Re: GCC revision history in CVSQL

[Daniel Jacobowitz <drow at mvista dot com>]

On Thu, Nov 14, 2002 at 06:39:23PM +1100, Ben Elliston wrote:
> I recently released a little package called CVSQL that, as its name
> implies, keeps CVS revision metadata in a database to be used as a
> discovery tool by users, by regression testing tools, or whatnot.
> Yesterday I imported all of the gcc.gnu.org revision history since day
> one into the database.
> 
> The database associates collections of patches into "changesets" so
> that patches can be meaningfully and completely extracted and, for
> regression testing, backed out.  The definition of a collection is
> "all revisions committed atomically by the user".  When changes are
> commited, CVSQL also discovers PR numbers and patch author email
> addresses in the commit messages and records those explicitly.
> 
> As you know, CVS operates on files and if you want to ask more complex
> questions, CVS employs recursion over the tree.  With this data in an
> auxillary SQL database, more expressive queries about the entire tree
> can be asked, and answered more quickly.
> 
> At present there is no externally available SQL database for this
> purpose, so I am using an internal database at Red Hat.  Here is a
> static page I prepared earlier :-) that shows commits over the last 24
> hours, just to give folks a feel for what it looks like:
> 
>         http://people.redhat.com/bje/cvsql/recent.html

Neat!  One thing - it would be nice to indicate branch commits
specially... but I really like this tool.  Is there any plan for an
externally accessible database?  Perhaps when the new machine for
sources is ready...

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer