Now setup as a driver and N anonymous workers that may become clients or servers.
Will convert async soon.

a proper join.

Also had been misusing EXPECT_EQ, as well as actually having an invalid
expectation on the ok field. Now it should be sane.

With this change the Python code generated by the Python code generator
uses the grpc.early_adopter package and not the grpc.framework.face
package.

This reverts commit 45b0bc4b.

This revert is being done because the compilers on Travis don't
understand the C++11 template/using syntax.

Also speed up a test, and disable some rarely touched but long running tests

To ensure that we actually probe different ports

bind and other functions

There may be some aspect of mortal sin in the way that context
management is now done in this test.

that it should be with multiple threads waiting on a single
completion queue.
The client currently uses a separate completion
queue per-thread, as trying to do a single unified queue was leading
to crashes for me. I need to figure that out.

Fixes module path finding in the Python code generator and the signatures
of generated servicer methods.

The test currently allocates a single port and reuses it through the
test. Given the timeouts in this test this leaves substantial race
windows where other processes on the same machine could steal the port
between subcases.

Instead, as a simple hack, allocate a new port before each test.

This string comes from an authority field, which is allowed to contain a
':' port (see https://tools.ietf.org/html/rfc3986#section-3.2).

We need to strip it before performing host name verification.

Dramatically lowers (eliminates maybe?) false negatives from ?SAN runs.