@SuppressWarnings("ThreadPriorityCheck") // doing our best to test for races
          public Boolean get() {
            // Check that this method is called exactly once, by the first
            // thread to synchronize.
            long t0 = System.nanoTime();
            while (waitingThreads() != numThreads - 1) {
              if (System.nanoTime() - t0 > timeout) {
                thrown.set(

        assertEquals(0x00020000, SmbConstants.READ_CONTROL);
        assertEquals(0x00040000, SmbConstants.WRITE_DAC);
        assertEquals(0x00080000, SmbConstants.WRITE_OWNER);
        assertEquals(0x00100000, SmbConstants.SYNCHRONIZE);
        assertEquals(0x10000000, SmbConstants.GENERIC_ALL);
        assertEquals(0x20000000, SmbConstants.GENERIC_EXECUTE);
        assertEquals(0x40000000, SmbConstants.GENERIC_WRITE);

Voici un petit bout de cette histoire.

## Alternatives

Je crée des API avec des exigences complexes depuis plusieurs années (Machine Learning, systèmes distribués, jobs asynchrones, bases de données NoSQL, etc), en dirigeant plusieurs équipes de développeurs.

Dans ce cadre, j'ai dû étudier, tester et utiliser de nombreuses alternatives.

### Locks

We have 3 different things that we synchronize on.

#### Http2Connection

        }
    }

    @Override
    public void run() {
        final Thread run_thread = Thread.currentThread();
        Exception ex0 = null;

        try {
            /* We cannot synchronize (run_thread) here or the caller's
             * thread.wait( timeout ) cannot reaquire the lock and
             * return which would render the timeout effectively useless.
             */
            doConnect();

				success := countErrs(errs, nil)
				if success < len(disks)/2+1 {
					retErr = fmt.Errorf("one or more errors reported during listing: %v", errors.Join(errs...))
				}
			},
		})
		jt.Wait() // synchronize all the concurrent heal jobs
		if err != nil {
			// Set this such that when we return this function
			// we let the caller retry this disk again for the
			// buckets it failed to list.
			retErr = err
		}

            0x00010000, // DELETE
            0x00020000, // READ_CONTROL
            0x00040000, // WRITE_DAC
            0x00080000, // WRITE_OWNER
            0x00100000, // SYNCHRONIZE
            0x10000000, // GENERIC_ALL
            0x20000000, // GENERIC_EXECUTE
            0x40000000, // GENERIC_WRITE
            0x80000000 // GENERIC_READ (as int will be negative)
    })

  //
  // Socket reads are unguarded but are only made by the reader thread.
  //
  // Certain operations (like SYN_STREAM) need to synchronize on both the frameWriter (to do
  // blocking I/O) and this (to create streams). Such operations must synchronize on 'this' last.
  // This ensures that we never wait for a blocking operation while holding 'this'.

  /** True if this peer initiated the connection. */

   * serial access, it is critical that <b>all</b> access to the backing table is accomplished
   * through the returned table.
   *
   * <p>It is imperative that the user manually synchronize on the returned table when accessing any
   * of its collection views:
   *
   * {@snippet :
   * Table<R, C, V> table = Tables.synchronizedTable(HashBasedTable.<R, C, V>create());
   * ...

## Installation

<div class="termy">

```console
$ pip install fastapi

---> 100%
```

</div>