That said, it is my aim to keep bsddb in stdlib, providing a stable and featureful module. I think keeping bsddb development inside python svn is not appropiate.
I think it would be helpful if you could analyze the crashes that bsddb caused on Windows. Just go back a few revisions in the subversion tree to reproduce the crashes.
These were particularly bad since they invoked the CRT abort() inside bsddb, namely inside __db_win32_mutex_unlock, which, in debug mode, would __db_panic with "Win32 unlock failed: lock already unlocked".
This problem has now been worked-around, by reformulating the test cases so that the situation doesn't occur anymore, but IMO, it should not be possible for an extension module to cause an interpreter abort.