*** brinzhang0 is now known as brinzhang | 01:36 | |
*** brinzhang0 is now known as brinzhang | 02:51 | |
*** brinzhang_ is now known as brinzhang | 09:12 | |
opendevreview | Stephen Finucane proposed openstack/oslo.db master: types: Set 'cache_ok' (redux) https://review.opendev.org/c/openstack/oslo.db/+/856453 | 09:48 |
---|---|---|
stephenfin | zzzeek: When you're about, could you give me a hint about how I can test that? ^ | 09:50 |
stephenfin | What I've tried clearly isn't useful | 09:51 |
stephenfin | >>> types.JsonEncodedDict()._static_cache_key | 09:51 |
stephenfin | symbol('no_cache') | 09:51 |
stephenfin | I tried binding also, but I don't think that's doing what I think it's doing 😇 | 09:55 |
stephenfin | >>> type_ = types.JsonEncodedDict() | 09:56 |
stephenfin | >>> bind_processor = type_.bind_processor(dialect=default.DefaultDialect()) | 09:56 |
stephenfin | >>> bind_processor({'a': 42, 'b': [1, 2, 3]}) | 09:56 |
stephenfin | '{"a": 42, "b": [1, 2, 3]}' | 09:56 |
dtantsur | huh, autocommit does not help ironic-inspector, broken now | 11:34 |
dtantsur | migrations tests are a horrible mess | 11:35 |
dtantsur | short of reverting the oslo.db bump, do we have any ideas? | 11:42 |
dtantsur | https://review.opendev.org/c/openstack/ironic-inspector/+/856488 should show-case the problem once it's through the CI | 11:43 |
stephenfin | dtantsur: I'll look before EOD. Looks like there's another (alembic-based?) context that's messed up | 13:35 |
stephenfin | (IMO) | 13:36 |
stephenfin | dtantsur: https://github.com/openstack/ironic-inspector/blob/master/ironic_inspector/test/unit/test_migrations.py#L60 | 13:39 |
stephenfin | you need to pass 'autocommit=True' to that | 13:39 |
stephenfin | dtantsur: https://github.com/openstack/oslo.db/blob/master/oslo_db/sqlalchemy/orm.py#L60 | 13:39 |
dtantsur | stephenfin: great find, thanks a lot! | 15:15 |
*** gibi is now known as gibi_off | 16:16 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!