No duplicates, no deletes, but insert order could be different.
Then hash collisions can differ. Hence resizing decisions can differ. Therefore database sizes (and internal order of free space re-use) can differ.
Still I notice that there was some discrepancy in the size of the database so that suggests something was different.
This is normal with hashing, for the reasons described above.
-jcw
_______________________________________________ metakit mailing list - [EMAIL PROTECTED] http://www.equi4.com/mailman/listinfo/metakit
