Format for JDK 6/7 changeset comments?

[Mark Reinhold]

> Date: Fri, 09 Nov 2007 14:28:54 +0300
> From: yuri.nesterenko at sun.com

> Should there be a convention for the name of author?
> A person should be easily found by the name.
> 
> I'm asking that because the name will be written in developer's
> .hgrc file and [another name or nick]  used for push.
> No sense for me to put there 'yuri' as in my OpenJDK public keys collection:
> it's too common a name in our part of the world.

The current thinking is that anyone who has authorship rights will have a
username in the OpenJDK web infrastructure.  That's the name that you'll
set in your .hgrc so that it appears in the author field of changesets
that you create, and it'll also be the name you use to push changesets
via ssh.

So while "yuri" may be common in your part of the world, within the
OpenJDK world it can be yours if you claim it first.

Most people will find it most convenient to use their regular Unix
username, assuming it isn't already taken.

(Note to Sun employees: Borgified usernames of the form xy12345 will,
 per internal policy, not be permitted.)

> E-mail? Will it be obfuscated?

The only people we need to identify by their e-mail addresses are
contributors who don't yet have authorship rights.  Is it worth
obfuscating those in the changeset comments themselves?  Any technique we
use will be obvious to anyone who's so desperate for e-mail addresses
that they're willing to clone an hg tree.  I'd be inclined to use actual
addresses in changeset comments but obfuscate them in the web interface.

Or we could identify such contributors only by their full names, but then
it'd be hard to find them.

>                                At least a gatekeeper should be able to
> quickly find the submitter without research.

People who have authorship rights will be able to look up the e-mail
address of anyone else who has such rights.

- Mark