## ¿Tienes prisa?

<abbr title="too long; didn't read"><strong>TL;DR:</strong></abbr>

Si estás utilizando libraries de terceros que te dicen que las llames con `await`, del tipo:

```Python
results = await some_library()
```

Entonces declara tus *path operation functions* con `async def` de la siguiente manera:

```Python hl_lines="2"
@app.get('/')
async def read_results():
    results = await some_library()

    call: Call,
    delay: Long,
  ): Thread {
    val thread =
      object : Thread("canceler") {
        override fun run() {
          try {
            sleep(delay)
          } catch (e: InterruptedException) {
            throw AssertionError()
          }
          call.cancel()
        }
      }
    thread.start()
    return thread
  }

    assertThat(recordedRequest!!.method).isEqualTo("GET")
    assertThat(recordedRequest.path).isEqualTo(
      "/lookup?ct&dns=AAABAAABAAAAAAAABmdvb2dsZQNjb20AAAEAAQ",
    )
    Thread.sleep(2000)
    server.enqueue(
      dnsResponse(
        "0000818000010003000000000567726170680866616365626f6f6b03636f6d0000010001c00c000500010" +

# To simplify the automations section, some calculations are placed under unique YAML keys defined here.
# Read the "|" not as "or", but as a "pipe", taking the output of the previous command and passing it to the next command.
# This section could also appear ahead of the automations section.

colors:

        args:
          approvals: 2

# To simplify the automations section, some calculations are placed under unique YAML keys defined here.
# Read the "|" not as "or", but as a "pipe", taking the output of the previous command and passing it to the next command.
# This section could also appear ahead of the automations section.

   * cache size. The application needs to be aware of calling this function during the
   * initialization phase and preferably in a background worker thread.
   *
   * Note that if the application chooses to not call this method to initialize the cache. By
   * default, OkHttp will perform lazy initialization upon the first usage of the cache.
   */

    // Enqueueing the red task starts a thread because the head of the queue changed.
    redQueue.execute("red task") {
      log += "red:starting@${taskFaker.nanoTime}"
      taskFaker.sleep(100.µs)
      log += "red:finishing@${taskFaker.nanoTime}"
    }
    assertThat(taskFaker.executeCallCount).isEqualTo(1)

    // Enqueueing the blue task doesn't start a thread because the red one is still starting.

#### 🗒

❌:

```Python
connect_args={"check_same_thread": False}
```

...💪 🕴 `SQLite`. ⚫️ 🚫 💪 🎏 💽.

!!! info "📡 ℹ"

    🔢 🗄 🔜 🕴 ✔ 1️⃣ 🧵 🔗 ⏮️ ⚫️, 🤔 👈 🔠 🧵 🔜 🍵 🔬 📨.

    👉 ❎ 😫 🤝 🎏 🔗 🎏 👜 (🎏 📨).

    ✋️ FastAPI, ⚙️ 😐 🔢 (`def`) 🌅 🌘 1️⃣ 🧵 💪 🔗 ⏮️ 💽 🎏 📨, 👥 💪 ⚒ 🗄 💭 👈 ⚫️ 🔜 ✔ 👈 ⏮️ `connect_args={"check_same_thread": False}`.

import org.jetbrains.kotlin.utils.exceptions.errorWithAttachment
import kotlin.contracts.ExperimentalContracts
import kotlin.contracts.contract

/**
 * Maps FirElement to KtSymbol & ConeType to KtType, thread safe
 */
internal class KtSymbolByFirBuilder(
    private val project: Project,
    val analysisSession: KtFirAnalysisSession,
    val token: KtLifetimeToken,
) {

```

#### 注意

参数:

```Python
connect_args={"check_same_thread": False}
```

...仅用于`SQLite`，在其他数据库不需要它。

!!! info "技术细节"

    默认情况下，SQLite 只允许一个线程与其通信，假设有多个线程的话，也只将处理一个独立的请求。

    这是为了防止意外地为不同的事物（不同的请求）共享相同的连接。

    但是在 FastAPI 中，普遍使用def函数，多个线程可以为同一个请求与数据库交互，所以我们需要使用`connect_args={"check_same_thread": False}`来让SQLite允许这样。

    此外，我们将确保每个请求都在依赖项中获得自己的数据库连接会话，因此不需要该默认机制。