On Fri, 3 May 2024 06:49:09 GMT, Alan Bateman <al...@openjdk.org> wrote:
> It seems maybe we should instead be trying to avoid these events by > preloading the classes as was suggested as an option in the CR. The problem is that all such cases are unknown, so it is just not realistic. As Alan commented, this exact debugging option will be removed soon. The other cases are extremely rare because they were not identified with the assert in place (in fact, these assert existed for a couple of releases for exactly this purpose). However, such classes can cause real problems (deadlocks and other issues as we see) if they are not ignored by `CFLH`, `ClassPrepare` and `ClassLoad` posting code. ------------- PR Comment: https://git.openjdk.org/jdk/pull/19054#issuecomment-2092596008
