Many resources are needed to download a project. Please understand that we have to compensate our server costs. Thank you in advance. Project price only 1 $
You can buy this project and download/modify it how often you want.
/*
* Copyright 2014 The gRPC Authors
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package io.grpc.testing.integration;
import com.google.common.base.Preconditions;
import com.google.common.collect.Queues;
import com.google.protobuf.ByteString;
import io.grpc.ForwardingServerCall.SimpleForwardingServerCall;
import io.grpc.Metadata;
import io.grpc.ServerCall;
import io.grpc.ServerCallHandler;
import io.grpc.ServerInterceptor;
import io.grpc.Status;
import io.grpc.internal.LogExceptionRunnable;
import io.grpc.stub.ServerCallStreamObserver;
import io.grpc.stub.StreamObserver;
import io.grpc.testing.integration.Messages.Payload;
import io.grpc.testing.integration.Messages.ResponseParameters;
import io.grpc.testing.integration.Messages.SimpleRequest;
import io.grpc.testing.integration.Messages.SimpleResponse;
import io.grpc.testing.integration.Messages.StreamingInputCallRequest;
import io.grpc.testing.integration.Messages.StreamingInputCallResponse;
import io.grpc.testing.integration.Messages.StreamingOutputCallRequest;
import io.grpc.testing.integration.Messages.StreamingOutputCallResponse;
import java.util.ArrayDeque;
import java.util.Arrays;
import java.util.HashSet;
import java.util.List;
import java.util.Queue;
import java.util.Random;
import java.util.Set;
import java.util.concurrent.Future;
import java.util.concurrent.ScheduledExecutorService;
import java.util.concurrent.TimeUnit;
import javax.annotation.concurrent.GuardedBy;
/**
* Implementation of the business logic for the TestService. Uses an executor to schedule chunks
* sent in response streams.
*/
public class TestServiceImpl extends TestServiceGrpc.TestServiceImplBase {
private final Random random = new Random();
private final ScheduledExecutorService executor;
private final ByteString compressableBuffer;
/**
* Constructs a controller using the given executor for scheduling response stream chunks.
*/
public TestServiceImpl(ScheduledExecutorService executor) {
this.executor = executor;
this.compressableBuffer = ByteString.copyFrom(new byte[1024]);
}
@Override
public void emptyCall(EmptyProtos.Empty request,
StreamObserver responseObserver) {
responseObserver.onNext(EmptyProtos.Empty.getDefaultInstance());
responseObserver.onCompleted();
}
/**
* Immediately responds with a payload of the type and size specified in the request.
*/
@Override
public void unaryCall(SimpleRequest req, StreamObserver responseObserver) {
ServerCallStreamObserver obs =
(ServerCallStreamObserver) responseObserver;
SimpleResponse.Builder responseBuilder = SimpleResponse.newBuilder();
try {
if (req.hasResponseCompressed() && req.getResponseCompressed().getValue()) {
obs.setCompression("gzip");
} else {
obs.setCompression("identity");
}
} catch (IllegalArgumentException e) {
obs.onError(Status.UNIMPLEMENTED
.withDescription("compression not supported.")
.withCause(e)
.asRuntimeException());
return;
}
if (req.getResponseSize() != 0) {
// For consistency with the c++ TestServiceImpl, use a random offset for unary calls.
// TODO(wonderfly): whether or not this is a good approach needs further discussion.
int offset = random.nextInt(compressableBuffer.size());
ByteString payload = generatePayload(compressableBuffer, offset, req.getResponseSize());
responseBuilder.setPayload(
Payload.newBuilder()
.setBody(payload));
}
if (req.hasResponseStatus()) {
obs.onError(Status.fromCodeValue(req.getResponseStatus().getCode())
.withDescription(req.getResponseStatus().getMessage())
.asRuntimeException());
return;
}
responseObserver.onNext(responseBuilder.build());
responseObserver.onCompleted();
}
/**
* Given a request that specifies chunk size and interval between responses, creates and schedules
* the response stream.
*/
@Override
public void streamingOutputCall(StreamingOutputCallRequest request,
StreamObserver responseObserver) {
// Create and start the response dispatcher.
new ResponseDispatcher(responseObserver).enqueue(toChunkQueue(request)).completeInput();
}
/**
* Waits until we have received all of the request messages and then returns the aggregate payload
* size for all of the received requests.
*/
@Override
public StreamObserver streamingInputCall(
final StreamObserver responseObserver) {
return new StreamObserver() {
private int totalPayloadSize;
@Override
public void onNext(StreamingInputCallRequest message) {
totalPayloadSize += message.getPayload().getBody().size();
}
@Override
public void onCompleted() {
responseObserver.onNext(StreamingInputCallResponse.newBuilder()
.setAggregatedPayloadSize(totalPayloadSize).build());
responseObserver.onCompleted();
}
@Override
public void onError(Throwable cause) {
responseObserver.onError(cause);
}
};
}
/**
* True bi-directional streaming. Processes requests as they come in. Begins streaming results
* immediately.
*/
@Override
public StreamObserver fullDuplexCall(
final StreamObserver responseObserver) {
final ResponseDispatcher dispatcher = new ResponseDispatcher(responseObserver);
return new StreamObserver() {
@Override
public void onNext(StreamingOutputCallRequest request) {
if (request.hasResponseStatus()) {
dispatcher.cancel();
dispatcher.onError(Status.fromCodeValue(request.getResponseStatus().getCode())
.withDescription(request.getResponseStatus().getMessage())
.asRuntimeException());
return;
}
dispatcher.enqueue(toChunkQueue(request));
}
@Override
public void onCompleted() {
if (!dispatcher.isCancelled()) {
// Tell the dispatcher that all input has been received.
dispatcher.completeInput();
}
}
@Override
public void onError(Throwable cause) {
dispatcher.onError(cause);
}
};
}
/**
* Similar to {@link #fullDuplexCall}, except that it waits for all streaming requests to be
* received before starting the streaming responses.
*/
@Override
public StreamObserver halfDuplexCall(
final StreamObserver responseObserver) {
final ResponseDispatcher dispatcher = new ResponseDispatcher(responseObserver);
final Queue chunks = new ArrayDeque<>();
return new StreamObserver() {
@Override
public void onNext(StreamingOutputCallRequest request) {
chunks.addAll(toChunkQueue(request));
}
@Override
public void onCompleted() {
// Dispatch all of the chunks in one shot.
dispatcher.enqueue(chunks).completeInput();
}
@Override
public void onError(Throwable cause) {
dispatcher.onError(cause);
}
};
}
/**
* Schedules the dispatch of a queue of chunks. Whenever chunks are added or input is completed,
* the next response chunk is scheduled for delivery to the client. When no more chunks are
* available, the stream is half-closed.
*/
private class ResponseDispatcher {
private final Chunk completionChunk = new Chunk(0, 0, 0);
private final Queue chunks;
private final StreamObserver responseStream;
private boolean scheduled;
@GuardedBy("this") private boolean cancelled;
private Throwable failure;
private Runnable dispatchTask = new Runnable() {
@Override
@SuppressWarnings("CatchAndPrintStackTrace")
public void run() {
try {
// Dispatch the current chunk to the client.
try {
dispatchChunk();
} catch (RuntimeException e) {
// Indicate that nothing is scheduled and re-throw.
synchronized (ResponseDispatcher.this) {
scheduled = false;
}
throw e;
}
// Schedule the next chunk if there is one.
synchronized (ResponseDispatcher.this) {
// Indicate that nothing is scheduled.
scheduled = false;
scheduleNextChunk();
}
} catch (Throwable t) {
t.printStackTrace();
}
}
};
/**
* The {@link StreamObserver} will be used to send the queue of response chunks. Since calls to
* {@link StreamObserver} must be synchronized across threads, no further calls should be made
* directly on {@code responseStream} after it is provided to the {@link ResponseDispatcher}.
*/
public ResponseDispatcher(StreamObserver responseStream) {
this.chunks = Queues.newLinkedBlockingQueue();
this.responseStream = responseStream;
}
/**
* Adds the given chunks to the response stream and schedules the next chunk to be delivered if
* needed.
*/
public synchronized ResponseDispatcher enqueue(Queue moreChunks) {
assertNotFailed();
chunks.addAll(moreChunks);
scheduleNextChunk();
return this;
}
/**
* Indicates that the input is completed and the currently enqueued response chunks are all that
* remain to be scheduled for dispatch to the client.
*/
public ResponseDispatcher completeInput() {
assertNotFailed();
chunks.add(completionChunk);
scheduleNextChunk();
return this;
}
/**
* Allows the service to cancel the remaining responses.
*/
public synchronized void cancel() {
Preconditions.checkState(!cancelled, "Dispatcher already cancelled");
chunks.clear();
cancelled = true;
}
public synchronized boolean isCancelled() {
return cancelled;
}
private synchronized void onError(Throwable cause) {
responseStream.onError(cause);
}
/**
* Dispatches the current response chunk to the client. This is only called by the executor. At
* any time, a given dispatch task should only be registered with the executor once.
*/
private synchronized void dispatchChunk() {
if (cancelled) {
return;
}
try {
// Pop off the next chunk and send it to the client.
Chunk chunk = chunks.remove();
if (chunk == completionChunk) {
responseStream.onCompleted();
} else {
responseStream.onNext(chunk.toResponse());
}
} catch (Throwable e) {
failure = e;
if (Status.fromThrowable(e).getCode() == Status.CANCELLED.getCode()) {
// Stream was cancelled by client, responseStream.onError() might be called already or
// will be called soon by inbounding StreamObserver.
chunks.clear();
} else {
responseStream.onError(e);
}
}
}
/**
* Schedules the next response chunk to be dispatched. If all input has been received and there
* are no more chunks in the queue, the stream is closed.
*/
private void scheduleNextChunk() {
synchronized (this) {
if (scheduled) {
// Dispatch task is already scheduled.
return;
}
// Schedule the next response chunk if there is one.
Chunk nextChunk = chunks.peek();
if (nextChunk != null) {
scheduled = true;
// TODO(ejona): cancel future if RPC is cancelled
Future> unused = executor.schedule(new LogExceptionRunnable(dispatchTask),
nextChunk.delayMicroseconds, TimeUnit.MICROSECONDS);
return;
}
}
}
private void assertNotFailed() {
if (failure != null) {
throw new IllegalStateException("Stream already failed", failure);
}
}
}
/**
* Breaks down the request and creates a queue of response chunks for the given request.
*/
public Queue toChunkQueue(StreamingOutputCallRequest request) {
Queue chunkQueue = new ArrayDeque<>();
int offset = 0;
for (ResponseParameters params : request.getResponseParametersList()) {
chunkQueue.add(new Chunk(params.getIntervalUs(), offset, params.getSize()));
// Increment the offset past this chunk. Buffer need to be circular.
offset = (offset + params.getSize()) % compressableBuffer.size();
}
return chunkQueue;
}
/**
* A single chunk of a response stream. Contains delivery information for the dispatcher and can
* be converted to a streaming response proto. A chunk just references it's payload in the
* {@link #compressableBuffer} array. The payload isn't actually created until {@link
* #toResponse()} is called.
*/
private class Chunk {
private final int delayMicroseconds;
private final int offset;
private final int length;
public Chunk(int delayMicroseconds, int offset, int length) {
this.delayMicroseconds = delayMicroseconds;
this.offset = offset;
this.length = length;
}
/**
* Convert this chunk into a streaming response proto.
*/
private StreamingOutputCallResponse toResponse() {
StreamingOutputCallResponse.Builder responseBuilder =
StreamingOutputCallResponse.newBuilder();
ByteString payload = generatePayload(compressableBuffer, offset, length);
responseBuilder.setPayload(
Payload.newBuilder()
.setBody(payload));
return responseBuilder.build();
}
}
/**
* Generates a payload of desired type and size. Reads compressableBuffer or
* uncompressableBuffer as a circular buffer.
*/
private ByteString generatePayload(ByteString dataBuffer, int offset, int size) {
ByteString payload = ByteString.EMPTY;
// This offset would never pass the array boundary.
int begin = offset;
int end = 0;
int bytesLeft = size;
while (bytesLeft > 0) {
end = Math.min(begin + bytesLeft, dataBuffer.size());
// ByteString.substring returns the substring from begin, inclusive, to end, exclusive.
payload = payload.concat(dataBuffer.substring(begin, end));
bytesLeft -= (end - begin);
begin = end % dataBuffer.size();
}
return payload;
}
/** Returns interceptors necessary for full service implementation. */
public static List interceptors() {
return Arrays.asList(
echoRequestHeadersInterceptor(Util.METADATA_KEY),
echoRequestMetadataInHeaders(Util.ECHO_INITIAL_METADATA_KEY),
echoRequestMetadataInTrailers(Util.ECHO_TRAILING_METADATA_KEY));
}
/**
* Echo the request headers from a client into response headers and trailers. Useful for
* testing end-to-end metadata propagation.
*/
private static ServerInterceptor echoRequestHeadersInterceptor(final Metadata.Key>... keys) {
final Set> keySet = new HashSet<>(Arrays.asList(keys));
return new ServerInterceptor() {
@Override
public ServerCall.Listener interceptCall(
ServerCall call,
final Metadata requestHeaders,
ServerCallHandler next) {
return next.startCall(new SimpleForwardingServerCall(call) {
@Override
public void sendHeaders(Metadata responseHeaders) {
responseHeaders.merge(requestHeaders, keySet);
super.sendHeaders(responseHeaders);
}
@Override
public void close(Status status, Metadata trailers) {
trailers.merge(requestHeaders, keySet);
super.close(status, trailers);
}
}, requestHeaders);
}
};
}
/**
* Echoes request headers with the specified key(s) from a client into response headers only.
*/
private static ServerInterceptor echoRequestMetadataInHeaders(final Metadata.Key>... keys) {
final Set> keySet = new HashSet<>(Arrays.asList(keys));
return new ServerInterceptor() {
@Override
public ServerCall.Listener interceptCall(
ServerCall call,
final Metadata requestHeaders,
ServerCallHandler next) {
return next.startCall(new SimpleForwardingServerCall(call) {
@Override
public void sendHeaders(Metadata responseHeaders) {
responseHeaders.merge(requestHeaders, keySet);
super.sendHeaders(responseHeaders);
}
@Override
public void close(Status status, Metadata trailers) {
super.close(status, trailers);
}
}, requestHeaders);
}
};
}
/**
* Echoes request headers with the specified key(s) from a client into response trailers only.
*/
private static ServerInterceptor echoRequestMetadataInTrailers(final Metadata.Key>... keys) {
final Set> keySet = new HashSet<>(Arrays.asList(keys));
return new ServerInterceptor() {
@Override
public ServerCall.Listener interceptCall(
ServerCall call,
final Metadata requestHeaders,
ServerCallHandler next) {
return next.startCall(new SimpleForwardingServerCall(call) {
@Override
public void sendHeaders(Metadata responseHeaders) {
super.sendHeaders(responseHeaders);
}
@Override
public void close(Status status, Metadata trailers) {
trailers.merge(requestHeaders, keySet);
super.close(status, trailers);
}
}, requestHeaders);
}
};
}
}