Thanks, Dave. I noticed the following comment in the code for the declaration of BTree#maxRowsPerPage:
RESOLVE (mikem) - this should not be static. Need to design a way in
debugging mode to get btree created with a persistent "maxRowsPerPage".
This hack makes all btrees get created with the "last" maxRowsPerPage
so that may be why the access pattern is the way it is. Mike, do you think the access code in LeafControlRow should be changed to reflect reality, or would you want to keep it the way it is until this gets resolved? It wouldn't do any hard to change it, were it changed to be a state variable, the compiler would flag the access via class.