Após isso, o cliente e o servidor possuem uma conexão TCP encriptada, que é provida pelo TLS. E então eles podem usar essa conexão para começar a comunicação HTTP propriamente dita.

E isso resume o que é HTTPS, apenas HTTP simples dentro de uma conexão TLS segura em vez de uma conexão TCP pura (não encriptada).

/// tip | Dica

接著，用戶端與 TLS 終止代理會以該憑證為基礎，**協商後續如何加密**整段 **TCP 通訊**。至此完成 **TLS 握手**。

之後，用戶端與伺服器之間就有一條**已加密的 TCP 連線**，這就是 TLS 所提供的能力。接著他們可以在這條連線上開始實際的 **HTTP** 通訊。

這也就是 **HTTPS** 的本質：在**安全的 TLS 連線**內傳送一般的 **HTTP**，而非在純（未加密）的 TCP 連線上。

/// tip

請注意，加密發生在 **TCP 層**，不是在 HTTP 層。

///

### HTTPS 請求 { #https-request }

現在用戶端與伺服器（更精確地說，是瀏覽器與 TLS 終止代理）之間已有**加密的 TCP 連線**，他們可以開始進行 **HTTP** 通訊。

        props.setProperty("jcifs.smb.client.useRDMA", "true");
        props.setProperty("jcifs.smb.client.rdmaProvider", "tcp");

        PropertyConfiguration config = new PropertyConfiguration(props);

        assertTrue(config.isUseRDMA(), "RDMA should be enabled");
        assertEquals("tcp", config.getRdmaProvider(), "Provider should be tcp");
    }

    @Test
    public void testRdmaAutoProvider() throws CIFSException {

cd "$SCRIPT_DIR"

which wstest
if [ $? != 0 ]; then
  echo "Run 'pip install autobahntestsuite', maybe with 'sudo'."
  exit 1
fi
which jq
if [ $? != 0 ]; then
  echo "Run 'brew install jq'"
  exit 1
fi

trap 'kill $(jobs -pr)' SIGINT SIGTERM EXIT

set -ex

wstest -m fuzzingserver -s fuzzingserver-config.json &
sleep 2 # wait for wstest to start

java -jar target/okhttp-tests-*-jar-with-dependencies.jar

然后，通过使用证书，客户端和 TLS 终止代理 **决定如何加密** **TCP 通信** 的其余部分。 这就完成了 **TLS 握手** 部分。

此后，客户端和服务器就拥有了**加密的 TCP 连接**，这就是 TLS 提供的功能。 然后他们可以使用该连接来启动实际的 **HTTP 通信**。

这就是 **HTTPS**，它只是 **安全 TLS 连接** 内的普通 **HTTP**，而不是纯粹的（未加密的）TCP 连接。

/// tip | 提示

请注意，通信加密发生在 **TCP 层**，而不是 HTTP 层。

///

### HTTPS 请求 { #https-request }

现在客户端和服务器（特别是浏览器和 TLS 终止代理）具有 **加密的 TCP 连接**，它们可以开始 **HTTP 通信**。

// not be used after the arena is freed. Accessing the underlying storage of the
// slice after free may result in a fault, but this fault is also not guaranteed.
func MakeSlice[T any](a *Arena, len, cap int) []T {
	var sl []T
	runtime_arena_arena_Slice(a.a, &sl, cap)
	return sl[:len]
}

// Clone makes a shallow copy of the input value that is no longer bound to any
// arena it may have been allocated from, returning the copy. If it was not

package http

import (
	"bufio"
	"bytes"
	"errors"
	"fmt"
	"io"
	"net"
	"net/http"
	"time"

	"github.com/klauspost/compress/gzip"
)

// ResponseRecorder - is a wrapper to trap the http response
// status code and to record the response body
type ResponseRecorder struct {
	http.ResponseWriter
	io.ReaderFrom
	StatusCode int
	// Log body of 4xx or 5xx responses
	LogErrBody bool

		"Total number of failed internode calls")
	internodeDialedErrorsTotalMD = NewCounterMD(internodeDialErrorsTotal,
		"Total number of internode TCP dial timeouts and errors")
	internodeDialAvgTimeNanosMD = NewGaugeMD(internodeDialAvgTimeNanos,
		"Average dial time of internode TCP calls in nanoseconds")
	internodeSentBytesTotalMD = NewCounterMD(internodeSentBytesTotal,
		"Total number of bytes sent to other peer nodes")

func (b *Buffer) Len() int { return len(b.buf) - b.off }

// Cap returns the capacity of the buffer's underlying byte slice, that is, the
// total space allocated for the buffer's data.
func (b *Buffer) Cap() int { return cap(b.buf) }

// Available returns how many bytes are unused in the buffer.
func (b *Buffer) Available() int { return cap(b.buf) - len(b.buf) }

done
set -e

# Single handler for all cleanup actions, triggered on an EXIT trap.
# TODO(angerson) Making this use different scripts may be overkill.
cleanup() {
  if [[ "$TFCI_DOCKER_ENABLE" == 1 ]]; then
    ./ci/official/utilities/cleanup_docker.sh
  fi
  ./ci/official/utilities/cleanup_summary.sh
}