How to access Subversion from Oracle PL/SQL?
I think the basic flow makes sense. I would recommend doing experiments to see where exactly the performance bottlenecks are and then take if from there. For example, is it crossing from PL/SQL to Oracle JVM? Is it JVM shelling out to execute the svn command? Is it the svn round trip? Is it the parsing of the XML?
Let's say, for example, it's the svn round trip. Maybe you could have a process on the oracle machine that caches answers from the svn server so that at times the round trip could be avoided? Maybe the svn round trip could be async?
But, like I said, you need to know where the bottleneck is.