Justin Van Patten

/run-acceptance-tests Please view the results of the acceptance tests [Here](https://github.com/pulumi/pulumi/actions/runs/7174633968)

Thanks for putting this together, @raelix, and thanks for your patience! I've kicked off a test run to see how it fares and we'll look more closely at the changes.

@raelix, in the issue, you mentioned: > It means that on a stack with ~2k resources it completes in 6m instead of 30m. Can you share any more details on...

@tobiashenkel, what version of the Pulumi Kubernetes plugin are you using (run `pulumi about` inside your stack directory)?

> The rule of "stack outputs are determined entirely by the last successful run" seems easier to reason about. After reading all the comments, I think this is my preference.

The fix for this has merged and will be part of the next CLI release this week. The fix implements the following behavior as noted [here](https://github.com/pulumi/pulumi/issues/14621#issuecomment-1828791907): If the stack did...

Private repos and local file paths may be considered sensitive.

This test is now consistently failing on macOS: https://github.com/pulumi/pulumi/actions/runs/7908712209/job/21588589686#step:38:214 ``` === FAIL: integration TestConstructComponentConfigureProviderPython (174.27s) io.go:432: + set -euo pipefail io.go:432: + DONEFILE=.done io.go:432: ++ git rev-parse HEAD:./ io.go:432: +...

I've temporarily skipped this test on macOS to unblock the release. This issue tracks addressing the underlying problem and unskipping.

Workaround would be to specify: ``` pulumi up --suppress-outputs ```