Und genau das ist **HTTPS**, es ist einfach **HTTP** innerhalb einer **sicheren TLS-Verbindung**, statt einer puren (unverschlüsselten) TCP-Verbindung.

/// tip | Tipp

Beachten Sie, dass die Verschlüsselung der Kommunikation auf der **TCP-Ebene** und nicht auf der HTTP-Ebene erfolgt.

///

/// info

Remarquez que `Offer` a une liste d’`Item`, qui à leur tour ont une liste optionnelle d’`Image`.

///

## Corps de listes pures { #bodies-of-pure-lists }

Si la valeur de premier niveau du corps JSON attendu est un `array` JSON (une `list` Python), vous pouvez déclarer le type dans le paramètre de la fonction, de la même manière que dans les modèles Pydantic :

  python3 tensorflow/tools/ci_build/update_version.py --nightly
fi

# This generates a pure python wheel of the format "*-py3-none-any.whl"
bazel run --HERMETIC_PYTHON_VERSION=3.13 //tensorflow/tools/pip_package:setup_py_binary -- bdist_wheel --dist-dir "$TFCI_OUTPUT_DIR"

# Get the name of the pure python wheel that was built. This should
# resolve to either
# 1. tensorflow-a.b.c-py3-none-any.whl  or, for nightly builds,

	{"3>>(1<<63)", "negative right shift count"},
	{"(1<<63)>>2", "right shift of value with high bit set"},
	{"(1<<62)>>2", ""},
	{`'\x80'`, "illegal UTF-8 encoding for character constant"},
	{"(23*4", "missing closing paren"},
	{")23*4", "unexpected ) evaluating expression"},
	{"18446744073709551616", "value out of range"},
}

func TestBadExpr(t *testing.T) {
	for i, test := range badExprTests {
		err := runBadTest(i, test, t)

    if (charset.equals(UTF_8)) {
      int utf16Length = input.length();
      int h1 = seed;
      int i = 0;
      int len = 0;

      // This loop optimizes for pure ASCII.
      while (i + 4 <= utf16Length) {
        char c0 = input.charAt(i);
        char c1 = input.charAt(i + 1);
        char c2 = input.charAt(i + 2);
        char c3 = input.charAt(i + 3);

        // ensure that bouncycastle is on classpath for the all of test types, must happen in evaluateAfter since the rest tests explicitly
        // set the class path to help maintain pure black box testing, and here we are adding to that classpath
        tasks.withType(Test).configureEach { Test test ->
          test.setClasspath(test.getClasspath().plus(extraFipsJars))
        }
      }

/// info | Bilgi

`Offer`’ın bir `Item` list’i olduğuna, `Item`’ların da opsiyonel bir `Image` list’ine sahip olduğuna dikkat edin.

///

## Sadece list olan body’ler { #bodies-of-pure-lists }

Beklediğiniz JSON body’nin en üst seviye değeri bir JSON `array` (Python’da `list`) ise, tipi Pydantic modellerinde olduğu gibi fonksiyonun parametresinde tanımlayabilirsiniz:

```Python
images: list[Image]

import com.google.errorprone.annotations.Immutable;
import java.nio.ByteBuffer;
import java.nio.charset.Charset;
import org.jspecify.annotations.Nullable;

/**
 * A hash function is a collision-averse pure function that maps an arbitrary block of data to a
 * number called a <i>hash code</i>.
 *
 * <h3>Definition</h3>
 *
 * <p>Unpacking this definition:
 *
 * <ul>

/// info | Info

Beachten Sie, wie `Offer` eine Liste von `Item`s hat, die ihrerseits eine optionale Liste von `Image`s haben.

///

## Bodys aus reinen Listen { #bodies-of-pure-lists }

Wenn das äußerste Element des JSON-Bodys, das Sie erwarten, ein JSON-`array` (eine Python-`list`) ist, können Sie den Typ im Funktionsparameter deklarieren, mit der gleichen Syntax wie in Pydantic-Modellen:

```Python

    // Warning to maintainers: this implementation is highly optimized.
    int utf16Length = sequence.length();
    int utf8Length = utf16Length;
    int i = 0;

    // This loop optimizes for pure ASCII.
    while (i < utf16Length && sequence.charAt(i) < 0x80) {
      i++;
    }

    // This loop optimizes for chars less than 0x800.
    for (; i < utf16Length; i++) {
      char c = sequence.charAt(i);