You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Ramkumar Ramachandra <ar...@gmail.com> on 2010/06/10 09:10:03 UTC

Getting data out of SVN quickly and efficiently

Hi,

I'm working on a project to convert SVN revisions to a stream that
git-fast-import can import. We have already built an excellent
exporter for an SVN dumpfile [1], because we figured that dumpfile is
the most reliable way to get data out of SVN. Now, to connect to the
SVN server and emit data in the dumpfile format, Jonas has already
written rsvndump [2], but I find it to be unnecessarily elaborate and
complicated for our purposes- I believe that by using
svn_ra_replay_range, it can be simplified to a great extent. I've
therefore just started out writing an SVN client from scratch [3], but
I have some conceptual difficulties. As I understand it, I should
attach an editor to replay revisions through, and map the actions
(like add_directory, add_file) to a printf statement that translates
that action to a representation in dumpfile format. (Why) do I need
delta builders and path hashers like the one that rsvndump implements?

Thanks.

-- Ram

[1]: http://github.com/artagnon/svn-dump-fast-export
[2]: http://github.com/jgehring/rsvndump
[3]: http://article.gmane.org/gmane.comp.version-control.git/148250