kevdoran commented on PR #11210: URL: https://github.com/apache/nifi/pull/11210#issuecomment-4389853785
Thanks @mcgilman. Pushed an update: Here is a summary of what was changed: 1. Test isolation (testVerifyConfigurationStepSkipsSecretReferenceWhenPropertyDependenciesNotMet) - Built the test's SecretReference with all four fields populated (pid, My Provider, my-secret, My Provider.my-secret) so it is not an empty stub. The new isEmptySecretReference filter cannot mask a regression in the dependency filter. - Wired in a strict SecretsManager whose getSecrets(anySet()) is stubbed to thenThrow(new AssertionError(...)). If the dependency-skip path ever stops short-circuiting, the test now fails loudly with a clear message. - To support per-test SecretsManagers, added an overload createConnectorNode(Connector, SecretsManager). The original single-arg overload now delegates to it, preserving existing tests. 2. Javadoc accuracy (getConfiguredValueForDependency) - Replaced the misleading "aligned with AbstractConnector.isDependencySatisfied(...)" claim with an explicit description of: "What this helper is for (controlling-property String value lookup against a pre-resolution StepConfiguration)." - The two cases where it intentionally diverges from the connector-side implementation (typed null StringLiteralValue → descriptor default; non-literal references → null/unsatisfied). - Why those divergences are acceptable (controlling properties are, in practice, STRING_LITERAL values from a fixed allowable-value set; null literals match the "switch back to default mode" UX). -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
