I sent him a driver that was recompiled with MAX_BUFFERED_QUERIES = 1 and he still appears to be having some problems, but I have no idea where.
I wonder if he is doing one transaction per batch update loop or something similar. If there's a bottleneck on the server side that's related to the number of updates in a transaction (deferred constraints or something similar?) then that could explain some of the behaviour. Don't know why changing MAX_BUFFERED_QUERIES would change things at all if that's the case, though.
-O
---------------------------(end of broadcast)--------------------------- TIP 5: Have you checked our extensive FAQ?
http://www.postgresql.org/docs/faqs/FAQ.html
