* size to reach 0, causing an infinite loop.
   */
  public void testCopyWithReaderThatDoesNotFillBuffer() throws IOException {
    // need a long enough string for the buffer to hit 0 remaining before the copy completes
    String string = Strings.repeat("0123456789", 100);
    StringBuilder b = new StringBuilder();

   * size to reach 0, causing an infinite loop.
   */
  public void testCopyWithReaderThatDoesNotFillBuffer() throws IOException {
    // need a long enough string for the buffer to hit 0 remaining before the copy completes
    String string = Strings.repeat("0123456789", 100);
    StringBuilder b = new StringBuilder();

    bytesIn.writeUtf8("custom-baz")
    bytesIn.writeByte(0x0d) // Literal value (len = 13)
    bytesIn.writeUtf8("custom-header")

    // Set to only support 110 bytes (enough for 2 headers).
    // Use a new Writer because we don't support change the dynamic table
    // size after Writer constructed.
    val writer = Hpack.Writer(110, false, bytesOut)
    writer.writeHeaders(headerBlock)

 *       {@code fail}, {@code assertTrue}, etc.) It is OK (but not particularly recommended) for
 *       other code to use these forms too. Only the most typically used JUnit assertion methods are
 *       defined this way, but enough to live with.
 *   <li>If you override {@link #setUp} or {@link #tearDown}, make sure to invoke {@code
 *       super.setUp} and {@code super.tearDown} within them. These methods are used to clear and

		has := int64(len(buf))
		if has >= n {
			return nil
		}
		if hasFull || n > size {
			return io.ErrUnexpectedEOF
		}
		extra := n - has
		if int64(cap(buf)) >= n {
			// Extend since we have enough space.
			buf = buf[:n]
		} else {
			buf = append(buf, make([]byte, extra)...)
		}
		_, err := io.ReadFull(r, buf[has:])
		if err != nil {
			if errors.Is(err, io.EOF) {

      checkNotNull(element);
      copyIfNecessary();
      if (n == elements.length) {
        sortAndDedup();
        /**
         * sortAndDedup may have made enough room for this element, but that's not necessarily good
         * enough. Consider, for example, the case where we have a buffer of size (n+1), add n
         * distinct elements, and add the last element over again many times over. We don't want a

      Request.Builder()
        .url(server.url("/"))
        .post(AsyncRequestBody())
        .build()
    val call = client.newCall(request)
    call.timeout()
      .timeout(500, TimeUnit.MILLISECONDS) // Long enough for the first TLS handshake.
    call.execute().use { response ->
      val requestBody = (call.request().body as AsyncRequestBody?)!!.takeSink()
      val responseBody = response.body.source()

 * `Function<@Nullable A, @Nullable B>` already had to find a workaround. Still, there is a *ton* of
 * fallout from trying to switch. I would be shocked if the switch would offer benefits to anywhere
 * near enough users to justify the costs.
 *
 * Fortunately, if anyone does want to use a Converter as a `Function<@Nullable A, @Nullable B>`,
 * it's easy to get one: `converter::convert`.
 *

///

/// warning

These are very technical details about the standards **JSON Schema** and **OpenAPI**.

If the ideas above already work for you, that might be enough, and you probably don't need these details, feel free to skip them.

///

Before OpenAPI 3.1.0, OpenAPI used an older and modified version of **JSON Schema**.

         * I'm not sure that we could actually get here for *Android*: I would expect us to enter
         * the POSIX code path instead. And if we tried this code path, we'd have trouble unless we
         * were running under a new enough version of Android to support NIO.
         *
         * So this is probably just the "Windows Java 8" case. In that case, if we wanted *another*
         * layer of fallback before consulting the system property, we could try