In _server.py start_server_batch_result is removed because
start_server_batch can only ever fail as a result of a programming
defect in gRPC Python and not the application. This differs from some
analogous-appearing points in _channel.py where the result of
start_client_batch is checked because at those points it is possible
for a failure to indicate a programming defect on the part of the
application.

This should have been removed in, and has done anything but fail on
import since, 4d425522.

client now conforms to spec. See
https://github.com/grpc/grpc/blob/master/doc/negative-http2-interop-test-descriptions.md

This should have been included in 48226a2f.

Useful for situations where we need to repeatedly trylock, not useful
for cases where we need to lock (due to spinning).

Add a variant of sync_test to test it (with the same tests we run for
gpr_mu).

Add a benchmark to bm_closure to demonstrate single threaded performance.

Based on soltanmm-google's #9114