[sqlalchemy] 0.9.7: Intermittently find a dead weakref to a listens_for function while checking _stored_in_collection of a registry

2015-04-02 Thread Evan James
Hi folks, While running my test suite, I hit an issue with the following stack trace: ERROR at setup of test_track_before_delete request = SubRequest 'database_session' for Function 'test_track_before_delete', engine = Engine(sqlite:///test.db) @pytest.fixture def

Re: [sqlalchemy] 0.9.7: Intermittently find a dead weakref to a listens_for function while checking _stored_in_collection of a registry

2015-04-02 Thread Mike Bayer
On 4/2/15 4:28 PM, Evan James wrote: Hi folks, While running my test suite, I hit an issue with the following stack trace: | ERROR at setup of test_track_before_delete request =SubRequest'database_session'forFunction'test_track_before_delete',engine =Engine(sqlite:///test.db)

Re: [sqlalchemy] Inherited class column override

2015-04-02 Thread Pierre B
Seems like exactly what I need! Many thanks for all your help, I will try that out right away :) On Wednesday, April 1, 2015 at 7:04:51 PM UTC+2, Michael Bayer wrote: On 4/1/15 10:28 AM, Pierre B wrote: Here's a simple visual of the schema OK, so that's called a polymorphic foreign

Re: [sqlalchemy] PostgreSQL UNION with ORDER or LIMIT parenthesis patch

2015-04-02 Thread Mike Bayer
haha it's not resolved :) UNIONs are really tough, as are SQLite's choice of quirks in this area. On 4/1/15 11:27 PM, Charles Leifer wrote: Funnily enough I ran into this particular question regarding Peewee ORM and, googling Postgresql UNION parentheses, I found this post. I ran into the

Re: [sqlalchemy] PostgreSQL UNION with ORDER or LIMIT parenthesis patch

2015-04-02 Thread Jonathan Vanasco
I wish I saw this earlier. This thread basically describes a situation I've had, and explains a few of the weird hacks I've had to use. If you need a workaround, I've been getting around this by using a union on 2 subqueries, querying the union columns, and using plaintext sql to order (the

Re: [sqlalchemy] Nested transaction rollback does not undo changes to instances

2015-04-02 Thread Mike Bayer
On 4/2/15 10:01 AM, Chris Wilson wrote: Dear SQLAlchemy developers, I think I've found a problem with SQLAlchemy not rolling back changes to instances that are committed in an inner nested transaction, when the outer nested transaction rolls back. The manual says: When

[sqlalchemy] Nested transaction rollback does not undo changes to instances

2015-04-02 Thread Chris Wilson
Dear SQLAlchemy developers, I think I've found a problem with SQLAlchemy not rolling back changes to instances that are committed in an inner nested transaction, when the outer nested transaction rolls back. The manual says: When begin_nested()

Re: [sqlalchemy] Nested transaction rollback does not undo changes to instances

2015-04-02 Thread Mike Bayer
On 4/2/15 11:54 AM, Mike Bayer wrote: it's a bug, and I've captured the origin, diagnosis and probable solution here: https://bitbucket.org/zzzeek/sqlalchemy/issue/3352/nested-begin_nested-blocks-dont-track this issue is fixed for 0.9.10 and 1.0.0b5, you can test now using either latest