On Sep 8, 2011, at 9:37 AM, Victor Olex wrote: > I never for a moment thought that your change was thoughtless. To the > contrary, I have huge respect for SQLAlchemy. I will try to test the > drop_all and your pyodbc issue with my setup and to report here later > today.
thanks ! Unfortunately I've tested this some more and things are looking very, very bad. For us to support 0.91, we'd need to figure out how to get all of our "table exists" functions to work. If you look at http://www.sqlalchemy.org/trac/ticket/2273, I've now added a patch that detects 0.82 vs. 0.91 and sets the flag, but you can see that we can't send u'' strings when we query INFORMATION_SCHEMA still - literally, the number of characters present in one of the bind parameters changes the behavior. So there is something very strange and arbitrary (seems basically like it's just making guesses about datatypes) going on with the internals of FreeTDS, and I'm not optimistic about being able to get clear answers from their list. Would you have any resources to evaluate the test cases on that ticket , both are now against pure PyODBC 2.1.9? Without being able to query information schema, none of our unit tests can run period with 0.91 - I need a reliable way to do so, hopefully without losing support for table names that contain non-ascii characters. A lot of adjustments to the MSSQL dialect and testing will be needed. -- 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.
