Re: GitHub “pull request” is proprietary, incompatible with Git ‘request-pull ’
Ben Finney writes ("Re: GitHub “pull request” is proprietary, incompatible with Git ‘request-pull ’"):
> A putative decentralised [0] Git pull request feature would IMO require
> that anyone with a Git repository can submit a pull request to any
> other, without any registration on a privileged central server.
Yes.
> My understanding is that Git ‘request-pull’ satisfies this requirement
> for decentralisation, by using two decentralised protocols: Git, and
> a formatted plain text file.
However, it has two important downsides:
* The person sending the pull request must have a git server. This
is not inherently a property of the problem, nor of the underlying
git machinery. I think it is undesirable.
* The vital statistics of the pull request are transmitted by email.
> Once a hosted service decides to use Git ‘request-pull’ on which to
> build its pull request feature, [...]
Perhaps this is the reason we seem to be talking past each other.
I don't think this is a good idea. git request-pull is not a
protocol, it's a tool for helping generate a moderately standard form
of email. Trying to turn it into a protocol is not going to end well.
Furthermore, you seem to be under the impression that it is necessary
for the world to agree on a single protocol. That's not true. It is
not even necessary for the UI to be particularly similar.
What's necessary is for the submitter to be able to do whatever the
maintainer asks, but that could easily be a web page written by the
maintainer saying "to make a pull request, run the following git push
rune".
> So what seems to be lacking is:
I think we have very different ideas of what the problem is and how it
might be solved. If you wish to go ahead and try to make some kind of
protocol layered on top of git request-pull, I wish you luck, but I'm
afraid I don't have effort to provide concrete help for such a
project.
Ian.
Reply to: