As you know I'm running with pyPgSQL as well. I actually haven't built
my connection
tracking system yet but I've got a pretty strong idea of how I'm going to do it.
My design currently
integrates web connections and database connections (which are short term) so it
doesn't
necessarily scale well but the design can be extended for a larger app. The main
concern I have is
a cookie-less implementation so that knocks me out of using the pre-existing
mechanisms. As far as
mutlithreading is concerned, I haven't seen any issues - the main consideration
is that your app
stays alive between requests (unlike a pure cgi model) which I like but puts
tracking who the app is
talking to more on the developer (within the connection tracking system). I'll
have this implemented
and in use sometime in May and would be happy to share the mechanism. Meanwhile
my screen
layout bites and I don't have my head around a good solution as of yet - thus my
question. :-)

        thanx & later,

                Ben Scherrey

4/24/2003 9:25:13 PM, Gerhard Häring  wrote:

>Benjamin Scherrey wrote:
>> I've got a fairly functional system going now with my medusa/quixote/postgres
app [...]
>
>I'm deloping such a beast myself currently (with pyPgSQL, of course ;-).
>How do you handle database connections? Do you use a connection pool of
>some sorts? Is there a usable premade one or did you build it yourself?
>
>Can I have concurrent requests with Quixote, anyway? In production, I
>currently use the Medusa thingie, but in production I'd like to use the
>SCGI adapter. Do each of these do multithreading? A look into the source
>code didn't get me an immediate answer.
>
>-- Gerhard
>
>_______________________________________________
>Quixote-users mailing list
>Quixote-users@mems-exchange.org
>http://mail.mems-exchange.org/mailman/listinfo/quixote-users
>