35cb22dd42
|
||
|---|---|---|
| all | ||
| auth | ||
| benchmarks | ||
| buildscripts | ||
| compiler | ||
| core | ||
| examples | ||
| gradle/wrapper | ||
| integration-testing | ||
| lib | ||
| netty | ||
| okhttp | ||
| protobuf | ||
| protobuf-nano | ||
| stub | ||
| testing | ||
| .gitignore | ||
| .gitmodules | ||
| .travis.yml | ||
| AUTH-README.md | ||
| CONTRIBUTING.md | ||
| DEPLOYING.md | ||
| LICENSE | ||
| PATENTS | ||
| README.md | ||
| build.gradle | ||
| checkstyle.license | ||
| checkstyle.xml | ||
| gradlew | ||
| gradlew.bat | ||
| run-test-client.sh | ||
| run-test-server.sh | ||
| settings.gradle | ||
README.md
grpc-java
How to Build
Build Netty
grpc-java requires Netty 4.1, which is still in flux. The version we need can be found in the lib/netty submodule, which requires Maven 3.2 or higher to build:
$ git submodule update --init
$ cd lib/netty
$ mvn install -pl codec-http2 -am -DskipTests=true
Build Protobuf
The codegen plugin requires protobuf 3.0.0-alpha-2.
For Linux, Mac and MinGW:
$ git clone https://github.com/google/protobuf.git
$ cd protobuf
$ git checkout v3.0.0-alpha-2
$ ./autogen.sh
$ ./configure
$ make
$ make check
$ sudo make install
If you are comfortable with C++ compilation and autotools, you can specify a
--prefix for Protobuf and use -I in CXXFLAGS, -L in
LDFLAGS, LD_LIBRARY_PATH, and PATH to reference it. The
environment variables will be used when building grpc-java.
Protobuf installs to /usr/local by default.
For Visual C++, please refer to the Protobuf README for how to compile Protobuf.
Linux and MinGW
If /usr/local/lib is not in your library search path, you can add it by running:
$ sudo sh -c 'echo /usr/local/lib >> /etc/ld.so.conf'
$ sudo ldconfig
Mac
Some versions of Mac OS X (e.g., 10.10) doesn't have /usr/local in the
default search paths for header files and libraries. You will need to set
environment variables:
$ export CXXFLAGS="-I/usr/local/include" LDFLAGS="-L/usr/local/lib"
Build GRPC
On Linux, Mac or MinGW:
$ ./gradlew install
Notes for Visual C++
When building on Windows and VC++, you need to specify project properties for Gradle to find protobuf:
.\gradlew install ^
-Pvc.protobuf.include=C:\path\to\protobuf-3.0.0-alpha-2\src ^
-Pvc.protobuf.libs=C:\path\to\protobuf-3.0.0-alpha-2\vsprojects\Release
Since specifying those properties every build is bothersome, you can instead
create %HOMEDRIVE%%HOMEPATH%\.gradle\gradle.properties with contents like:
vc.protobuf.include=C:\\path\\to\\protobuf-3.0.0-alpha-2\\src
vc.protobuf.libs=C:\\path\\to\\protobuf-3.0.0-alpha-2\\vsprojects\\Release
The build script will build the codegen for the same architecture as the Java runtime installed on your system. If you are using 64-bit JVM, the codegen will be compiled for 64-bit, that means you must have compiled Protobuf in 64-bit.
Notes for MinGW on Windows
If you have both MinGW and VC++ installed on Windows, VC++ will be used by
default. To override this default and use MinGW, add -Dvc.disable to your
Gradle command line.
Navigating Around the Source
Heres a quick readers guide to the code to help folks get started. At a high level there are three distinct layers to the library: stub, channel & transport.
Stub
The 'stub' layer is what is exposed to most developers and provides type-safe bindings to whatever datamodel/IDL/interface you are adapting. An example is provided of a binding to code generated by the protocol-buffers compiler but others should be trivial to add and are welcome.
Key Interfaces
Channel
The 'channel' layer is an abstraction over transport handling that is suitable for interception/decoration and exposes more behavior to the application than the stub layer. It is intended to be easy for application frameworks to use this layer to address cross-cutting concerns such as logging, monitoring, auth etc. Flow-control is also exposed at this layer to allow more sophisticated applications to interact with it directly.
Common
Client
Server
Transport
The 'transport' layer does the heavy lifting of putting & taking bytes off the wire. The interfaces to it are abstract just enough to allow plugging in of different implementations. Transports are modeled as 'Stream' factories. The variation in interface between a server stream and a client stream exists to codify their differing semantics for cancellation and error reporting.
Common
Client
Server
Examples
Tests showing how these layers are composed to execute calls using protobuf messages can be found here https://github.com/google/grpc-java/tree/master/integration-testing/src/main/java/io/grpc/testing/integration