Great write-up!

Something I'm not getting, perhaps I'm missing something terribly obvious:

> during a full GC, we should also accept l_markVersion that is off-by-one. In that case, we know the isMarked bit accurately reflects whether or not a cell is live, since that is the result of the last GC cycle.

Here's part of the related footnote:

> any number of eden GC cycles may have run between the last full GC cycle and the current full GC cycle, but eden GC does not bump mark version. So for any object born before the last GC cycle (no matter eden or full), the isMarked bit honestly reflects if it is live

Couldn't an object have become non-live since the last GC cycle (whether eden or full)?