* Si estás comparando Starlette, compáralo con Sanic, Flask, Django, etc. Frameworks web (o microframeworks).
* **FastAPI**:
    * De la misma manera que Starlette usa Uvicorn y no puede ser más rápido que él, **FastAPI** usa Starlette, por lo que no puede ser más rápido que él.

    * Cependant il nous apporte les outils pour construire une application web simple, avec un routage basé sur des chemins, etc.
    * Si on compare Starlette, il faut le comparer à d'autres frameworks web (ou micorframework) comme Sanic, Flask, Django, etc.
* **FastAPI** :
    * Comme Starlette, FastAPI utilise Uvicorn et ne peut donc pas être plus rapide que ce dernier.

Нужна мощная система внедрения зависимостей. Необходим способ для уменьшения повторов кода.

///

### <a href="https://sanic.readthedocs.io/en/latest/" class="external-link" target="_blank">Sanic</a>

Sanic был одним из первых чрезвычайно быстрых Python-фреймворков основанных на `asyncio`.
Он был сделан очень похожим на Flask.

/// note | "Технические детали"

// Exported for testing only.
import (
	"unicode/utf8"
)

var IsSpace = isSpace

const DefaultBufSize = defaultBufSize

func (s *Scanner) MaxTokenSize(n int) {
	if n < utf8.UTFMax || n > 1e9 {
		panic("bad max token size")
	}
	if n < len(s.buf) {
		s.buf = make([]byte, n)
	}
	s.maxTokenSize = n
}

// ErrOrEOF is like Err, but returns EOF. Used to test a corner case.
func (s *Scanner) ErrOrEOF() error {

		if globalLocalNodeName != "" {
			w.Header().Set(xhttp.AmzRequestHostID, globalLocalNodeNameHex)
		}
		h.ServeHTTP(w, r)
	})
}

// criticalErrorHandler handles panics and fatal errors by
// `panic(logger.ErrCritical)` as done by `logger.CriticalIf`.
//
// It should be always the first / highest HTTP handler.
func setCriticalErrorHandler(h http.Handler) http.Handler {

of the location.
</p>


<h3 id="Handling_panics">Handling panics</h3>

<p> Two built-in functions, <code>panic</code> and <code>recover</code>,
assist in reporting and handling <a href="#Run_time_panics">run-time panics</a>
and program-defined error conditions.
</p>

<pre class="grammar">
func panic(interface{})
func recover() interface{}
</pre>

<p>

func close(c chan<- Type)

// The panic built-in function stops normal execution of the current
// goroutine. When a function F calls panic, normal execution of F stops
// immediately. Any functions whose execution was deferred by F are run in
// the usual way, and then F returns to its caller. To the caller G, the
// invocation of F then behaves like a call to panic, terminating G's

// as we only guarantee a lookback buffer size of one, any subsequent calls to back()
// before calling next() may panic
func (s *scanner) back() {
	if s.ipos <= 0 {
		panic("back buffer exhausted")
	}
	s.ipos--
	s.pos--

//
// It panics if count is negative or if the result of (len(b) * count)
// overflows.
func Repeat(b []byte, count int) []byte {
	if count == 0 {
		return []byte{}
	}

	// Since we cannot return an error on overflow,
	// we should panic if the repeat will generate an overflow.
	// See golang.org/issue/16237.
	if count < 0 {
		panic("bytes: negative Repeat count")
	}

		if debugPrint {
			fmt.Println("Mux", m.ID, "Handler took", time.Since(start).Round(time.Millisecond))
		}
		if r := recover(); r != nil {
			gridLogIf(ctx, fmt.Errorf("grid handler (%v) panic: %v", msg.Handler, r))
			err := RemoteErr(fmt.Sprintf("handler panic: %v", r))
			handlerErr = &err
		}
		if debugPrint {
			fmt.Println("muxServer: Mux", m.ID, "Returned with", handlerErr)
		}
	}()