On 6/15/06, James Butts <[EMAIL PROTECTED]> wrote: > > In the main thread, I open a connection to database and create a timer > function and start a worker thread. The worker thread establishes its own > separate connection to database. The worker thread creates a table and > then signals the main thread. The timer function in the main thread drops > the table and signals back to the worker thread. The worker thread > creates the table again and receives no error. However, if the worker > thread tries to insert a record into that table, I get back a SQLITE_ERROR > error. >
>Are you perhaps getting a "schema changed" error? >It has to recache the new schema after the drop. >If the other thread has it open that can easily happen. I'm not sure how that is possible. My example code used sqlite3_exec which internally handles schema changed errors by retrying. The API call sqlite3_exec also finalizes each command. My code is set up lock-step so work is only being performed in one thread at any point in time. -- View this message in context: http://www.nabble.com/Multiple-Threads-with-a-DROP-command-t1792339.html#a4883955 Sent from the SQLite forum at Nabble.com.
