}

    /**
     * The version of this test supplied by {@link MapInterfaceTest} fails for this particular map
     * implementation, because {@code map.get()} returns a view collection that changes in the
     * course of a call to {@code remove()}. Thus, the expectation doesn't hold that {@code
     * map.remove(x)} returns the same value which {@code map.get(x)} did immediately beforehand.
     */
    @Override

    }

    /**
     * The version of this test supplied by {@link MapInterfaceTest} fails for this particular map
     * implementation, because {@code map.get()} returns a view collection that changes in the
     * course of a call to {@code remove()}. Thus, the expectation doesn't hold that {@code
     * map.remove(x)} returns the same value which {@code map.get(x)} did immediately beforehand.
     */
    @Override

It would depend mainly on the tool you use to **install** those requirements.

The most common way to do it is to have a file `requirements.txt` with the package names and their versions, one per line.

You would of course use the same ideas you read in [About FastAPI versions](versions.md){.internal-link target=_blank} to set the ranges of versions.

For example, your `requirements.txt` could look like:

```

    checkState(xSumOfSquaresOfDeltas > 0.0);
    checkState(ySumOfSquaresOfDeltas > 0.0);
    // The product of two positive numbers can be zero if the multiplication underflowed. We
    // force a positive value by effectively rounding up to MIN_VALUE.
    double productOfSumsOfSquaresOfDeltas =
        ensurePositive(xSumOfSquaresOfDeltas * ySumOfSquaresOfDeltas);

		switch {
		case c == '"':
			return string(d.scratch.bytes()), nil
		case c == '\\':
			c = d.next()
			goto scan_esc
		case c < 0x20:
			return "", d.mkError(ErrSyntax, "in string literal")
		// Coerce to well-formed UTF-8.
		default:
			d.scratch.add(c)
			if d.remaining() == 0 {
				return "", d.mkError(ErrSyntax, "in string literal")
			}
			c = d.next()
		}
	}

scan_esc:
	switch c {

            },
            directExecutor()));
    thingToCancel[0] = serializer.submit(Callables.returning(null), directExecutor());
    results.add(thingToCancel[0]);
    // Enqueue more than enough tasks to force reentrancy.
    for (int i = 0; i < 5; i++) {
      results.add(serializer.submit(Callables.returning(null), directExecutor()));
    }

    manualExecutorTask[0].run();

    for (Future<?> result : results) {

        });
        indexUpdater.start();

        int startedCrawlerNum = 0;
        int activeCrawlerNum = 0;
        try {
            while (startedCrawlerNum < crawlerList.size()) {
                // Force to stop crawl
                if (systemHelper.isForceStop()) {
                    for (final Crawler crawler : crawlerList) {
                        crawler.stop();
                    }
                    break;

 * </ul>
 *
 * <p>Disadvantages compared to {@code long[]}:
 *
 * <ul>
 *   <li>Memory footprint has a fixed overhead (about 24 bytes per instance).
 *   <li><i>Some</i> construction use cases force the data to be copied (though several construction
 *       APIs are offered that don't).
 *   <li>Can't be passed directly to methods that expect {@code long[]} (though the most common
 *       utilities do have replacements here).

 * </ul>
 *
 * <p>Disadvantages compared to {@code int[]}:
 *
 * <ul>
 *   <li>Memory footprint has a fixed overhead (about 24 bytes per instance).
 *   <li><i>Some</i> construction use cases force the data to be copied (though several construction
 *       APIs are offered that don't).
 *   <li>Can't be passed directly to methods that expect {@code int[]} (though the most common
 *       utilities do have replacements here).

C'est pourquoi il est logique d'utiliser du code asynchrone ⏸🔀⏯ pour des APIs web.

Ce type d'asynchronicité est ce qui a rendu NodeJS populaire (bien que NodeJS ne soit pas parallèle) et c'est la force du Go en tant que langage de programmation.

Et c'est le même niveau de performance que celui obtenu avec **FastAPI**.