,
-- 
Mateusz Loskot, http://mateusz.loskot.net