I want to set up an isolated lab where I read the software sources, execute for a bit, and see what happens.

In my particular scenario I will use an earlier version of the reference bitcoin client and download about 10,000 blocks.  The difficulty will be 1 at this point.

I will use cgminer and ckpool as well, to develop what skill I can at understanding stratum and how to configure the servers and what they do.

Toward that end I have made the most recent version of ckpool, and started it with
  ckpool -l 7 -A -H

and ckpool.conf points only to my local setup.

I appear to retreive getblocktemplates, pass info up to cpuminer/minerd, do the magic and send results back to ckpool.  ckpool.log and the consoles for cpuminer and ckpool make sense to me.

I don't grok how to get the new block submitted to my local bitcoind.

For what it's worth, I am looking at line 6225 of stratum.c.  That seems to be one place where I might expect a submitblock to happen, but it does not.

My understanding of "upstream" and whether it is necessary / possible to set up a proxy to make this happen is a bit wobbly.

Oh, and while I am posting, I thought I would mention that I am pretty familiar with GitHub, but not so much with the repository where ck has his source located.  Is there a GitHub bridge (fork) to his efforts, or do I need to figure out the new repository as part of all this effort.