Prevent ClassLatch from latching wrappers with defaults methods

[bric3]

What Does This Do

Prevents an AbstractMethodError thrown by a wrapper's delegate from latching the wrapper class when interface "evolution" exposes an abstract declaration alongside an inherited default implementation.

Motivation

On Java 8, these delegate failures have no message. The fallback added in #12799 inspected only abstract public methods, which could treat the wrapper as missing the method and disable calls for every healthy instance of that class.

Granted, this case seems a tad exotic, but its fix looks easy enough to handle that case.

Suppose the following, in version A, compiled together.

interface Api {}

interface Forwarding {
  Supplier<String> delegate();

  default String read() {
    return delegate().get();
  }
}

final class Wrapper implements Api, Forwarding {
  private final Supplier<String> delegate;

  Wrapper(Supplier<String> delegate) {
    this.delegate = delegate;
  }

  @Override
  public Supplier<String> delegate() {
    return delegate;
  }
}

Then there's version B of Api with a new method that has the same signature as Forwarding's default method, while the existing Wrapper.class remains as is:

interface Api {
  String read();
}

When invoking wrapper.read() the JVM would still call Forwarding.read(). However, Wrapper.class.getMethods() now exposes both:

• public abstract String Api.read()
• public default String Forwarding.read()

With the following latch

ClassLatch<Api, String, RuntimeException> latch =
    new ClassLatch<Api, String, RuntimeException>() {
      @Override
      protected String apply(Api target) {
        return handleAbstractMethod(target, "read", Api::read);
      }
    };

Api healthy = new Wrapper(() -> "ok");
Api failing = new Wrapper(() -> {
  throw new AbstractMethodError(); // No message, as on Java 8
});

latch.tryApply(healthy); // "ok"
latch.tryApply(failing); // null; incorrectly latches Wrapper.class
latch.tryApply(healthy); // null; expected "ok"

Upon error the cache sees the abstract Api.read() method and concludes that Wrapper lacks a read() method. It overlooked the working Forwarding.read(), so this permanently disables the operation for every Wrapper.
The explicit throw keeps this example small; the reproduction I ran obtained the message-less error from the JVM. This requires keeping the old wrapper binary: recompiling it against the upgraded Api would reject the conflicting declarations.

Additional Notes

The cache remains a String[]. The fix filters the array in place instead of building a HashSet, so it preserves the existing allocation shape. It adds a second bounded scan only when the cache is built after the first message-less failure for a class. A temporary Java 8 JMH comparison found the array scan faster than the HashSet variant on method-heavy types. Healthy and cached paths are unchanged.

Contributor Checklist

• Format the title according to the contribution guidelines
• Assign the type: and (comp: or inst:) labels in addition to any other useful labels
• Avoid using close, fix, or any linking keywords when referencing an issue
  Use solves instead, and assign the PR milestone to the issue
• Update the CODEOWNERS file on source file addition, migration, or deletion
• Update public documentation with any new configuration flags or behaviors
• Once approved, use merge queue to merge the PR

[dd-octo-sts]

🟡 Java Benchmark SLOs — Performance SLO warning (near threshold)

Suite	Status
Startup	🟡 warning

SLO thresholds are defined here based on automatically generated metrics. A warning is raised when results are within 5% of the threshold.

PR vs. master results

Scenario	Candidate	master	Δ (95% CI of mean)
startup:insecure-bank:iast:Agent	13.89 s	13.87 s	[-0.6%; +0.8%] (no difference)
startup:insecure-bank:tracing:Agent	12.87 s	12.89 s	[-0.8%; +0.5%] (no difference)
startup:petclinic:appsec:Agent	17.75 s	17.56 s	[+0.1%; +2.1%] (maybe worse)
startup:petclinic:iast:Agent	16.78 s	16.83 s	[-8.2%; +7.6%] (unstable)
startup:petclinic:profiling:Agent	17.30 s	17.39 s	[-1.8%; +0.8%] (no difference)
startup:petclinic:sca:Agent	17.71 s	17.67 s	[-0.7%; +1.1%] (no difference)
startup:petclinic:tracing:Agent	16.69 s	16.68 s	[-1.1%; +1.2%] (no difference)

Commit: 70380246 · CI Pipeline · Benchmarking Platform UI

---

Load and DaCapo benchmarks can be triggered manually in the GitLab pipeline. Results will appear in the Benchmarking Platform UI after completion.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-12T10:16:01.878777Z	15910e3	Draft marked ready
🔒 Security Review	✅ Completed 2026-10-12T10:21:29.412746Z	15910e3	Draft marked ready

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[datadog-prod-us1-4]

More details

The message-less AbstractMethodError fallback preserves wrappers with an inherited concrete default method, preventing delegate failures from disabling healthy instances of the same class. The completed static review found no reportable defect.

Was this helpful? React 👍 or 👎

_{Open Bits AI session}

_{🤖 Bits Code Review · Commit 15910e3}

[dougqh]

Thanks for fixing my oversight