On Jun 1, 2011, at 7:00 PM, Jon Nelson wrote:
> On Wed, Jun 1, 2011 at 4:47 PM, Michael Bayer <mike...@zzzcomputing.com>
> wrote:
>> metadata.reflect() would close() the connection used for reflection as it
>> assumed it was passed an Engine, not a Connection, fixed in r926ee70b67ff.
>> Nothing to do with begin_nested() or anything like that.
>
> Why did this work in 0.6.[1,2,3,4,5,6,7] ?
> What's the "right" way to do this?
This is a brief summary of metadata.reflect() from 0.5-0.6 (was only introduced
in 0.5)
def reflect(bind):
conn = bind.connect()
go_reflect_the_tables(conn)
Conspicuously missing is a corresponding close() of the connection. In 0.7
we got all 3000 something tests to pass on Pypy. Pypy uses generational
garbage collection, which means it does not garbage collect anything until the
GC runs periodically. A half dozen tests against metadata.reflect() and the
connection pool would freeze, as each test left a connection hanging open.
We'd run out of connections.
So here was the fix in 0.7:
def reflect(bind):
conn = bind.connect()
go_reflect_the_tables(conn)
conn.close()
fixed ! Except it broke your app that happened to be passing a Connection, not
an Engine (Connection has a connect() method too so "bind" can be either), and
we closed your connection on you. We had no tests for that. Now we do. Fixed
again !
--
You received this message because you are subscribed to the Google Groups
"sqlalchemy" group.
To post to this group, send email to sqlalchemy@googlegroups.com.
To unsubscribe from this group, send email to
sqlalchemy+unsubscr...@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/sqlalchemy?hl=en.
