bpo-33530: Implement Happy Eyeballs in asyncio, v2 (GH-7237)

Added two keyword arguments, `delay` and `interleave`, to
`BaseEventLoop.create_connection`. Happy eyeballs is activated if
`delay` is specified.

We now have documentation for the new arguments. `staggered_race()` is in its own module, but not exported to the main asyncio package.


https://bugs.python.org/issue33530

Lib/asyncio/staggered.py

@@ -0,0 +1,147 @@
+"""Support for running coroutines in parallel with staggered start times."""
+
+__all__ = 'staggered_race',
+
+import contextlib
+import typing
+
+from . import events
+from . import futures
+from . import locks
+from . import tasks
+
+
+async def staggered_race(
+        coro_fns: typing.Iterable[typing.Callable[[], typing.Awaitable]],
+        delay: typing.Optional[float],
+        *,
+        loop: events.AbstractEventLoop = None,
+) -> typing.Tuple[
+    typing.Any,
+    typing.Optional[int],
+    typing.List[typing.Optional[Exception]]
+]:
+    """Run coroutines with staggered start times and take the first to finish.
+
+    This method takes an iterable of coroutine functions. The first one is
+    started immediately. From then on, whenever the immediately preceding one
+    fails (raises an exception), or when *delay* seconds has passed, the next
+    coroutine is started. This continues until one of the coroutines complete
+    successfully, in which case all others are cancelled, or until all
+    coroutines fail.
+
+    The coroutines provided should be well-behaved in the following way:
+
+    * They should only ``return`` if completed successfully.
+
+    * They should always raise an exception if they did not complete
+      successfully. In particular, if they handle cancellation, they should
+      probably reraise, like this::
+
+        try:
+            # do work
+        except asyncio.CancelledError:
+            # undo partially completed work
+            raise
+
+    Args:
+        coro_fns: an iterable of coroutine functions, i.e. callables that
+            return a coroutine object when called. Use ``functools.partial`` or
+            lambdas to pass arguments.
+
+        delay: amount of time, in seconds, between starting coroutines. If
+            ``None``, the coroutines will run sequentially.
+
+        loop: the event loop to use.
+
+    Returns:
+        tuple *(winner_result, winner_index, exceptions)* where
+
+        - *winner_result*: the result of the winning coroutine, or ``None``
+          if no coroutines won.
+
+        - *winner_index*: the index of the winning coroutine in
+          ``coro_fns``, or ``None`` if no coroutines won. If the winning
+          coroutine may return None on success, *winner_index* can be used
+          to definitively determine whether any coroutine won.
+
+        - *exceptions*: list of exceptions returned by the coroutines.
+          ``len(exceptions)`` is equal to the number of coroutines actually
+          started, and the order is the same as in ``coro_fns``. The winning
+          coroutine's entry is ``None``.
+
+    """
+    # TODO: when we have aiter() and anext(), allow async iterables in coro_fns.
+    loop = loop or events.get_running_loop()
+    enum_coro_fns = enumerate(coro_fns)
+    winner_result = None
+    winner_index = None
+    exceptions = []
+    running_tasks = []
+
+    async def run_one_coro(
+            previous_failed: typing.Optional[locks.Event]) -> None:
+        # Wait for the previous task to finish, or for delay seconds
+        if previous_failed is not None:
+            with contextlib.suppress(futures.TimeoutError):
+                # Use asyncio.wait_for() instead of asyncio.wait() here, so
+                # that if we get cancelled at this point, Event.wait() is also
+                # cancelled, otherwise there will be a "Task destroyed but it is
+                # pending" later.
+                await tasks.wait_for(previous_failed.wait(), delay)
+        # Get the next coroutine to run
+        try:
+            this_index, coro_fn = next(enum_coro_fns)
+        except StopIteration:
+            return
+        # Start task that will run the next coroutine
+        this_failed = locks.Event()
+        next_task = loop.create_task(run_one_coro(this_failed))
+        running_tasks.append(next_task)
+        assert len(running_tasks) == this_index + 2
+        # Prepare place to put this coroutine's exceptions if not won
+        exceptions.append(None)
+        assert len(exceptions) == this_index + 1
+
+        try:
+            result = await coro_fn()
+        except Exception as e:
+            exceptions[this_index] = e
+            this_failed.set()  # Kickstart the next coroutine
+        else:
+            # Store winner's results
+            nonlocal winner_index, winner_result
+            assert winner_index is None
+            winner_index = this_index
+            winner_result = result
+            # Cancel all other tasks. We take care to not cancel the current
+            # task as well. If we do so, then since there is no `await` after
+            # here and CancelledError are usually thrown at one, we will
+            # encounter a curious corner case where the current task will end
+            # up as done() == True, cancelled() == False, exception() ==
+            # asyncio.CancelledError. This behavior is specified in
+            # https://bugs.python.org/issue30048
+            for i, t in enumerate(running_tasks):
+                if i != this_index:
+                    t.cancel()
+
+    first_task = loop.create_task(run_one_coro(None))
+    running_tasks.append(first_task)
+    try:
+        # Wait for a growing list of tasks to all finish: poor man's version of
+        # curio's TaskGroup or trio's nursery
+        done_count = 0
+        while done_count != len(running_tasks):
+            done, _ = await tasks.wait(running_tasks)
+            done_count = len(done)
+            # If run_one_coro raises an unhandled exception, it's probably a
+            # programming error, and I want to see it.
+            if __debug__:
+                for d in done:
+                    if d.done() and not d.cancelled() and d.exception():
+                        raise d.exception()
+        return winner_result, winner_index, exceptions
+    finally:
+        # Make sure no tasks are left running if we leave this function
+        for t in running_tasks:
+            t.cancel()