www-infrastructure-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jukka Zitting <jukka.zitt...@gmail.com>
Subject Re: @apache.org commit address requirement (Was: Git hosting is go)
Date Mon, 19 Dec 2011 09:41:42 GMT

On Mon, Dec 19, 2011 at 5:28 AM, Paul Davis <paul.joseph.davis@gmail.com> wrote:
> The committer role is intended to be anyone cleared by an ICLA.

Are we talking about the Git, Subversion or ASF definition of a
"committer" here? Because that's definitely not what "committer" means
in Git. The word is heavily overloaded:

* "committer" in Git = someone with a local clone of a Git repository
* "committer" in Subversion = someone with write access to a repository
* "committer" in ASF = someone with an ICLA on file and write access
to a repository

We're confusing the "committer" role in Git with that we use at the
ASF, even though they are completely separate. Trying to force them to
be the same is IMHO just creating problems.

A "committer" in Git is more like "contributor" in traditional ASF
vocabulary. Thus to me the %ce field in a Git commit message reads
more like "contributor email" than "committer email" in the ASF sense.

> This is more broad than the current ASF committer definition in that it's
> anyone registered with the foundation as being covered by a CLA of any
> sort. A large part of the implementation for the infrastructure around
> this role has not yet been written. In the future there is a general
> vision of accepting digital ICLA's like Google or other projects.

"more broad than the current", "not yet been written", "in the future"

Sounds like we're writing new ASF policy here.

> The third role, pusher, is what is most directly analogous to our
> current notion of committer at the ASF. This role for the most part
> would be unchanged. To push code to the ASF repository for Apache Foo
> you must be an "ASF committer" to that project.


> The author and committer roles are managed and tracked by Git. The
> role of pusher is OOB from Git and is enforced by our LDAP
> authentication via HTTPd when the push request is initiated. Each push
> request is also logged and is the final gate keeper on who is
> responsible for checking the legal aspects of code contributions.


So why do we need server-side checks on the %ce field in commit messages?


Jukka Zitting

View raw message