Because of that, it's now recommended to instead use the `lifespan` as explained above.

## Technical Details { #technical-details }

Just a technical detail for the curious nerds. 🤓

/*
 * Written by Doug Lea with assistance from members of JCP JSR-166
 * Expert Group and released to the public domain, as explained at
 * http://creativecommons.org/publicdomain/zero/1.0/
 */

/*
 * Source:
 * http://gee.cs.oswego.edu/cgi-bin/viewcvs.cgi/jsr166/src/jsr166e/extra/AtomicDoubleArray.java?revision=1.5
 * (Modified to adapt to guava coding conventions and
 * to use AtomicLongArray instead of sun.misc.Unsafe)
 */

/*
 * Written by Doug Lea with assistance from members of JCP JSR-166
 * Expert Group and released to the public domain, as explained at
 * http://creativecommons.org/publicdomain/zero/1.0/
 */

/*
 * Source:
 * http://gee.cs.oswego.edu/cgi-bin/viewcvs.cgi/jsr166/src/jsr166e/Striped64.java?revision=1.9
 */

package com.google.common.cache;

import com.google.common.annotations.GwtIncompatible;
import java.lang.reflect.Field;

/*
 * Written by Doug Lea and Martin Buchholz with assistance from
 * members of JCP JSR-166 Expert Group and released to the public
 * domain, as explained at
 * http://creativecommons.org/publicdomain/zero/1.0/
 */

/*
 * Source:
 * http://gee.cs.oswego.edu/cgi-bin/viewcvs.cgi/jsr166/src/jsr166e/extra/AtomicDouble.java?revision=1.13
 * (Modified to adapt to guava coding conventions and
 * to use AtomicLongFieldUpdater instead of sun.misc.Unsafe)
 */

## How to Test Your Changes to TensorFlow

You may check how your changes will affect TensorFlow by:

1. Creating a PR and observing the presubmit test results
2. Running the CI scripts locally, as explained below
3. **Google employees only**: Google employees can use an internal-only tool
called "MLCI" that makes testing more convenient: it can execute any full CI job
against a pending change. Search for "MLCI" internally to find it.

/*
 * Written by Doug Lea and Martin Buchholz with assistance from
 * members of JCP JSR-166 Expert Group and released to the public
 * domain, as explained at
 * http://creativecommons.org/publicdomain/zero/1.0/
 */

/*
 * Source:
 * http://gee.cs.oswego.edu/cgi-bin/viewcvs.cgi/jsr166/src/test/tck-jsr166e/AtomicDoubleArrayTest.java?revision=1.13
 * (Modified to adapt to guava coding conventions)
 */

package com.google.common.util.concurrent;

This is because by default, FastAPI will inspect every item inside and make sure it is serializable as JSON, using the same [JSON Compatible Encoder](../tutorial/encoder.md){.internal-link target=_blank} explained in the tutorial. This is what allows you to return **arbitrary objects**, for example database models.

/// tip

Review what each line does by clicking each number bubble in the code. 👆

///

/// warning

Make sure to **always** use the **exec form** of the `CMD` instruction, as explained below.

///

#### Use `CMD` - Exec Form { #use-cmd-exec-form }

 * which contained the following notice:
 *
 * Written by Doug Lea with assistance from members of JCP JSR-166
 * Expert Group and released to the public domain, as explained at
 * http://creativecommons.org/publicdomain/zero/1.0/
 * Other contributors include Andrew Wright, Jeffrey Hayes,
 * Pat Fisher, Mike Judd.
 */

package com.google.common.util.concurrent;

### Return a Response Directly { #return-a-response-directly }

The most common case would be [returning a Response directly as explained later in the advanced docs](../advanced/response-directly.md){.internal-link target=_blank}.

{* ../../docs_src/response_model/tutorial003_02.py hl[8,10:11] *}