On Thu, 2004-08-05 at 11:03, Mark Phippard wrote:
> François Beausoleil <fbeausoleil@ftml.net> wrote on 08/05/2004 10:40:38
> AM:
>
> > Mark Phippard wrote:
> > > I want to create a database schema where I will log all of my commits,
> via
> > > a hook script, and perhaps also capture and store other statistics in
> the
> > > process such as diff stats. My ultimate goal is to drive a
> > > ViewCVS/WebSVN-type web application written in Java/JSP. Ideally, I
> would
> > > like to have the web app work completely off the database, as opposed
> to
> > > talking to the actual repository. Exception would be pages that
> display
> > > file contents etc...
> >
> > Take a look at Trac:
> > http://trac.edgewall.com/
> >
> > It has everything you mention, and quite a bit more.
> >
>
> I do not think Trac stores Subversion info in their database (with the
> possible exception of revision numbers), it looks like its browser is
> using the Python bindings to browse the repository directly. That is not
> what I want. The Trac schema just has tables for the the Trac stuff like
> their tickets and Wiki. I am looking for an example schema that can
> mirror the Subversion repository structue, in addition to the commit info.
>
> Mark
>
i don't think you looked very closely then. trac indexes into its
database each svn revision, to allow for unified searching across wiki,
issues, and revisions as well other nicetites, like handling issue
followups in log messages. it does this indexing (delta index against
new revisions) every time it opens the repository, but it can optionally
do this from a commit script, see the following link for the schema.
http://projects.edgewall.com/trac/file/trunk/trac/db_default.py
-kapil
Received on Mon Aug 16 06:44:56 2004