jdaugherty commented on PR #16456: URL: https://github.com/apache/grails-core/pull/16456#issuecomment-5935773615
@matrei split as you suggested: #16462 on `8.0.x` carries the builder, the hierarchy-member flag (package-scoped there), the development-watch guard, their specs, both examples and a short doc section. This PR was rebuilt on current `8.1.x` as a merge of that branch plus one feature commit, so its diff shows both until #16462 is merged up; after that only the feature commit remains and applies on its own. Your findings are addressed in the feature commit: 1. The loan is now withdrawn in `DisposableBean.destroy()` on `GrailsApplicationPostProcessor`, which Spring runs on a failed refresh as well as on a normal close; the close-event path is gone. `GrailsAppBuilderSpec` covers a child with a failing `@Bean` followed by a working child beneath the same parent (it fails with the event-only withdrawal). 2. A hierarchy member without an application class logs at `INFO`, naming its sources, and the docs say the Grails context must have the `Application` class among its sources (#16462 carries both). 3. Destroying dependent parent beans is documented as intended, in the code and in the docs with an `ObjectProvider` example; a spec shows the dependent bean destroyed with the loan and recreated against the next one. 4. Specs added for `properties()` reaching the parent, the child shutdown hook (observed through Boot's close listener on the context), `sibling(Class[], String...)`, a parent's own `pluginManager` being left alone, and both contexts of the CLI example being started by a `GrailsApp`. 5. The docs now say to keep the parent to plain configuration and leave auto-configuration to the Grails child. Nits: `build()` and `build(String...)` return `GrailsApp`, the example controller is `@GrailsCompileStatic`, and the run-status change is in the what's new entry and the description. `contextHierarchyMember` stays public on 8.1 as the hook for a hierarchy assembled by hand, and is package-scoped on 8.0.x as you suggested. Could you review both? -- 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]
