On Thu, 15 Oct 2026 18:08:17 GMT, Matthew Donovan <[email protected]> wrote:
> The JBS issue says this is an issue for Linux/loongarch64 architecture. This
> change causes a test to be skipped on MacOSX aarch64 when it can successfully
> run and pass without legacy.so. Can you update your change to apply the logic
> only for the loongarch64?
Thanks for the review. You're right that the legacy-provider check shouldn't
be applied unconditionally — I've narrowed it so it only runs on Linux.
It's actually not loongarch64-specific: I reproduced the same failure on
Linux/aarch64 and Linux/x64 as well, where the system OpenSSL is also built
without the legacy provider. So gating on Platform.isLinux() covers all of
those, while leaving macOS aarch64 (which uses the downloaded bundle and
passes) unaffected.
-------------
PR Comment: https://git.openjdk.org/jdk/pull/33380#issuecomment-6206661557
