On Tue, 27 Oct 2026 22:07:53 GMT, David Holmes <[email protected]> wrote:
> That said, wouldn't taking the core dump immediately address both native and > Java issues and provide stacks dumps that are consistent with each other? I think it could, but I see a few issues with an approach like that: 1) In at least on one of our platforms, the process is killed as a part of the core dumping mechanism used. 2) If this is the only thread dump mechanism, we would loose the ability to make multiple, consecutive dumps, which could be useful for tracking movements of the threads. (The reason why we have iteration_0, .., iteration_5) 3) Something needs to materialize the Java frames of the thread dumps. Today that's something the SA can do for core dumps. Personally, I don't want to rely on that tool for future features. ------------- PR Comment: https://git.openjdk.org/jdk/pull/33591#issuecomment-6464075374
