Juergen Hoeller

A quick follow-up: We actually do have test failures in the test suite when the AspectJ patch is applied but not in the core AOP tests, rather in `AopNamespaceHandlerTests` for...

That `if` block in `KindedPointcut` is still commented out as of AspectJ 1.9.21.1, so there must be some other change improving the matching performance there. @edavedian, great to hear that...

A quick update: AspectJ 1.9.22 is out in the meantime, so please rather re-run your benchmarks against that version.

I assume there is nothing more we can do about this from the Spring Framework side. For any further benchmarks, please run them against the latest AspectJ version and let...

This should have significantly improved in 5.2 already, based on #23336 and #23337. So this is effectively a duplicate at this point.

It looks like we'll address both this and #23473 through a new `@TransactionConfig` annotation, analogous to our existing `@CacheConfig`, or similar global settings which are not trying to rely on...

As a rather design-oriented measure, we are going to take type-level `@Qualifier("...")` values into account for transaction manager selection, effectively serving as a `@Transactional("...")` qualifier value for transaction definitions without...

Hi Greg, Thanks for that suggestion and the initial work! This looks sensible indeed for providing even closer Jetty 12 integration. And since Jetty's Servlet adapters live entirely outside of...

In WebFlux, the Servlet request/response does not leak through to the application at all, so migrating from the Servlet-based adapter to the Jetty Core adapter should be entirely transparent to...

@gregw a general non-blocking label should be fine there since this is exclusively for WebFlux as a reactive stack. Any kind of blocking in that stack would be a mistake...