Eclipse editor falsely flags error when using 'proceed()' in lambda expression

[kriegaex]

I am not sure if anyone still looks at AJDT Bugzilla, so I am recording this problem here, creating tracker bug #576017 in AJDT Bugzilla only.

While answering StackOverflow #69196740, I noticed the following problem in AJDT.

This aspect compiles and runs fine in Eclipse 2021-06 and AspectJ 1.9.7:

package de.scrum_master.aspect;

import java.util.concurrent.Callable;

public aspect MemoizeAspect {
  Object around(): execution(* doesNotExist()) {
    MyCache<String, Object> cache = new MyCache<>();
    return cache.get("key", () -> proceed());
  }

  Object around(): execution(* doesNotExist()) {
    MyCache<String, Object> cache = new MyCache<>();
    Callable<?> callable = () -> proceed();
    return cache.get("key", callable);
  }

  Object around(): execution(* doesNotExist()) {
    MyCache<String, Object> cache = new MyCache<>();
    return cache.get("key", () -> "value");
  }

  public static class MyCache<K, V> {
    public V get(K key, Callable<? extends V> loader) {
      return null;
    }
  }
}

Eclipse does not show any errors in the "Problems" view. But the editor falsely flags an alleged error:

The error message is:

The method get(String, Callable<? extends Object>) in the type MemoizeAspect.MyCache<String,Object> is not applicable for the arguments (String, () -> {})

Please note:

• As you can see in the third around-advice, when returning a fixed value instead of proceed() or proceed(arg1, args2) there is no problem.
• When transforming the aspect into annotation-style, it also works. But in that case, we also need to wrap proceed() into a try-catch block handling Throwable, so the source code looks different.
• Factoring out the Callable into a local variable like in the second around-advice works, too.