So, in our case, if we set the Peewee state variable (with a default `dict`) in the `async` dependency, all the rest of the internal code in our app will see this value and will be able to reuse it for the whole request.

And the context variable would be set again for the next request, even if they are concurrent.

### Set database state in the dependency `get_db()`

- [`KT-60132`](https://youtrack.jetbrains.com/issue/KT-60132) K2: properties and functions without a name should be re-analyzable as well
- [`KT-59199`](https://youtrack.jetbrains.com/issue/KT-59199) K2 IDE: PSI changes which do not cause OOB modifications can be unseen from the FIR elements
- [`KT-59667`](https://youtrack.jetbrains.com/issue/KT-59667) Analysis API: PsiInvalidElementAccessException from JavaClassifierTypeImpl.substitutor

            PluginEntry(id, implementationClass)
        }.toList()
    }
} catch (cause: IOException) {
    throw IllegalArgumentException(
        "Failed to extract plugin metadata from '" + jar.path + "'",
        cause
    )
}


private
fun isGradlePluginPropertiesFile(entry: JarEntry) = entry.run {

    Note over client,tasks: Can raise exception for dependency, handled after response is sent
    Note over client,operation: Can raise HTTPException and can change the response
    client ->> dep: Start request
    Note over dep: Run code up to yield
    opt raise
        dep -->> handler: Raise HTTPException
        handler -->> client: HTTP error response
        dep -->> dep: Raise other exception
    end

     *
     * If too many type arguments are provided, a mapping is still created. Extra type arguments are simply ignored. If this wasn't the
     * case, the resulting [KtCall] would contain no type arguments at all, which can cause problems later. If too few type arguments are
     * provided, an empty map is returned defensively so that [toTypeArgumentsMapping] doesn't conjure any error types. If you want to map

  // daemonset on any given node can double if the readiness check fails, and
  // so resource intensive daemonsets should take into account that they may
  // cause evictions during disruption.
  // +optional
  optional k8s.io.apimachinery.pkg.util.intstr.IntOrString maxSurge = 2;
}

// Spec to control the desired behavior of rolling update.
message RollingUpdateDeployment {

* Fix a race condition in the scheduler preemption logic that could cause nominatedNodeName of a pod not to be considered in one or more scheduling cycles. ([#72504](https://github.com/kubernetes/kubernetes/pull/72504), [@bsalamat](https://github.com/bsalamat))

    ```Python hl_lines="68-69  81"
    {!> ../../../docs_src/websockets/tutorial002.py!}
    ```

!!! info
    As this is a WebSocket it doesn't really make sense to raise an `HTTPException`, instead we raise a `WebSocketException`.

    You can use a closing code from the <a href="https://tools.ietf.org/html/rfc6455#section-7.4.1" class="external-link" target="_blank">valid codes defined in the specification</a>.

  // daemonset on any given node can double if the readiness check fails, and
  // so resource intensive daemonsets should take into account that they may
  // cause evictions during disruption.
  // This is an alpha field and requires enabling DaemonSetUpdateSurge feature gate.
  // +optional
  optional k8s.io.apimachinery.pkg.util.intstr.IntOrString maxSurge = 2;
}

`HTTPException`, ao fundo, nada mais é do que a conjunção entre uma exceção comum do Python e informações adicionais relevantes para APIs.

E porque é uma exceção do Python, você não **retorna** (return) o `HTTPException`, você lança o (raise) no seu código.