> \[a\] commit says >> This provides a way for threads to share an in-memory database <b>without the use</b> of shared-cache mode > where is the truth? Regarding my statement: > To me, looking at the code, it appears that the memdb specification variations all bottom-out in the common sqlite_open*() implementation with SQLITE_OPEN_SHAREDCACHE passed in among the flags. I am going to document what these open options do, after a much more careful study of the code to see what combinations of flags reach the underlying open(). That study may induce a walk-back of the above. Or not. In my opinion, a shared in-memory store and a shared-cache mode are largely redundant. If the performance advantages all favor one or the other, one should vanish as a needless pitfall. And if they trade against each other, that should also be documented.