Hi John, On Thu, 15 Aug 2024 at 17:33, John Engebretson <jengebr...@gmail.com> wrote: > Thanks Ralph! Unfortunately the test you attached doesn't attempt any > log statements against thread thread context. The only test performed on > this implementation loops through the keys that are present, and provides > identical results for CopyOnWriteSortedArrayThreadContextMap. Note that I > had to strip the file down to work against 2.24.0, which doesn't contain > some of the other implementations you tested. > So, I believe "null" is the correct result from the test, and I've > confirmed that result is provided by both an existing implementation and > the new string array implementation. The key-scanning test also passes.
The `null` result might have been caused by the "ScopedContext PR"[1] that also changed how `ContextDataInjector` was used, but recently I postponed the PR to 2.25.0. I am no longer observing the `null` result in PR#2690[2] and I have similar results as you, which is reassuring. I plan to merge that PR in about a week and release 2.24.0 soon after. Piotr [1] https://github.com/apache/logging-log4j2/pull/2438 [2] https://github.com/apache/logging-log4j2/pull/2690
