rzo1 commented on PR #305: URL: https://github.com/apache/openjpa/pull/305#issuecomment-6849850318
Ran the whole suite (lib, kernel, jdbc, persistence, persistence-jdbc, locking; Derby) 13 times with random method *and* class order (`MethodOrderer$Random` + `surefire.runOrder=random`). 8 seeds were clean; the failures came from 5 classes, fixed in 61c52313f and d1c3c8ca7: - `TestAutoApplyConverted*`: unenhanced entities are only registered when a factory creates its first EntityManager, so the metamodel tests failed when they ran first (seeds 11, 2). - `TestRangeQuery`: `TestSchemaGenerationProperties` left its drop-script tables in `SchemaTool`'s JVM-wide `_droppedTables` for PU `TestConv`, so a later `buildSchema` factory did not recreate `HELPERPC` (seed 11). - `TestInMemoryQueryMatchEscapes`: `TestEntityManagerFactory.emfCreation` left an `AllFieldTypes` row whose id came from the shared sequence table, which other tests reset → duplicate key (seed 505). - `TestInheritedPropertyAccessPersist`, `TestFieldAccessWithNonStandardAccessors`: fixed ids inserted and never deleted; the Derby DB under `target/` survives builds, so every second local run without `clean` failed (not visible in CI). So +1 for random order; I'd still keep a fixed order (or a fixed seed) in the main build and run random order with a logged seed in a separate job, so new order issues do not turn unrelated PRs red. -- 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]
