Re: [sqlite] Of shared cache, table locks and transactions

On 1/14/07, Dan Kennedy <[EMAIL PROTECTED]> wrote: So you're right, when you use shared-cache mode there is less concurrency in some circumstances. Thanks for your confirmation and additional comments on the situation! It's good to know that I'm interpreting the source correctly at least.

AW: [sqlite] Of shared cache, table locks and transactions

] Of shared cache, table locks and transactions Yes exactly my point for some time now!!! I think it would be nice if sqlite could optionally maintain a pre-write state version of the cached page in memory. This really means maintaining some form of page versioning, which is already done via

[sqlite] Of shared cache, table locks and transactions

Hey folks... I have a situation that caused me a little head-scratching and I'm wondering if it's intended behavior or not. I'm running a server thread (roughly based on test_server.c in the distro) on top of the 3.3.6 library. The effectve call sequence in question (all from one thread) looks