coda on the jump address stuff:

it turns out my southbridge reads a data table to initialise the cpu before boot. providing this table seems to get the new layout to run at boot (although earlier changing the jump address seemed to do it, and no longer does. I infer I was cross-eyed again earlier, but a third check is needed.)

I still don't understand all the parts, but the southbridge information is enough to make the problem much more manageable.

--

regarding the programmer, the headers that arrived are square pin. I can use male-female jumper wires though. apparently I also ordered backup sockets, so I can in theory troubleshoot the hardware programmer issue now, by swapping components.

to do these tests I was flashing with the internal functionality of the board itself which involves a boot cycle every test.