On Mon, 10 Aug 2026 20:42:23 GMT, Chen Liang <[email protected]> wrote:
>> We should allow unset fields when `uninitializedThis` is in the stack, in >> addition to only checking locals, as a result of >> https://bugs.openjdk.org/browse/JDK-8375481. >> >> Noticed that hotspot's verifier is not up-to-date with this spec; this patch >> therefore only fixes classfile API and javap and leaves classfile API's >> verifier and hotspot in place. >> >> --------- >> - [x] I confirm that I make this contribution in accordance with the >> [OpenJDK Interim AI Policy](https://openjdk.org/legal/ai). > > Chen Liang has updated the pull request with a new target base due to a merge > or a rebase. The incremental webrev excludes the unrelated changes brought in > by the merge/rebase. The pull request contains three additional commits since > the last revision: > > - Add 8389843 > - Merge branch 'master' of https://github.com/openjdk/jdk into > fix/cf-unset-restriction > - unset field allowed for uninitThis on stack Mixed-purpose PRs are difficult to review and maintain. I would strongly recommend splitting this into separate PRs. > This patch currently has these contents: > > 1. Fix `StackMapFrameInfo` and stack maps parsing/generation to correctly > accept a `flagThisUninit` frame with `uninitializedThis` in the stack only. The Class-File API should perform only critical structural validation while parsing a class file. I recommend removing: uninitializedThisCheck: if (!unsetFields.isEmpty()) { for (var local : locals) { if (local == SimpleVerificationTypeInfo.UNINITIALIZED_THIS) { break uninitializedThisCheck; } } throw new IllegalArgumentException("unset fields requires uninitializedThis in locals"); } from the `StackMapDecoder.StackMapFrameImpl` constructor, as this condition is not a structural class file validation. > 2. Incorporate the ClassFile API port of verifier changes in [8389840: > Verifier is not setting flagThisUninit when uninitializedThis is on the stack > #32242](https://github.com/openjdk/jdk/pull/32242) This is an important synchronization with the native verifier. It should have its own bug and PR, with all relevant references. > 3. Fix typo in `StackMapGenerator.Type.UNITIALIZED_THIS_TYPE` This could be treated as a small, unrelated implementation change. > 4. Validate the u2 checks for the list arguments to `StackMapFrameInfo.of` > with unset fields overload This could be treated as a small, independent change. > 5. Fix javap to print unset fields when it encounters a `flagThisUninit` > frame with `uninitializedThis` in the stack only. This belongs to a different JDK module and is unrelated to the other changes. > 6. Clean ups to StrictStackMapsTest, use constant for EARLY_LARVAL frame > type, clean up other IDE warnings. These could be treated as minor test and cleanup changes. ------------- PR Comment: https://git.openjdk.org/jdk/pull/32224#issuecomment-5254176779
