Matt Jankowski
Matt Jankowski
> What's the gain here? It makes the spec even harder to follow: first you need to check the `before` callback and then you need to read a private method...
Rebased, updated the method name a bit to make it clear WHY we need to only return a subset of the checks, and added an assertion about the html page...
Added a commit here which adds explicit ordering and (I think) handles this intermittent order-based failure: - https://github.com/mastodon/mastodon/actions/runs/7730831147/job/21077168755 - https://github.com/mjankowski/mastodon/actions/runs/7740058517/job/21104277140
Pulled out a bunch of smaller PRs (linked) on the assumption that this may be too large to review all at once.
Rebased this after snowflake one merged. This is now just: - The range change (also in it's own linked PR) - One snowflake change I missed in the other PR...
All the smaller linked PRs are merged, and this one is now just: - One snowflake change I missed in that PR - The spec coverage for all the classes...
Closing in favour of linked PRs which split this up by measure/dimension groups.
Rebased this against main to get the model concern module changes, and moved the file. Also removed all of the non-module-extraction style changes that were previously here in favor of...
Rebased again, I suspect rubocop stuff is resolved now.
Rebased to include the composite key PR merge.