intellij-idea - gRPC UnsupportedAddressTypeException，但仅当与 shadowJar 打包时

Question

我有一个简单的 gRPC 客户端，由 Gradle 使用 shadowJar 打包在一个 jar 中。当我运行 main()使用 IntelliJ，RPC 发送成功。当我用 java -jar 运行它时，我得到一个异常(exception):

更新:我想我已经确定 IntelliJ 和 ./gradlew shadowJar 都提供了同一组类，但顺序不同。我怀疑这会导致不同的行为，但我真的不明白为什么，或者在这两种情况下如何控制类路径的顺序。如果我能弄清楚哪些类在这里真正相关，那将是一个很大的帮助。完整扩展包括 19k 类文件。

Exception in thread "main" io.grpc.StatusException: UNKNOWN
        at io.grpc.Status.asException(Status.java:550)
        at io.grpc.kotlin.ClientCalls$rpcImpl$1$1$1.onClose(ClientCalls.kt:296)
        at io.grpc.internal.ClientCallImpl.closeObserver(ClientCallImpl.java:562)
        at io.grpc.internal.ClientCallImpl.access$300(ClientCallImpl.java:70)
        at io.grpc.internal.ClientCallImpl$ClientStreamListenerImpl$1StreamClosed.runInternal(ClientCallImpl.java:743)
        at io.grpc.internal.ClientCallImpl$ClientStreamListenerImpl$1StreamClosed.runInContext(ClientCallImpl.java:722)
        at io.grpc.internal.ContextRunnable.run(ContextRunnable.java:37)
        at io.grpc.internal.SerializingExecutor.run(SerializingExecutor.java:133)
        at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1136)
        at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:635)
        at java.base/java.lang.Thread.run(Thread.java:833)
Caused by: java.nio.channels.UnsupportedAddressTypeException
        at java.base/sun.nio.ch.Net.checkAddress(Net.java:146)
        at java.base/sun.nio.ch.Net.checkAddress(Net.java:157)
        at java.base/sun.nio.ch.SocketChannelImpl.checkRemote(SocketChannelImpl.java:816)
        at java.base/sun.nio.ch.SocketChannelImpl.connect(SocketChannelImpl.java:839)
        at io.netty.util.internal.SocketUtils$3.run(SocketUtils.java:91)
        at io.netty.util.internal.SocketUtils$3.run(SocketUtils.java:88)
        at java.base/java.security.AccessController.doPrivileged(AccessController.java:569)
        at io.netty.util.internal.SocketUtils.connect(SocketUtils.java:88)
        at io.netty.channel.socket.nio.NioSocketChannel.doConnect(NioSocketChannel.java:322)
        at io.netty.channel.nio.AbstractNioChannel$AbstractNioUnsafe.connect(AbstractNioChannel.java:248)
        at io.netty.channel.DefaultChannelPipeline$HeadContext.connect(DefaultChannelPipeline.java:1342)
        at io.netty.channel.AbstractChannelHandlerContext.invokeConnect(AbstractChannelHandlerContext.java:548)
        at io.netty.channel.AbstractChannelHandlerContext.connect(AbstractChannelHandlerContext.java:533)
        at io.netty.channel.ChannelDuplexHandler.connect(ChannelDuplexHandler.java:54)
        at io.grpc.netty.WriteBufferingAndExceptionHandler.connect(WriteBufferingAndExceptionHandler.java:157)
        at io.netty.channel.AbstractChannelHandlerContext.invokeConnect(AbstractChannelHandlerContext.java:548)
        at io.netty.channel.AbstractChannelHandlerContext.access$1000(AbstractChannelHandlerContext.java:61)
        at io.netty.channel.AbstractChannelHandlerContext$9.run(AbstractChannelHandlerContext.java:538)
        at io.netty.util.concurrent.AbstractEventExecutor.runTask(AbstractEventExecutor.java:174)
        at io.netty.util.concurrent.AbstractEventExecutor.safeExecute(AbstractEventExecutor.java:167)
        at io.netty.util.concurrent.SingleThreadEventExecutor.runAllTasks(SingleThreadEventExecutor.java:470)
        at io.netty.channel.nio.NioEventLoop.run(NioEventLoop.java:503)
        at io.netty.util.concurrent.SingleThreadEventExecutor$4.run(SingleThreadEventExecutor.java:995)
        at io.netty.util.internal.ThreadExecutorMap$2.run(ThreadExecutorMap.java:74)
        at io.netty.util.concurrent.FastThreadLocalRunnable.run(FastThreadLocalRunnable.java:30)
        ... 1 more

我的客户(这似乎不是问题所在):

fun main() {
    runBlocking {
        SampleClient().hello("hello")
    }
}
class SampleClient() : Closeable {
    private val channel = ManagedChannelBuilder.forAddress("localhost", StationController.PORT).usePlaintext().build()
    private val stub = StationServiceGrpcKt.StationServiceCoroutineStub(channel)
    suspend fun hello(msg: String) {
        val request = helloRequest { message = msg }
        stub.hello(request)
    }
    override fun close() {
        channel.shutdown().awaitTermination(5, TimeUnit.SECONDS)
    }
}

运行java -jar build/libs/test-0.1-all.jar因上述异常而失败。使用 java -classpath <very long classpath> com.test.MainKt 运行成功。

我试图最小化仍然显示此行为的 gradle.build.kts:

import org.jetbrains.kotlin.gradle.tasks.KotlinCompile
import com.google.protobuf.gradle.*

plugins {
    kotlin("jvm") version "1.7.10"
    id("com.github.johnrengelman.shadow") version "6.0.0"
    id("com.google.protobuf") version "0.8.19"
}

// These address this error:
// 'compileJava' task (current target is 17) and 'compileKotlin' task (current target is 1.8) jvm target compatibility
// should be set to the same Java version.
java { sourceCompatibility = JavaVersion.VERSION_17; targetCompatibility = JavaVersion.VERSION_17 }
tasks.withType<KotlinCompile> { kotlinOptions { jvmTarget = "17" } }

group = "com.test"
version = "0.1"

repositories { mavenCentral() }

dependencies {
    implementation("org.jetbrains.kotlinx:kotlinx-coroutines-core:1.6.3")
    implementation("io.grpc:grpc-protobuf:1.48.1")
    implementation("io.grpc:grpc-kotlin-stub:1.3.0")

    // The rpc fails only when I run it from the fat jar. running from intellij works. the stacktrace is all in netty,
    // but when I compile without netty there is no netty in the jar, making me think there is no transitive dependency
    // on netty. it fails with both netty and netty-shaded, but with different stack traces.
    // with netty-shaded, it says Caused by: io.grpc.netty.shaded.io.netty.channel.AbstractChannel$AnnotatedConnectException: connect(..) failed: Address family not supported by protocol: /localhost:8980
    // with netty, it just says Caused by: java.nio.channels.UnsupportedAddressTypeException
    implementation("io.grpc:grpc-netty:1.48.1")

    implementation("com.google.protobuf:protobuf-kotlin:3.21.4")
}

// for gRPC
protobuf {
    protoc { artifact = "com.google.protobuf:protoc:3.21.4" }
    plugins {
        // We only need this for Kotlin, but it doesn't fully link without the java sources.
        id("grpc") { artifact = "io.grpc:protoc-gen-grpc-java:1.48.1" }
        id("grpckt") { artifact = "io.grpc:protoc-gen-grpc-kotlin:1.3.0:jdk8@jar" }
    }
    generateProtoTasks {
        all().forEach {
            it.plugins { id("grpc"); id("grpckt") }
            // The plugins above (I think) generate the Java proto code, but not the extra Kotlin stuff like DSLs.
            // Including this adds a /kotlin folder to the generated code which contains all the extra stuff.
            it.builtins { id("kotlin") }
        }
    }
}


sourceSets {
    main {
        java {
            // These are the default generated source dirs, but they have to be added manually currently:
            // https://github.com/google/protobuf-gradle-plugin/issues/109
            // The plugin generates both a grpc dir with Java files, and grpckt with Kotlin files. I believe the grpckt
            // contents includes all the functionality of the grpc (java) contents, but I'm not positive. The kotlin
            // files might just be stubs for the java version? so maybe this compiles but doesn't run?
            srcDirs("build/generated/source/proto/main/grpc")
            srcDirs("build/generated/source/proto/main/grpckt")
            // This contains the Java classes like the request and response messages.
            srcDirs("build/generated/source/proto/main/java")
            // This contains the Kotlin extensions to those java classes, like DSLs.
            srcDirs("build/generated/source/proto/main/kotlin")
        }
    }
}

tasks.jar { manifest.attributes["Main-Class"] = "com.test.MainKt" }

我不知道如何继续调试它。

Answer 1

一位 friend 找到了答案。这个对 build.gradle.kts 的添加最终成为让它工作的全部:

tasks.named("shadowJar", ShadowJar::class) {
    mergeServiceFiles()
}

参见 shadowJar 的 documentation关于这里发生的事情。摘要:“Java 库通常在 JAR 的 META-INF/services 目录中包含服务描述 rune 件。服务描述符通常包含以行分隔的特定服务支持的类列表。在运行时，读取并使用此文件配置库或应用程序行为。

多个依赖项可以使用相同的服务描述 rune 件名。在这种情况下，通常希望将文件的每个实例的内容合并到一个输出文件中。” mergeServiceFiles() 就是这样做的。