Closed VenelinMartinov closed 3 months ago
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 62.04%. Comparing base (
5d60418
) to head (de66cb6
).:exclamation: Current head de66cb6 differs from pull request most recent head 012cd5b
Please upload reports for the commit 012cd5b to get more accurate results.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
Note CI used 3.119 not 3.120
Short example of difference in behaviour with new pulumi CLI.
Two resource:
prov_test
- this one has a fully computed property -out_val
prov_aux
- this one will use the computed property as an input.Note the read behaviour on
prov_test
- it produces a different value forout_val
than Create.Test one:
This just uses prov_test -
out_val
is only an output in this program. This is meant to represent how most of our provider tests look like.Test two:
this uses
out_val
as an input to the second resource. This might be a user program.Under old refresh both tests failed at the refresh step - the
outVal
output is changed. Under new refresh only test 2 fails an it only fails at preview, refresh --expect-no-changes is fine.Most of our tests in providers don't exercise properties as an input to another resource, so we might miss regressions in any refresh behaviour.