Make it possible to run CVSAnalY in parallel

[andygrunwald]

With the current version of CVSAnaly you can only run the tool once per database at the same time.
But you can analyze various repositories in one database.

For large analysis suites and a message driven and scalable systems it might be useful to run several cvsanaly process at the same time.
I got a real world use case for this.

One problem is the creation of a temp table with a fixed name. We should add a random postfix to the table name in such way to enable parallel processing.
I do not know if there are more bottlenecks. This has to be researched.

[sduenas]

I have the memory that with the current architecture it will really hard. But it is just a feeling. I do not remember well the details, although I remember problems with temporal tables (you also point at it) and key tables, that are generated by CVSAnalY and not by the DBMSs.