This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: 33 unknowns left
- From: Joseph Myers <joseph at codesourcery dot com>
- To: Ian Lance Taylor <iant at google dot com>
- Cc: Eric Raymond <esr at thyrsus dot com>, GCC Development <gcc at gcc dot gnu dot org>
- Date: Wed, 26 Aug 2015 21:04:41 +0000
- Subject: Re: 33 unknowns left
- Authentication-results: sourceware.org; auth=none
- References: <20150826193120 dot GA4593 at thyrsus dot com> <alpine dot DEB dot 2 dot 10 dot 1508261955070 dot 6767 at digraph dot polyomino dot org dot uk> <20150826203507 dot GC4945 at thyrsus dot com> <CAKOQZ8wC=bL+LFsxKpAXThKH1zR2vL_eyKRnf3vfKyPnpGaTAw at mail dot gmail dot com>
On Wed, 26 Aug 2015, Ian Lance Taylor wrote:
> On Wed, Aug 26, 2015 at 1:35 PM, Eric S. Raymond <esr@thyrsus.com> wrote:
>
> > root = root <root>
>
> Hmmm. root. This was probably Danny Berlin logged into the root
> account on sourceware.
The root commits I see from October 2005 (implementing parts of the
CVS-to-SVN transition) are pretty definitely him, since he was in charge
of that transition.
Two of the three root commits then are to the /hooks directory. That
directory should be omitted from the conversion (it doesn't make sense to
convert SVN hooks to git, any more than we converted the CVSROOT files to
SVN - Jason will be working on the git hooks needed).
At some point we need to work out the mapping for which subdirectories of
/branches are actual branches and which contain branches a level lower
down. In general I expect that subdirectories containing a ChangeLog file
are branches and those that don't are not branches, but the list of those
that are not branches should be reviewed manually (premerge-fsf-branch at
least *is* a branch not a container, despite only containing a directory).
In general, we should have a public git repository somewhere with all the
configuration for the conversion, such as the author map and the
configuration for branches.
--
Joseph S. Myers
joseph@codesourcery.com