///

* Si la PR peut être simplifiée d'une certaine manière, vous pouvez le demander, mais il n'est pas nécessaire d'être trop pointilleux, il peut y avoir beaucoup de points de vue subjectifs (et j'aurai les miens aussi 🙈), donc il est préférable de vous concentrer sur les choses fondamentales.

### Tests { #tests }

* Aidez‑moi à vérifier que la PR a des **tests**.

        print(f"Hi {name}!")
```

`typing` propose également un raccourci pour déclarer que quelque chose peut être `None`, avec `Optional`.

Voici un conseil issu de mon point de vue très subjectif :

- 🚨 Évitez d'utiliser `Optional[SomeType]`
- À la place ✨ **utilisez `Union[SomeType, None]`** ✨.

import com.google.common.annotations.GwtCompatible;
import org.jspecify.annotations.NullMarked;
import org.jspecify.annotations.Nullable;

/**
 * To be implemented by test generators that can produce test subjects without requiring any
 * parameters.
 *
 * @param <T> the type created by this generator.
 * @author George van den Driessche
 */
@GwtCompatible
@NullMarked

def say_hi(name: Union[str, None]):
        print(f"Hi {name}!")
```

`typing` also has a shortcut to declare that something could be `None`, with `Optional`.

Here's a tip from my very **subjective** point of view:

* 🚨 Avoid using `Optional[SomeType]`
* Instead ✨ **use `Union[SomeType, None]`** ✨.

import com.google.common.annotations.GwtCompatible;
import org.jspecify.annotations.NullMarked;
import org.jspecify.annotations.Nullable;

/**
 * To be implemented by test generators that can produce test subjects without requiring any
 * parameters.
 *
 * @param <T> the type created by this generator.
 * @author George van den Driessche
 */
@GwtCompatible
@NullMarked

    // these checks test alternative subjects. The test data contains an
    // alternative subject starting with a japanese kanji character. This is
    // not supported by Android because the underlying implementation from
    // harmony follows the definition from rfc 1034 page 10 for alternative
    // subject names. This causes the code to drop all alternative subjects.
    assertThat(verifier.verify("bar.com", session)).isTrue()

    void equalsAndHashCode_subjectSemantics() {
        Kerb5Authenticator a = new Kerb5Authenticator((Subject) null);
        Kerb5Authenticator b = new Kerb5Authenticator((Subject) null);

        // Both null subjects -> equal
        assertEquals(a, b);
        assertEquals(a.hashCode(), b.hashCode());

        // Mixed null/non-null -> not equal
        Kerb5Authenticator c = new Kerb5Authenticator(new Subject());

        assertEquals(orig.getUserDomain(), copy.getUserDomain());
        assertEquals(orig.getPassword(), copy.getPassword());

        // Test the cloning behavior with both null and non-null cached subjects
        if (first != null) {
            // If we got a valid subject, test that it's properly cached and cloned
            Subject cachedOrig = orig.getSubject();
            assertNotNull(cachedOrig);

///

* If the PR can be simplified in a way, you can ask for that, but there's no need to be too picky, there might be a lot of subjective points of view (and I will have my own as well 🙈), so it's better if you can focus on the fundamental things.

### Tests { #tests }

* Help me check that the PR has **tests**.

* Fixes an issue with RBAC reconciliation that could cause duplicated subjects in some bootstrapped rolebindings on each restart of the API server. ([#53239](https://github.com/kubernetes/kubernetes/pull/53239), [@enj](https://github.com/enj))