-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Roger Binns wrote: > I was under the impression that SQLite 3.5.2 is completely threadsafe, > meaning you can make any relevant API call in any thread.
I should be clearer what I meant by threadsafe. The most important notion is that there is no way to corrupt SQLite itself no matter which APIs are called in any thread or concurrently. I'd expect the misuse error if trying to do bad things such as simultaneously calling sqlite3_step in two different threads on the same statement at the same time (one call should succeed, the other getting misuse). The second is that APIs will work if called in the correct order but with the calls happening in different threads. If there is any possibility for corrupting SQLite internals based on what threads calls are happening in, or if calls can't happen in different threads then I'll need to leave the thread protection in. Roger -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFHSfHSmOOfHg372QQRAnNQAJ0T8uWac3zE9kaJ9c43bfuqE5SGCwCgwS/1 2m+FAVbJ4qTYOctlEaFyKjY= =67SP -----END PGP SIGNATURE----- ----------------------------------------------------------------------------- To unsubscribe, send email to [EMAIL PROTECTED] -----------------------------------------------------------------------------
