In our code example above, we don't use it directly, but we pass it to FastAPI for it to use it.

The `lifespan` parameter of the `FastAPI` app takes an **async context manager**, so we can pass our new `lifespan` async context manager to it.

```Python hl_lines="22"
{!../../../docs_src/events/tutorial003.py!}
```

    // we pad because the dummy hash function we use to test this, merely copies the input into
    // the output, so the input must be at least 32 bits, since the output has to be that long
    assertPutString(new char[] {'p', HashTestUtils.randomLowSurrogate(new Random())});
  }

  public void testPutStringWithHighSurrogate() {
    // we pad because the dummy hash function we use to test this, merely copies the input into

    // we pad because the dummy hash function we use to test this, merely copies the input into
    // the output, so the input must be at least 32 bits, since the output has to be that long
    assertPutString(new char[] {'p', HashTestUtils.randomLowSurrogate(new Random())});
  }

  public void testPutStringWithHighSurrogate() {
    // we pad because the dummy hash function we use to test this, merely copies the input into

// gofmtLineReplacer is used to put a gofmt-formatted string for an
// AST expression onto a single line. The lexer normally inserts a
// semicolon at each newline, so we can replace newline with semicolon.
// However, we can't do that in cases where the lexer would not insert
// a semicolon. We only have to worry about cases that can occur in an
// expression passed through gofmt, which means composite literals and

      if (!firstTimeSeen) {
        /*
         * We've seen this, so we've seen its causes, too. No need to re-add them. (There's one case
         * where this isn't true, but we ignore it: If we record an exception, then someone calls
         * initCause() on it, and then we examine it again, we'll conclude that we've seen the whole
         * chain before when in fact we haven't. But this should be rare.)
         */

Contributing
============

Keeping the project small and stable limits our ability to accept new contributors. We are not
seeking new committers at this time, but some small contributions are welcome.

If you've found a security problem, please follow our [bug bounty][security] program.

If you've found a bug, please contribute a failing test case so we can study and fix it.

If you have a new feature idea, please build it in an external library. There are

Contributing
============

Keeping the project small and stable limits our ability to accept new contributors. We are not
seeking new committers at this time, but some small contributions are welcome.

If you've found a security problem, please follow our [bug bounty][security] program.

If you've found a bug, please contribute a failing test case so we can study and fix it.

If you have a new feature idea, please build it in an external library. There are

  </plugin>
  .
  .
  .
</plugins>

   * reclaimed.
   *
   * If any of this fails along the way, we fall back to loading Finalizer directly in the
   * application class loader.
   *
   * NOTE: The tests for this behavior (FinalizableReferenceQueueClassLoaderUnloadingTest) fail
   * strangely when run in JDK 9. We are considering this a known issue. Please see
   * https://github.com/google/guava/issues/3086 for more information.

   * reclaimed.
   *
   * If any of this fails along the way, we fall back to loading Finalizer directly in the
   * application class loader.
   *
   * NOTE: The tests for this behavior (FinalizableReferenceQueueClassLoaderUnloadingTest) fail
   * strangely when run in JDK 9. We are considering this a known issue. Please see
   * https://github.com/google/guava/issues/3086 for more information.