On Wed, 24 Feb 2021 02:39:57 GMT, Jamil Nimeh <jni...@openjdk.org> wrote:
>> All of the "isAssignableFrom" checks in "engineGetKeySpec" appear to be >> backwards in Java's KeyFactorySpi.engineGetKeySpec implementations. In most >> cases, the requested KeySpec is equal to the concrete implementation so the >> inversion does not matter. But there are few cases, as presented in the >> added jtreg test, will cause unexpected behavior (e.g., ClassCastException >> rather than an InvalidKeySpecException). The fix is trivial. >> >> Co-author @SalusaSecondus > > The fix itself makes sense and looks good to me. However I think it will > cause two other tests to break. Please try running the following two tests > with your changes and see if they fail for you as they did for me: > open/test/jdk/sun/security/rsa/TestKeyFactory.java > open/test/jdk/sun/security/rsa/pss/TestPSSKeySupport.java > > Also just a minor nit: some of the modified files should have their copyright > dates updated to 2021. Hi Jamil, Thanks for reviewing this PR. > The fix itself makes sense and looks good to me. However I think it will > cause two other tests to break. Please try running the following two tests > with your changes and see if they fail for you as they did for me: > open/test/jdk/sun/security/rsa/TestKeyFactory.java > open/test/jdk/sun/security/rsa/pss/TestPSSKeySupport.java Nice catch. I updated both tests in rev-2. > Also just a minor nit: some of the modified files should have their copyright > dates updated to 2021. Done ------------- PR: https://git.openjdk.java.net/jdk/pull/2682
