*       thread cleared it, we still know that it's not associated with our thread
     *   <li>If this field's value == null because it was associated with our thread and was
     *       cleared, we know that we're not executing inline any more
     * </ul>
     *
     * All the states where thread != currentThread are identical for our purposes, and so even
     * though it's racy, we don't care which of those values we get, so no need to synchronize.

  chmod +x "/usr/local/bin/bazel"
fi

# "TFCI_MACOS_UPGRADE_PYENV_ENABLE" is used to decide if we need to upgrade the
# Pyenv version. We enable this for macOS x86 builds as the default Pyenv on
# those VMs does not support installing Python 3.12 and above which we need
# for running smoke tests in nightly/release wheel builds.
if [[ "${TFCI_MACOS_UPGRADE_PYENV_ENABLE}" == 1 ]]; then

- **Reduce the size of the Gradle distribution** - We no longer need to carry multiple standard libraries for different languages.

## Decision

We do not introduce new public API methods that include Groovy types in their signatures.
Existing Groovy methods will not be removed immediately.

## Status

ACCEPTED

## Consequences

**Experimental!**

We'd like to capture our architectural decisions about the build tool as [Architectural Decision Records (ADRs)](https://adr.github.io/).
For now we just have this global repository of ADRs.
If we see fit, we can break these out to per-platform ones, or keep a hybrid approach to having global and platform-specific ADSs.

Our aim is to keep the process lightweight and approachable.

So, if we create a Pydantic object `user_in` like:

```Python
user_in = UserIn(username="john", password="secret", email="******@****.***")
```

and then we call:

```Python
user_dict = user_in.dict()
```

we now have a `dict` with the data in the variable `user_dict` (it's a `dict` instead of a Pydantic model object).

And if we call:

```Python
print(user_dict)
```

	// once the pod restarts (in CrashLoopBackoff), which can take some time.
	// We don't want to constantly try to apply the iptables rules, which is unneeded and will fail.
	// Instead, we track which UIDs we repaired and skip them if already repaired.
	//
	// An alternative would be to write something to the Pod (status, annotation, etc).
	// However, this requires elevated privileges we want to avoid
	if uid, f := c.repairedPods[key]; f {

 *  Breaking: Drop support for Kotlin Multiplatform.

    We planned to support multiplatform in OkHttp 5.0, but after building it, we weren't happy with
    the implementation trade-offs. We can't use our HTTP client engine on Kotlin/JS, and we weren't
    prepared to build a TLS API for Kotlin/Native.

    We'd prefer a multiplatform HTTP client API that's backed by OkHttp on Android and JVM, and

    could use new flow control capacity before acknowledging it, causing strict HTTP/2 servers to
    fail the call.


## Version 4.1.1

_2019-09-05_

 *  Fix: Don't drop repeated headers when validating cached responses. In our Kotlin upgrade we
    introduced a regression where we iterated the number of unique header names rather than then

   * @return a long of a concatenated 8 bytes
   */
  static long load64Safely(byte[] input, int offset, int length) {
    long result = 0;
    // Due to the way we shift, we can stop iterating once we've run out of data, the rest
    // of the result already being filled with zeros.

    // This loop is critical to performance, so please check HashBenchmark if altering it.
    int limit = Math.min(length, 8);

## Decision

We do not use Java serialization. Instead, we use custom serialization where we explicitly describe how data objects should be serialized and deserialized.

For internal purposes, we use binary formats for their brevity.
We use the `Serializer` abstraction to separate the actual implementation of serialization from its uses.

When sharing data with external tools, we use JSON.

## Status

PROPOSED