**Enhancements**
- Drop support for Python 2.7, convert codebase to Python 3.6+.
- CI: drop coverage for Python 2.7 and Django <= 2.1, which are no longer maintained upstream.
- CI: add coverage for Python 3.7 and 3.8, as well as Django long term support (LTS) and the latest released versions.
- CI: un-pin dependency on agronholm/apscheduler149, which has since been merged and released upstream.
- Rename Django `test_settings.py` file to prevent collision with actual test scripts.
- Clean up unused dependencies / update dependencies to latest available versions.
- Switch to Black code formatting.
- Align package layout with official [Django recommendations](https://docs.djangoproject.com/en/dev/intro/reusable-apps/#packaging-your-app)
- Move UI-related `DjangoJobExecution.html_status` out of model definition and in to the associated model admin definition.
- Add `help_text` to model fields to document their use.
- Remove unused code fragments.
- Add Python type annotations.
- Implement various Django best practices for QuerySet management and model instance creation / updates.
- Drop `DjangoJob.name` field in favor of aligning with using APScheduler's `id` field. NOTE: please run your Django
migrations again - might take a while depending on the number of `DjangoJobExecutions` in your database.
- Acquire a DB lock when updating `DjangoJob` or `DjangoJobExecution` instances. This should be safer for multi-threaded
usage.
- Switch to using `BigAutoField` for `DjangoJobExecution`'s primary keys. This should prevent running out of usable ID's
for deployments with a very large number of job executions in the database (
Resolves [36](https://github.com/jcass77/django-apscheduler/issues/36)).
- Implement `DjangoJob.shutdown()` method to close database connection when scheduler is shut down.
- `jobstores.register_events` has been deprecated and will be removed in a future release. Calling this method is no
longer necessary as the `DjangoJobStore` will automatically register for events that it cares about when the scheduler
is started.
- Ensure that Django and APScheduler always use the same timezones when passing datetimes between the two.
- Use the configured scheduler's locking mechanism to keep the creation of `DjangoJobExecution` in sync with APScheduler
events.
- Update README on recommended usage, which includes using a `BlockingScheduler` with a custom Django management command
instead of running a `BackgroundScheduler` directly in a Django application.
- Remove `ignore_database_error` decorator. All database errors will be raised so that users can decide on the best
course of action for their specific use case (Resolves [79](https://github.com/jcass77/django-apscheduler/issues/79))
.
- Remove `DjangoJobManager`: users should be allowed to manage the DB connection themselves based on their
implementation-specific use case. See the official Django recommendations at: https://code.djangoproject.com/ticket/21597#comment:29.
- Add AUTHORS file.
- Increase test coverage.
- Remove the `DjangoJobExecution.started` field. It appears that APScheduler only fires an event when the job is
submitted to the scheduler (not when job execution actually starts). We now calculate the job `duration` as the
elapsed time in seconds between the scheduled `run_time` and when we receive the `events.EVENT_EXECUTED`
APScheduler event.
**Fixes**
- Fix PEP8 code formatting violations.
- Implement locking mechanism to prevent duplicate `DjangoJobExecution`s from being created (
Fixes [28](https://github.com/jcass77/django-apscheduler/issues/28)
, [30](https://github.com/jcass77/django-apscheduler/issues/30)
, [44](https://github.com/jcass77/django-apscheduler/issues/44)).
- `DjangoJobStore.add_job` now raises a `ConflictingIdError` if a job with that particular ID already exists in the job
store. This aligns with the behavior expected by the APScheduler interface. Use the `replace_existing` parameter to
update existing jobs instead.