Øyvind Harboe wrote:
> A hardware engineer that tinkers with software should be
> able to contribute without becoming a git user.
Depends.. If they tinker with the git version of the software then
I think they *should* be a git user.
Tormod Volden wrote:
> letting the contributor discover the advantages of git step by step
> might be better than intimidating him with the legendary "git
> learning curve".
In my experience the git learning curve has nothing to do with what
commands to run. Anyone can run a sequence of commands. But the
difficulties I've seen people have with git are commonly about
workflow and processes. Even if git is only used on small scale it
helps to embrace the big scale workflows. This can feel like
unneccessary overhead, and git gets the blame for it.
I agree with a README.patches or HACKING file with brief
instructions. IMO, the important ones are:
git config --global user.name
git config --global user.email
while(1) {
work
git add work
git commit
}
git pull --rebase # to catch any conflicts
git format-patch origin/master..
This is a tradeoff between overhead when reverting changes, and
having to create a new branch for doing work. I really like to avoid
the latter for some reason. Branches can also be confusing, if the
whole git picture is not known (yet) - in those cases it's nice to
avoid creating a branch all together.
//Peter
_______________________________________________
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development
