On Mon, 2 Nov 2026 22:01:39 GMT, Artur Barashev <[email protected]> wrote:
>> Timofei Fedotov has updated the pull request incrementally with two >> additional commits since the last revision: >> >> - Verify complete DTLS handshake fragment reassembly >> - Test inconsistent DTLS handshake fragment types > > src/java.base/share/classes/sun/security/ssl/DTLSInputRecord.java line 586: > >> 584: } >> 585: >> 586: private record HandshakeHoles( > > I suggest to rename this class to `HandshakeReassemblyState` Done. > src/java.base/share/classes/sun/security/ssl/DTLSInputRecord.java line 751: > >> 749: (hsf.fragmentLength != hsf.messageLength); >> 750: >> 751: HandshakeHoles message = > > I suggest to rename `message` to `reassemblyState` here and in other places, > `message` looks rather confusing. Done. > test/jdk/javax/net/ssl/DTLS/InconsistentFragmentLength.java line 29: > >> 27: * @summary Reject inconsistent metadata in DTLS handshake fragments >> 28: * @library /test/lib >> 29: * @run main InconsistentFragmentLength > > Nit: this line is not really needed, `main()` is called by default. Done. > test/jdk/javax/net/ssl/DTLS/InconsistentFragmentLength.java line 129: > >> 127: } >> 128: Asserts.assertEquals(engine.getHandshakeStatus(), >> 129: SSLEngineResult.HandshakeStatus.NEED_WRAP); > > Here is the sequence of events: > > 1. `unwrap()` returns NEED_TASK. > 2. The application executes `DelegatedTask.run()`, which catches and stores > an exception. > 3. A later `wrap()` or `unwrap()` calls `checkTaskThrown()` and rethrows that > exception. > > Thus, after running the tasks, call `wrap()` and assert successful > ClientHello output. This also exposes any saved exception. Done. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/33633#discussion_r4451198269 PR Review Comment: https://git.openjdk.org/jdk/pull/33633#discussion_r4451197930 PR Review Comment: https://git.openjdk.org/jdk/pull/33633#discussion_r4451198600 PR Review Comment: https://git.openjdk.org/jdk/pull/33633#discussion_r4451198893
