[Csync2] recursive selects with sqlite

[Art -kwaak- van Breemen]

On Fri, Feb 20, 2009 at 11:03:27AM +0100, Lars Ellenberg wrote:
> > Anyway: the sqlite locking is even in my version still a pain. I
> > guess that for high update and > M files setups we should look at a
> > different approach for keeping syncs.
> 
> suggestions?

I think the final hack that is possible is to split the database
into seperate files, and have separate handles per file. For
each peer a seperate dirty db.
(I am talking about 10...200 updates/second with a duration of an
hour with 3 peers.)
The next step is to keep it a single database file, and have a
seperate daemon to handle database queries.
Or better: a single daemon handles incoming connections and using
a socket handles -h, -m and -c .
On incoming connections, that daemon should handle file state
while children are handling incoming connections or -x/seperate
updates.

-- 
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in email?