Pulse #19 highlights: Scenario 7.0 released, documentation and continuous integration (CI) improvements, preparing for future releases (ops 2.17 coming soon!), and continued work on the proof of concept to remove the AllWatcher from python-libjuju.
A big thank you to all the charm authors who responded to the charm authors survey!
Our office hours experiment for this cycle has come to an end - thanks to everyone that came along. We’ll be reviewing feedback and soliciting ideas about whether and how to continue with this in the future.
The ops library
- In partnership with Pietro, we released Scenario 7.0. This has a few new features, but the main focus is on a set of API adjustments that we’ve been working on over the last few months. There are quite a few changes (and a guide to help you upgrade) but the intention is that the project will now shift to a change cycle more like ops, where we maximise backwards compatibility as long as possible, and you’re encouraged to upgrade frequently. We’ve also been working on installing Scenario as an optional dependency of ops, where you access the framework inside ops.testing, and can find API reference documentation alongside the rest of ops.
- CI improvement: ops will now automatically run some tests against all published charms each month before our typical release window.
- Coming in ops 2.17, a
DeprecationWarningwill be emitted when observing the events that will no longer be supported in Juju 4 (collect-metrics,pre-series-upgrade, andpost-series-upgrade). - Coming in ops 2.17, various improvements include expanding the secret ID out to the full URI when only given the ID, adding a JujuVersion property for Pebble log forwarding to Loki, and documenting and validating settable status values in _ModelBackend.set_status – changes will be covered in more detail next pulse report.
Charming
- Tony continued investigating the merits of a first-class/native Juju defer.
- A couple of team members will be contributing to python-libjuju on a regular basis for the rest of the cycle.
- The scope of python-libjuju 4 has been decided on.
- There was a great turnout for the recent charm authors survey, and we’ve begun collating and analzying the responses to inform our work leading up to and in the next cycle. Many thanks to everyone who took the time to fill in the survey!
- Continuing improvements to super-tox, our tool for testing changes to
opsagainst 150+ charms.
Pebble
- Initial work on SSDLC threat modelling.
- Integration tests for Pebble - almost complete.
Documentation
- The Charm SDK tutorial now features fully type annotated code, and various how-to guides have seen updates .
- Dima added a new top-level intro to the ops API reference.