On Fri, 13 Nov 2026 10:08:58 GMT, Andrew Haley <[email protected]> wrote:
>>> You're saying that people who don't have an AArch64 machine that supports >>> pointer authentication won't be able to detect pointer authentication >>> failures? But at least they'll know there's a problem. How is knowing bad? 🤔 >> >> Oh, certainly, knowing that a problem exists is great, but if there isn't >> any way to reproduce it locally, I worry that the CI failure could get >> downplayed and ignored. > >> Oh, certainly, knowing that a problem exists is great, but if there isn't >> any way to reproduce it locally, I worry that the CI failure could get >> downplayed and ignored. > > I see. Well, yes, but that's a different domain. Let's leave the > psychoanalysis of our fellow maintainers out of this review. > @theRealAph I went ahead and updated the CI build rules to pass > "--enable-branch-protection" during the build and > "-XX:UseBranchProtection=standard" during the tests for both Linux/AArch64 > and Windows/ARM64. Why on both? If we enable branch protection on only one of these, then we're testing both possibilities with every commit, which is exactly what we need. ------------- PR Comment: https://git.openjdk.org/jdk/pull/33595#issuecomment-6875946322
