Rob Lohman wrote:
This all starts to feel a bit "unsafe". What would happen if a program would load a library that houses the callback and it gets unloaded after the first prepare? Or is it a rule you need to have the callback available for the entire run of the virtual machine?
It would have to be a "rule" as you say. It is really no different than what we have now. You can't unload an authorizer between the calls to sqlite3_set_authorizer(func) and sqlite3_set_authorizer(NULL). This will be basically the same except the calls are sqlite3_prepare_v2() and sqlite3_finalize().
Dennis Cote
