// tells the computing thread when to start computing
    CountDownLatch computeSignal = new CountDownLatch(1);
    // tells the main thread when computation is pending
    CountDownLatch secondSignal = new CountDownLatch(1);
    // tells the main thread when the second get has started
    CountDownLatch thirdSignal = new CountDownLatch(1);
    // tells the main thread when the third get has started

There are a few differences:

* `table=True` tells SQLModel that this is a *table model*, it should represent a **table** in the SQL database, it's not just a *data model* (as would be any other regular Pydantic class).

* `Field(primary_key=True)` tells SQLModel that the `id` is the **primary key** in the SQL database (you can learn more about SQL primary keys in the SQLModel docs).

   * update the underlying table, and vice versa. The cell set does not support the {@code add} or
   * {@code addAll} methods.
   *
   * @return set of table cells consisting of row key / column key / value triplets
   */
  Set<Cell<R, C, V>> cellSet();

  /**
   * Returns a set of row keys that have one or more values in the table. Changes to the set will

We are going to call them "**operations**" too.

#### Define a *path operation decorator* { #define-a-path-operation-decorator }

{* ../../docs_src/first_steps/tutorial001.py hl[6] *}

The `@app.get("/")` tells **FastAPI** that the function right below is in charge of handling requests that go to:

* the path `/`
* using a <abbr title="an HTTP GET method"><code>get</code> operation</abbr>

/// info | `@decorator` Info

				return i + 1, data[:i], nil
			}
		}
		if !atEOF {
			return 0, nil, nil
		}
		// There is one final token to be delivered, which may be the empty string.
		// Returning bufio.ErrFinalToken here tells Scan there are no more tokens after this
		// but does not trigger an error to be returned from Scan itself.
		return 0, data, bufio.ErrFinalToken
	}
	scanner.Split(onComma)
	// Scan.
	for scanner.Scan() {

      C columnKey = cell.getColumnKey();
      V value = cell.getValue();

      /*
       * These requireNonNull calls are safe because we construct the maps to hold all the provided
       * cells.
       */
      cellRowIndices[i] = requireNonNull(rowIndex.get(rowKey));
      Map<C, V> thisRow = requireNonNull(rows.get(rowKey));
      cellColumnInRowIndices[i] = thisRow.size();

You can run your tests as usual via:

<div class="termy">

```console
$ pytest

---> 100%
```

</div>

## In Detail { #in-detail }

The marker `@pytest.mark.anyio` tells pytest that this test function should be called asynchronously:

{* ../../docs_src/async_tests/test_main.py hl[7] *}

/// tip

If it doesn't receive it, it returns an HTTP 401 "Unauthorized" error.

And returns a header `WWW-Authenticate` with a value of `Basic`, and an optional `realm` parameter.

That tells the browser to show the integrated prompt for a username and password.

Then, when you type that username and password, the browser sends them in the header automatically.

## Simple HTTP Basic Auth { #simple-http-basic-auth }

 *   }
 * }
 *
 * public void testMySubtypes() throws Exception {
 *   new MySubtypeTests().testAllDeclarations();
 * }
 * }
 *
 * The calls to {@link #isSubtype} and {@link #notSubtype} tells the framework what assertions need
 * to be made.
 *
 * <p>The declaration methods must be public.
 */
@AndroidIncompatible // only used by android incompatible tests.
@NullUnmarked

          }
        };

    recursiveCache = CacheBuilder.newBuilder().weakKeys().weakValues().build(recursiveLoader);
    cacheRef.set(recursiveCache);

    // tells the test when the computation has completed
    CountDownLatch doneSignal = new CountDownLatch(1);

    Thread thread =
        new Thread() {
          @Override
          public void run() {
            try {