Besides API changes, this implementation is also up-to-date with the
latest design:
1. Delegate to round-robin and pick-first policies if requested by
the naming system.
2. OOB channels to LoadBalancer always use the LB authority provided by
the naming system.
3. Never send application RPCs to balancer addresses, even if the
address returns UNIMPLEMENTED error.
not necessary to synchronze every time calling
getServiceDescriptor(), if the descriptor has been created already;
go with the double-checked locking idom
"Because of spurious wakeups, wait() must always be called in a loop".
Got rid of wait().
"If you return or throw from a finally, then values returned or thrown from the try-catch block will be ignored. Consider using try-with-resources instead."
Ignored the exception thrown from finally.
These tests are for the new HTTP/2 test server (grpc/grpc#8900). They are designed to test client behavior when the server sends unexpected responses during rpcs.
TransportSet mandates that transportShutdown() has been called prior to
transportTerminated(). OkHttp transport tries to meet this requirement
by calling startGoAway() which calls transportShutdown() in both the
normal path and the error path. However, the error path only caught
Exception. If an Error is thrown from the try-block, calling
transportTerminated() in the finally block will cause a check failure in
TransportSet, shadowing the original Error, which is undesirable for
debugging. Catching Throwable here is more helpful.
Not to expose dependency of `io.netty.handler.ssl.SslContext` when implementing `TransportCreationParamsFilter`
Change the `TransportCreationParamsFilter` API
````
ProtocolNegotiator getProtocolNegotiator(NegotiationType negotiationType, SslContext sslContext);
````
into
````
ProtocolNegotiator getProtocolNegotiator();
````
1. Use ChannelExecutor to run callbacks. Now callbacks are no longer
run under the delayed transport lock.
2. Use explicit picker version instead of relying on identity equality
to detect new pickers. Rationale: if reprocess() is called again with
the same picker, all pending streams will be reprocessed by this picker
again, thus there is no reason to leave out the racing new stream.
* Fork DelayedClientTransport into DelayedClientTransport2, and fix a race
on it. Consider the following sequence:
1. Channel is created. Picker is initially null
2. Channel has a new RPC. Because picker is null, the delayed transport
is picked, but newStream() is not called yet.
3. LoadBalancer updates a new picker to Channel. Channel runs
reprocess() but does nothing because no pending stream is there.
4. newStream() called on the delayed transport.
In previous implementation, newStream() on step 4 will not see the
picker. It will only use the next picker.
After this change, delayed transport would save the latest picker and
use it on newStream(), which is the desired behavior.
Also deleted all the code that will not be used after the LB refactor.
* Also fixed a bug: newStream() should always return a failing stream if it's shutdown. Previously it's not doing so if picker is not null.
GrpcServerRule configures an in-process server and channel. It is
useful for asserting requests being made to a service. A consumer can
create a mock implementation of their service that records each
request, then make assertions on those records in their test.
OutputStreamAdapter is a private class, and is only ever called by two
places: ByteStreams.copy, which never calls the single byte method, and
DrainTo, which potentially can. There are two classes that implement
DrainTo, which is primarily ProtoInputStream. It calls
MessageLite.writeTo(OutputStream) or back again to ByteStreams.copy.
MessageLite.writeTo in turn wraps the OutputStream in a
CodedOutputStream.OutputStreamEncoder, which then never calls the single
byte version. Thus, all know implementations never call the single
byte override.
Additionally, it is well known that the single byte write is slow, and
is expected to be wrapped in a BufferedOutputStream if there are many
small writes.
Currently Context propagate in-thread by its own ThreadLocal, and
cross-thread propagation must be done with the mechanisms provied by
gRPC Context. However, there are frameworks (e.g. what we are using
inside Google) which have already established context-propagation
mechanisms. If gRPC Context may ride on top of them, it would be
propagated automatically without additional effort from the application.
The Storage API allows gRPC Context to be attached to anything. The
default implementation still uses its own ThreadLocal. If an override
implementation is present, gRPC Context will use it.
Kun Zhang
ZHANG Dapeng
Eric Anderson
Eric Gribkoff
Łukasz Strzałkowski
Carl Mastrangelo
Anuraag Agrawal
Lukasz Strzalkowski
Kenji Kaneda
Joey Bratton
Kailash Sethuraman
Michael Avrukin