On Sep 12, 2014, at 12:26 PM, Matthijs Blaas <[email protected]> wrote:
> > set the search path within the connection pool using the "connect" event; > > all connections used by the pool will then have this search path setting > > the moment they connect. Don't ever put multiple SQL expressions in a DBAPI > > execute (for this very reason - driver changes, hilarity ensues. We are > > having a crisis in Openstack right now because they made the same mistake). > Thanks, i'll update our code to use the NullPool's connect event. > > > no part of your code should be changing search_path - I'd grep for it. > > I did and made sure there is only one place where it's set. But i have good > hopes the connect event takes care of it, i haven't been able to reproduce > the error anymore. > > I'm curious though, i was under the impression that when you start every > request in a transaction (pyramid_tm in this case), and set the search_path > on every transaction (using the sessions after_begin event), is it even > possible that a new connection is triggered somewhere that isn't included in > the main transaction? SQLAlchemy never does that but I don't know what you'd be doing in your code, it's certainly a pattern that is supported and possible, just wouldn't happen implicitly. -- You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/sqlalchemy. For more options, visit https://groups.google.com/d/optout.
