Re: In support of Jonas Bernoulli's Magit

[phillip.lord@russet.org.uk (Phillip Lord)]

Richard Stallman <rms@gnu.org> writes:

> [[[ To any NSA and FBI agents reading my email: please consider    ]]]
> [[[ whether defending the US Constitution against all enemies,     ]]]
> [[[ foreign or domestic, requires you to follow Snowden's example. ]]]
>
>   > No this is no good. If I am going to do this, I want notifications -- push
>   > not pull. So, there needs to be a way to open this up beyond maintainers.
>
> I am working on getting this set up.

Good. From what Stefan says, if there is already a ticket system
involved in copyright assignment, this might be as simple as enabling a
non-maintainer to start a ticket when they email someone a request.


>   > Having a system to maintain links to the package we are trying to get
>   > assignment for would be good.
>
> I am not sure what that means -- could you say it more concretely?

Yes. The copyright assignment process says "I assign copyright for
Emacs"; obviously this makes sense, but if someone is doing this so that
we can get magit into Emacs or ELPA, then we need to know that also, so
we can work out when a new assignment happens which package (i.e. magit)
it has implications for.


>   > Having something which handles something like blocking would be nice also.
>   > i.e. we start a bug wrt to magit for assignment, then all the assignment
>   > requests would be blockers. Then we would know when we are finished.
>
> How would this be more convenient
> than putting a list of the contributors in a file
> and making a note when each one comes in?

Say we have 200 contributors for magit, and 500 people with copyright
assignment for Emacs. New assignment comes in for Emacs, I now have to
check 200 magit contributors to see whether its one of them. To do
magit, we also need to do 5 dependencies, so we need to check against
those lists as well. It's scaling something like quadratically.

It's also worth remembering here: copyright assignment works based on
peoples names, while the list of contributors for magit that I have is
based on their github names, or the email(s) that they used for their
git commits. So making sure that any data that is stored, also includes
peoples aliases.

To re-iterate, this is just my ideas based on my last experience; it's
going to take several attempts and refinements to get the process
working well.

Phil