byte[] data = new byte[] { 1, 2, 3, 4, 5 };
        InputStream in = new ByteArrayInputStream(concat(messageHeader(5), data));
        SocketInputStream sis = new SocketInputStream(in);

        // Read in chunks
        byte[] buffer = new byte[10];

        // The read will either:
        // 1. Successfully read all 5 bytes
        // 2. Throw IOException if it tries to read past the message
        try {

    }

    /**
     * Performs a streaming chat completion request using the configured LLM client.
     *
     * @param request the chat request
     * @param callback the callback to receive streaming chunks
     * @throws LlmException if LLM is not available or an error occurs
     */
    public void streamChat(final LlmChatRequest request, final LlmStreamCallback callback) {

                callback.onPhaseStart(ChatPhaseCallback.PHASE_ANSWER, "Generating response...");
                llmClientManager.generateUnclearIntentResponse(userMessage, history, (chunk, done) -> {
                    fullResponse.append(chunk);
                    callback.onChunk(chunk, done);
                });
                callback.onPhaseComplete(ChatPhaseCallback.PHASE_ANSWER);
                if (logger.isDebugEnabled()) {

    String regexPattern = "(?<=:)";
    Iterable<String> split = Splitter.onPattern(regexPattern).split(toSplit);
    assertThat(split).containsExactly(":", "foo:", ":", "barbaz:").inOrder();
    // splits into chunks ending in :
  }

  @J2ktIncompatible // Kotlin Native's regex is based on Apache Harmony, like old Android
  @GwtIncompatible // java.util.regex.Pattern

- `kubectl describe` will by default fetch large lists of resources in chunks of up to 500 items rather than requesting all resources up front from the server. A new flag `--chunk-size=SIZE` may be used to alter the number of items or disable this feature when `0` is passed.  This is a beta feature. ([#101171](https://github.com/kubernetes/kubernetes/pull/101171), [@KnVere...

* `kubectl get` will by default fetch large lists of resources in chunks of up to 500 items rather than requesting all resources up front from the server. This reduces the perceived latency of managing large clusters since the server returns the first set of results to the client much more quickly.  A new flag `--chunk-size=SIZE` may be used to alter the number of items or disable this feature when `0` is passed.  This is a beta feature. (...

The Kubernetes API server now supports the ability to break large LIST calls into multiple smaller chunks. A client can specify a limit to the number of results to return. If more results exist, a token is returned that allows the client to continue the previous list call repeatedly until all results are retrieved.  The resulting list is identical to a list call that does not perform chunking, thanks to capabilities provided by etcd3.  This allows the server to use less memory and CPU when very large...

            // Given
            SrvCopychunk chunk1 = new SrvCopychunk(100, 200, 300);
            SrvCopychunk chunk2 = new SrvCopychunk(400, 500, 600);
            SrvCopychunk chunk3 = new SrvCopychunk(700, 800, 900);
            byte[] buffer = new byte[EXPECTED_SIZE * 3];

            // When
            int offset1 = chunk1.encode(buffer, 0);
            int offset2 = chunk2.encode(buffer, offset1);

    ByteSource source = out.asByteSource();
    int chunk1 = min(dataSize, fileThreshold);
    int chunk2 = dataSize - chunk1;

    // Write just enough to not trip the threshold
    if (chunk1 > 0) {
      write(out, data, 0, chunk1, singleByte);
      assertTrue(ByteSource.wrap(data).slice(0, chunk1).contentEquals(source));
    }
    File file = out.getFile();
    assertThat(file).isNull();

* Add --image-repository flag to "kubeadm config images". ([#75866](https://github.com/kubernetes/kubernetes/pull/75866), [@jmkeyes](https://github.com/jmkeyes))
* Paginate requests from the kube-apiserver watch cache to etcd in chunks. ([#75389](https://github.com/kubernetes/kubernetes/pull/75389), [@jpbetz](https://github.com/jpbetz))
    * Paginate reflector init and resync List calls that are not served by watch cache.