Re: onwards to git

[Thomas Schwinge <tschwinge at gnu dot org>]

Hello!

On Thu, Apr 30, 2009 at 02:19:40PM -0700, Roland McGrath wrote:
> > > We'll recreate the git archive once more from scratch.
> > 
> > Doing so, consider fixing this: there are, in May 1996 for example, a
> > bunch of commits attributed to mib <uco-mib@gmx.at>, which is wrong -- as
> > it can be seen from the committed ChangeLog entries, these commits should
> > be attributed to Michael I. Bushnell, p/BSG <mib@gnu.ai.mit.edu>.
> 
> In general we are not putting much effort into this sort of trivia.

Sure, but as Ulrich said that the repository would be recreated
nevertheless, and this seemed easy enough to fix, I thought that I'd
bring it up right now.

> The
> git metadata (authors, dates, etc.) from the past is not something we'll
> ever rely on.  (Indeed, if you go back far enough, the cvs metadata for
> that stuff is not entirely reliable either.)

Sure -- login names can change over time and that's all CVS / RCS
records.

> This particular case is easy to fix, since it's just an entry in the name
> translation table Jim already has.  That was a case where the sourceware
> user name changed (perhaps the only such case): the old "mib" user in cvs
> is the "tb" user today.  I think the most useful translation now is:
> 	Thomas Bushnell <thomas@gnu.org>
> So use that for mib, tb, and thomas.

Or use Thomas Bushnell <thomas@gnu.org> for tb and thomas, and Michael I.
Bushnell <mib@gnu.org> for mib -- which is, in fact, what I've been using
when converting the Hurd CVS repositories to Git.  <mib@gnu.org> is still
a vaild email address, and using it, plus the name Michael I. B.  is
historically more correct, I would say.  But now we discussed this to an
end, I guess.  :-)



Regards,
 Thomas

Attachment: signature.asc
Description: Digital signature