[pysqlite] Concurrency - first steps

For a long time we've used Python locks to control access to a database
from multiple threads in PySQLite.

However, given the latest work on GIL in PySQLite 2.5.0 I'm looking at
moving to letting SQLite handle locking itself. I'm starting with a
subsidiary database which needs fast access and might have a lot of
throughput from many different worker threads.

A couple of questions:
The page
http://oss.itsystementwicklung.de/download/pysqlite/doc/sqlite3.html#sqlit
e3-controlling-transactions
 
talks about "SQLite’s supported isolation levels: DEFERRED, IMMEDIATE or
EXCLUSIVE". (Note, by the way, that EXCLUSIVE is given as EXLUSIVE
earlier on in the document).

However, http://www.sqlite.org/lockingv3.html lists UNLOCKED, SHARED,  
RESERVED, PENDING and EXCLUSIVE. What is the mapping to PySQLite's modes?


The timeout parameter to a connection enables automatic retry of a
statement. If that fails after a long time an exception is thrown. What
sort of behaviour can reasonably be implemented if the timeout period
expires? In our case I would just try to recommit - but should I look at
sending some sort of "too busy" message back to the caller? What do other
people do?

Hugh