Problem turns out to have been with my ISAPI WSGI interface, it looks
like it has a broken thread local model. More details here if anyone
is interested:
http://groups.google.com/group/sqlalchemy/browse_thread/thread/fbca1399020f6a2e
On Nov 6, 5:19 pm, Randy Syring [EMAIL PROTECTED] wrote:
Yes, I would also like to know what is the appropriate way to use SQLAlchemy
with respect to a WSGI server. I've been using Django with SQLAlchemy (not
yet supported, but the recipe here
http://lethain.com/entry/2008/jul/23/replacing-django-s-orm-with-sqlalchemy/got
me moving), and it's not clear
On Nov 6, 2008, at 1:29 PM, Randy Syring wrote:
I am developing a WSGI based web framework with sqlalchemy. I am
unclear about when create_engine() should be called. I initially
thought that engine creation and metadata would be initialized per
process and each thread/request would just
if you're using create_engine(), you're using a connection pool. The
diagram at http://www.sqlalchemy.org/docs/05/dbengine.html illustrates
this.
On Nov 6, 2008, at 1:35 PM, John Fries wrote:
Yes, I would also like to know what is the appropriate way to use
SQLAlchemy with respect to a
Thank you for taking the time to respond, I really appreciate it!
On Nov 6, 2:46 pm, Michael Bayer [EMAIL PROTECTED] wrote:
you should definitely create the engine and metadata on a per-process
basis. When using SQLite, the engine automatically chooses the
SingletonThreadPool
