[Building Sakai] Installing salora-csv - problems

[Gregory Guthrie]

After several iterations, I wanted to clarify and re-query the questions that I have in this case.

I installed Sakai-2.9.1 (source), and tomcat 7.0.21, all work fine. Then I downloaded and installed Sakora-csv (into the sakai-src home directory), did the mvn build in its own home (sub-)directory, and then did the mvn deploy from sakai-src again.

All seemed to go well and report success, but the sakora-csv tests fail. I find that the additional database tables it uses were never created in MySql and that causes the failures.

1) Who creates (or should have created) these new tables, and when?

2) Once I install/deploy sakora into the sakai-src home directory, do I need to manually rebuild/deploy csv from its home (sub-)directory ever again, or will the sakai master mvn automatically pick-up all such sub-projects and run the mvn builds from them?
That is a common approach in make - but I don't know how the mvn configurations is setup.
I presume this same question would apply to any additional contrib projects that get added to sakai before a source build; do they automatically get handled by the master mvn of sakai or do they require manual re-makes.

3) Is the order important here? Did Sakai build all tables (only) at its first startup, and so later deploy's skip that step? I.e. do I need to install svn and deploy before the first sakai mvn-deploy?
If so, what does that mean, I have to remove all previous installs and start over anytime I add a sub-feature to Sakai?

4) Also all of the cm_* tables are empty - I presume that is ok?

-------------------------------------------