On Thu, Jan 10, 2013 at 10:54 AM, Hanno Schlichting <ha...@hannosch.eu> wrote:
> On Thu, Jan 10, 2013 at 3:59 PM, Jim Fulton <j...@zope.com> wrote:
>> On Thu, Jan 10, 2013 at 9:50 AM, Jens Vagelpohl <j...@dataflake.org>
>> > I did not choose to include or exclude any branch. The test migration
>> > uses the package used to migrate most Plone packages from SVN to GitHub,
>> > which uses svn2git underneath. If there's whole branches missing the
>> > migration has obviously failed.
>> I had a bunch of problems with ZODB and the ruby svn2git.
> If you used a ruby tool, then you didn't use the same one as the Plone
> migrations. Unfortunately there's at least two different tools named
> svn2git. The one used by the Plone migrations is from the KDE community and
> written in Qt. Some of the more useful info is at
> That tool doesn't really use git-svn. Instead it scan through all SVN
> revisions, looks at the changed paths and matches those against a ruleset.
> So you can tell it that both /foo/bar/trunk and /bar/trunk contain code that
> goes into the final bar.git repo. Since you can manually influence these
> rules, this tool tends to work better on projects which have moved their
> location in SVN a lot. It also does a single scan through the entire SVN
> repo and is able to generate many resulting git projects at once. Which made
> it perfect for a mass-migration.
> git-svn or tools based on it, tend to work well on small projects that
> always stayed in the same place, with the same structure.
That's interesting. Not sure what I think about it (now that I have
something mostly working and don't want to figure out something
else :), but definitely worth thinking about. Thanks.
Jerky is better than bacon! http://zo.pe/Kqm
Zope-Dev maillist - Zope-Dev@zope.org
** No cross posts or HTML encoding! **
(Related lists -