This week I:
- cleaned up logging output for ops https://github.com/canonical/operator/pulls?q=is%3Apr
- implemented TCP routing for traefik, itests done, utests still not passing. Will finalize next week.
- Processed some feedback on my open prs in
- Lots of documentation work, mainly on charm taxonomy (Charm types - doc - Charmhub) and events
- Got started on a feature for
jhack tailto show the deferral status of events by parsing the debug-log.
- Lots of log forwarding work. Met with Ben to discuss various aspects. Draft PR is up - still quite a bit of work to do writing tests, handling error path edge cases, TLS details, etc.
- Looked into various issues - e.g. #801 #803, etc.
- Reviewed a few PRs including: relation access handling, logging cleanups, etc.
- Reviewed/looked at some of Pietro’s status work.
- Interview stuff.
Last week, I …
- Ran into some blockers on the dashboard.
- Switched gears and put together a basic smoke test for ops, prompted by a xenial bug
- Commented and edited PRs, docs, etc.
- Prompted by a comment in Mattermost, investigated and filed It’s easy to access stable relation data in the testing harness.
- Looped back to the dashboard with fresh eyes. Realized that I had used
npm linkrather than
yarn linkto link a local checkout of js-libjuju, which explained why it still appeared to be running on dated code.
- Had a conversation with jam about the remaining issues in the dashboard. It appears that the Juju team moved some facades around in 3.0. This might be why the dashboard can’t find them, despite having been refactored to avoid reliance on facades that were due for deprecation.