| Commit message (Collapse) | Author | Age | Lines |
| ... | |
| |\ \ \ \ \
| |/ / / /
|/| | | | |
Correctly initialise redis tests
|
| |/ / / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Calling the cog_load from within the setUp function resulted in interaction with a RedisSession before it was initialised.
This wasn't noticed in CI as it only error under certain concurrency timings due to xdist.
To resolve this, we moved the setup and async setup logic to a base class.
Co-authored-by: Hassan Abouelela <[email protected]>
|
| |\ \ \ \
| |_|/ /
|/| | | |
Globally decode binary responses from redis to strings
|
| |/ / / |
|
| |\ \ \ |
|
| | | | |
| | | |
| | | |
| | | |
| | | | |
The new versions introduce conversions which causes the doc command
embed to be formatted improperly
|
| | | | | |
|
| | | | |
| | | |
| | | |
| | | | |
This helper ensures that a fresh RedisSession is given to each test case that inherits from it.
|
| | | | | |
|
| | | | |
| | | |
| | | |
| | | | |
pep-naming now supports these functions being in camel case.
|
| | | | | |
|
| | | | |
| | | |
| | | |
| | | | |
This was a new lint rule added in the latest bugbear.
|
| | | | | |
|
| | | | | |
|
| | | | | |
|
| | | | |
| | | |
| | | |
| | | | |
If a float is given, Redis will assume the expiry is in milliseconds and must be multiplied by 1000. This is undesirable, as we are already passing the expiry in seconds.
|
| | | | | |
|
| | | | |
| | | |
| | | |
| | | | |
This has been abstracted away, the correct way to do this now is to directly access the client.
|
| | | | |
| | | |
| | | |
| | | | |
This commit resolves all the breaking changes from the aioredis -> redis-py migration.
|
| | | | | |
|
| | | | |
| | | |
| | | |
| | | | |
This bump comes with a move to redis-py over aioredis. As such, pin new transitive dependancies to exact versions.
|
| | | | |
| | | |
| | | |
| | | | |
These warnings were caused by the setup coro from error_handler.py being imported directly, causing a warning about an un-awaited coro whenever the Cog was accessed from the same file.
|
| | | | |
| | | |
| | | |
| | | | |
get_event_loop is deprecated as of 3.10 if there is no running loop.
|
| | | | | |
|
| |/ / /
| | |
| | |
| | | |
The bot service was still configured to depend on the snekbox service, even though this service is now optional, in favour of the snekbox-311 service.
|
| |\ \ \
| | | |
| | | | |
Updated `purge` to require at least one user
|
| | | | | |
|
| | | | | |
|
| |/ / / |
|
| | | | |
|
| | | |
| | |
| | |
| | |
| | |
| | | |
- Added new infraction delta calculations to updated infractions.
- Updates of infraction durations now also update the `last_applied` field.
- `inserted_at` is now sent by the bot client to denote the original unmodified infraction application time
|
| | | |
| | |
| | |
| | | |
Corrected test case to use `datetime.utcnow()` to be consistent with target
|
| | | | |
|
| | | |
| | |
| | |
| | | |
- Used new method of dict subset comparison instead of datetime patching for better compat. with argument types
|
| | | |
| | |
| | |
| | | |
Given API updates, the fallback is not needed
|
| | | |
| | |
| | |
| | | |
- Used arrow.utcnow to reduce complexity and import
|
| | | | |
|
| | | | |
|
| | | |
| | |
| | |
| | | |
- Added new usage of `last_applied` time for duration calculation in `apply_infraction`
|
| | | | |
|
| | | | |
|
| | | |
| | |
| | |
| | | |
- Corrected datetime patching
|
| | | |
| | |
| | |
| | | |
- Updated docstring to be more explicit on parameter fields
|
| | | |
| | |
| | |
| | | |
- Changed `duration` parameter names to `duration_or_expiry` to more accurately reflect options for help
|
| | | | |
|
| | | | |
|
| | | |
| | |
| | |
| | | |
- Refactored tests for new time duration arguments
|
| | | | |
|
| | | | |
|
| | |/
|/| |
|