On Tue, Oct 05, 2010 at 12:44:59PM +0200, Zaher Dirkey scratched on the wall: > On Tue, Oct 5, 2010 at 4:55 AM, Jay A. Kreibich <j...@kreibi.ch> wrote: > > > On Mon, Oct 04, 2010 at 07:25:05PM -0700, Dustin Sallings scratched on the > > wall: > > > > > > The main difference between =1 and =2 is that =2 assumes you more or > > less know what you're doing and will either lock a database handle as > > you pass it between threads or you'll keep it private to a thread. > > > > =1 is designed to be more or less idiot proof, and will simply not > > let you do something stupid with a database handle, like have two > > threads try to execute two different statements at the same time. > > > > What if two therad make insert or update then commit, what is happen to the > second thread after the first made commit?.
The second thread will block if it attempts to use a database connection that the first thread is already using. Once the first thread finishes (either with an explicit COMMIT/ROLLBACK, or because all auto-commit transactions go out of scope) then the first thread will release the database connection and the second thread will wake up and allowed to proceed. That's why the =1 mode is called "serial"... it automatically serializes the database statements. At least, I'm pretty sure that's how it works. I generally avoid threaded code, and when I do use it, I tend to use thread-specific resources that are carefully locked and guarded. -j -- Jay A. Kreibich < J A Y @ K R E I B I.C H > "Intelligence is like underwear: it is important that you have it, but showing it to the wrong people has the tendency to make them feel uncomfortable." -- Angela Johnson _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
