On 14 Jan 2019, at 6:34pm, Thomas Kurz <sqlite.2...@t-net.ruhr> wrote:
>> pragma_busy_timeout > > Does setting the busy_timeout retry periodically (e.g. every x milliseconds), > or is there some automatism that ensures that the requested operation is done > just-in-time as soon as the previous/blocking operation is finished? The busy_timeout setting is the maximum amount of time that the busy_timeout routine is allowed to run. That routine is not expected to sleep for that amount of time you set in a single sleep. SQLite's own busy_timeout routine (the one you get if you don't supply your own) uses exponential backoff. It first sleeps for a very small amount of time, then checks the lock. If access is still prevented it sleeps longer, then checks again. Then longer still, then longer still, until the amount of time /since the routine started/ exceeds the timeout you set. The "just-in-time" idea mentioned in your question doesn't work in real life, since constantly checking mutex status keeps one core completely busy, using lots of power and generating lots of heat. Simon. _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
