Closed VenelinMartinov closed 3 weeks ago
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 61.48%. Comparing base (
b5efa0d
) to head (f2a07a8
). Report is 3 commits behind head on master.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
Both of these assert on the current behaviour for both plan resource change and non-plan resource change.
I built it out to compare behaviour in https://github.com/pulumi/pulumi-terraform-bridge/pull/1971 and make sure we aren't regressing.
I'd like to merge as is and leave refactoring for later as https://github.com/pulumi/pulumi-terraform-bridge/pull/1971 depends on this and it has been hanging for quite a while
The structure of the tests it that for each of the property it injects an unknown at every possible level and asserts on the response.
Comparing PlanResourceChange and non-PlanResourceChange is not the point here
GH token rate limits ❤️
@iwahbe we need some plumbing work to do to allow cross-tests to express unknowns in the meanwhile we test at this level. I requested a few additional comments/cross-links but it looks good. It was very important to me to have some edge cases covered here with regression tests and some explanation as to why we think it behaves the way it does.
I've done some further digging around the results here in https://github.com/pulumi/pulumi-terraform-bridge/issues/2032 and https://github.com/pulumi/pulumi-terraform-bridge/pull/2060/files and https://github.com/pulumi/pulumi-terraform-bridge/pull/2060/files
I suspect we can do better with unknowns
This adds GRPC around unknowns for both attributes and collections for both PlanResourceChange and non-PlanResourceChange