We could not demonstrate via actual benchmarking that the tick reduction
from awaiting a promise prior to returning it translates into real-world
performance gains. See graphql#3796.

In any case, the promise must be awaited (irrespective of performance)
to allow the catch block to handle rejection errors. Simply returning
`completed` would result in test failures.

We could not demonstrate via actual benchmarking that the tick reduction
from awaiting a promise prior to returning it translates into real-world
performance gains. See graphql#3796.

In any case, the promise must be awaited (irrespective of performance)
to allow the catch block to handle rejection errors. Simply returning
`completed` would result in test failures.

We could not demonstrate via actual benchmarking that the tick reduction
from awaiting a promise prior to returning it translates into real-world
performance gains. See graphql#3796.

In any case, the promise must be awaited (irrespective of performance)
to allow the catch block to handle rejection errors. Simply returning
`completed` would result in test failures.

We could not demonstrate via actual benchmarking that the tick reduction
from awaiting a promise prior to returning it translates into real-world
performance gains. See graphql#3796.

In any case, the promise must be awaited (irrespective of performance)
to allow the catch block to handle rejection errors. Simply returning
`completed` would result in test failures.

We could not demonstrate via actual benchmarking that the tick reduction
from awaiting a promise prior to returning it translates into real-world
performance gains. See graphql#3796.

In any case, the promise must be awaited (irrespective of performance)
to allow the catch block to handle rejection errors. Simply returning
`completed` would result in test failures.

We could not demonstrate via actual benchmarking that the tick reduction
from awaiting a promise prior to returning it translates into real-world
performance gains. See graphql#3796.

In any case, the promise must be awaited (irrespective of performance)
to allow the catch block to handle rejection errors. Simply returning
`completed` would result in test failures.

We could not demonstrate via actual benchmarking that the tick reduction
from awaiting a promise prior to returning it translates into real-world
performance gains. See #3796.

In any case, the promise must be awaited (irrespective of performance)
to allow the catch block to handle rejection errors. Simply returning
`completed` would result in test failures.

We could not demonstrate via actual benchmarking that the tick reduction
from awaiting a promise prior to returning it translates into real-world
performance gains. See #3796.

In any case, the promise must be awaited (irrespective of performance)
to allow the catch block to handle rejection errors. Simply returning
`completed` would result in test failures.

We could not demonstrate via actual benchmarking that the tick reduction
from awaiting a promise prior to returning it translates into real-world
performance gains. See graphql#3796.

In any case, the promise must be awaited (irrespective of performance)
to allow the catch block to handle rejection errors. Simply returning
`completed` would result in test failures.

We could not demonstrate via actual benchmarking that the tick reduction
from awaiting a promise prior to returning it translates into real-world
performance gains. See #3796.

In any case, the promise must be awaited (irrespective of performance)
to allow the catch block to handle rejection errors. Simply returning
`completed` would result in test failures.

We could not demonstrate via actual benchmarking that the tick reduction
from awaiting a promise prior to returning it translates into real-world
performance gains. See #3796.

In any case, the promise must be awaited (irrespective of performance)
to allow the catch block to handle rejection errors. Simply returning
`completed` would result in test failures.

We could not demonstrate via actual benchmarking that the tick reduction
from awaiting a promise prior to returning it translates into real-world
performance gains. See #3796.

In any case, the promise must be awaited (irrespective of performance)
to allow the catch block to handle rejection errors. Simply returning
`completed` would result in test failures.

We could not demonstrate via actual benchmarking that the tick reduction
from awaiting a promise prior to returning it translates into real-world
performance gains. See #3796.

In any case, the promise must be awaited (irrespective of performance)
to allow the catch block to handle rejection errors. Simply returning
`completed` would result in test failures.

We could not demonstrate via actual benchmarking that the tick reduction
from awaiting a promise prior to returning it translates into real-world
performance gains. See #3796.

In any case, the promise must be awaited (irrespective of performance)
to allow the catch block to handle rejection errors. Simply returning
`completed` would result in test failures.

We could not demonstrate via actual benchmarking that the tick reduction
from awaiting a promise prior to returning it translates into real-world
performance gains. See #3796.

In any case, the promise must be awaited (irrespective of performance)
to allow the catch block to handle rejection errors. Simply returning
`completed` would result in test failures.