* to reduce cache contention on most processors. Padding is
   * overkill for most Atomics because they are usually irregularly
   * scattered in memory and thus don't interfere much with each
   * other. But Atomic objects residing in arrays will tend to be
   * placed adjacent to each other, and so will most often share
   * cache lines (with a huge negative performance impact) without
   * this precaution.

   * <p>It should be noted that using a {@code Striped<L>} with relatively few stripes, and {@code
   * bulkGet(keys)} with a relative large number of keys can cause an excessive number of shared
   * stripes (much like the birthday paradox, where much fewer than anticipated birthdays are needed
   * for a pair of them to match). Please consider carefully the implications of the number of

---

If you just don't know, use normal `def`.

---

**Note**: You can mix `def` and `async def` in your *path operation functions* as much as you need and define each one using the best option for you. FastAPI will do the right thing with them.

Anyway, in any of the cases above, FastAPI will still work asynchronously and be extremely fast.

### Don't worry about style { #dont-worry-about-style }

* Don't worry too much about things like commit message styles, I will squash and merge customizing the commit manually.

* Also don't worry about style rules, there are already automatized tools checking that.

  /**
   * This strategy uses all 128 bits of {@link Hashing#murmur3_128} when hashing. It looks different
   * from the implementation in MURMUR128_MITZ_32 because we're avoiding the multiplication in the
   * loop and doing a (much simpler) += hash2. We're also changing the index to a positive number by
   * AND'ing with Long.MAX_VALUE instead of flipping the bits.
   */
  MURMUR128_MITZ_64() {
    @Override

  private const val TMAX = 26
  private const val SKEW = 38
  private const val DAMP = 700
  private const val INITIAL_BIAS = 72
  private const val INITIAL_N = 0x80

  /**
   * Returns null if any label is oversized so much that the encoder cannot encode it without
   * integer overflow. This will not return null for labels that fit within the DNS size
   * limits.
   */
  fun encode(string: String): String? {
    var pos = 0

   * we've chosen to instead use a signature that benefits existing users -- and is still safe.
   *
   * For convertAll, I haven't looked as closely at *how* much existing users benefit, so we should
   * keep an eye out for problems that new users encounter. Note also that convertAll could support
   * both use cases by using @PolyNull. (By contrast, we can't use @PolyNull for our superinterface

    return -1;
  }

  /**
   * Returns the start position of the first occurrence of the specified {@code target} within
   * {@code array}, or {@code -1} if there is no such occurrence.
   *
   * <p>More formally, returns the lowest index {@code i} such that {@code Arrays.copyOfRange(array,
   * i, i + target.length)} contains exactly the same elements as {@code target}.
   *

        // snapshot timestamp, or some other timestamp later encoded into the metadata.
        // TODO this needs to be repeated here so the merging doesn't interfere with the written metadata
        //  - we'd be much better having a pristine input, and an ongoing metadata for merging instead

        Map<ArtifactRepository, Metadata> previousMetadata = new HashMap<>();
        ArtifactRepository selected = null;

 * entries may be counted by {@link Map#size}, but will never be visible to read or write
 * operations. A partially-reclaimed entry is never exposed to the user. Any {@link Map.Entry}
 * instance retrieved from the map's {@linkplain Map#entrySet entry set} is a snapshot of that
 * entry's state at the time of retrieval; such entries do, however, support {@link