> There seems to be a few misconceptions in the article regarding the > difference between DEFERRED/IMMEDIATE/EXCLUSIVE
Looks like I mistakes in the 2PL section, it had contradictory content about `RESERVED` locks. Thanks for your feedback, have pushed corrections. > Specifically, your diagrams suggest that once a writer obtains a RESERVED lock (as happens when an IMMEDIATE transaction begins), no readers will be able to obtain a SHARED lock. The idea was to portray a scenario which forces serial execution of transactions. If all transactions are started in either `IMMEDIATE`/`EXCLUSIVE` modes, they try to acquire `RESERVED` /`EXCLUSIVE` locks which block other `IMMEDIATE` and `EXCLUSIVE` transactions. > It looks like you go into more detail on the locks later in the article - > only got to the "Shared cache mode" section. The sections on locking and WAL were meant to build up concepts like deadlocks and stale snapshots (discussed later in article), and the fact that SQLite can’t make progress on transactions by retrying individual queries in those cases. Rahul _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
