Merge datastax python-driver 3.24.0 #74
Conversation
PYTHON-1228 Provide log wrapper to asyncore so it doesnt error
…rsing Python 1207 transient parsing
- Add TestCluster class to integration tests to allow easier setting of default options and updated all call points - Removed setting __defaults__ on stuff - Changed a bunch of '4.0' checks to be '4.0-a' - Enabled materialized views, sasi, and transient replication in standard test startup options - Updated materialized views to specify not null for primary keys as required in 4.0 - Added new protocol error message received from 4.0 in test_no_connection_refused_on_timeout
Cassandra 4.0 test changes
More test fixes
CASSANDRA-15663 Add 'default' to a list of reserved keywords
PYTHON-700: Support port discovery for C* 4.0
…emoval Add core graph documentation
…-fix PYTHON-1207: transient replication support (phase 2)
PYTHON-1225: add initial Jenkinsfile
PYTHON-1220: Drop python 3.4 support
PYTHON-1260: Ensure the driver can connect when invalid peer hosts are in system.peers
PYTHON-1266: Fix asyncore race condition cause logging exception on shutdown
PYTHON-1258: Implement protocol v5 checksumming
We set out to build a Flask app that talks to Astra. The CQL plugin we used turned out to be crashing if we didn't give it hosts to connect to (https://github.com/thegeorgeous/flask-cqlalchemy/blob/master/flask_cqlalchemy/__init__.py#L48). But we also had to use the `cloud` key to pass in Astra credentials (via `setup_kwargs` nearby the link above). This exploded on us in cqlengine.connection#setup() where the hosts passed into Cluster() made that constructor blow up because: `ValueError: contact_points, endpoint_factory, ssl_context, and ssl_options cannot be specified with a cloud configuration` We're not sure this is the right place to patch, but it helped us at least. We figured coming out with a patch directly, instead of an issue, might save some time.
PYTHON-1265: Fix the default cqlengine connection mechanism to work with Astra
| Doing it with PR like this might be problematic, we'll need to solve a few of the conflicts I've written down the procedure to do that: https://github.com/scylladb/python-driver/blob/master/scripts/merge_next_tag_from_upstream.sh But please don't push tag a new version at least not yet. |
| I tried and think to have solved them but I may very well have missed something What I can say is that the shard-aware driver works good with this But we can use your script too ofc |
We can consider doing that on a PR, it would help with testing it before tagging a new version |
| @ultrabug seem like all the tests (both unittest and integration) were broken here... |
| version released: https://pypi.org/project/scylla-driver/3.24.0/ |
It sounds like a good time to sync with upstream