Android真机安装失败 CMake报expected buildTargetsCommandComponents错误

开发带简单人脸识别功能的Android应用时，出现以下异常现象：

• 应用可在模拟器正常构建、运行
• 应用无法安装到真机设备
• Gradle执行CMake相关任务时抛出错误，错误信息如下：

Task :engine:configureCMakeDebug[arm64-v8a] FAILED
Execution failed for task ':engine:configureCMakeDebug[arm64-v8a]'.
[CXX1410] D:\KAM\code\AET\AET\engine.cxx\Debug\3p465u2i\arm64-v8a\android_gradle_build.json debug|arm64-v8a : expected buildTargetsCommandComponents or ncnn-release-arm64-v8a.buildCommandComponents to exist

CMakeLists.txt 配置 #

cmake_minimum_required(VERSION 3.4.1)

set(CMAKE_BUILD_TYPE "release")
set(CMAKE_LIBRARY_OUTPUT_DIRECTORY ${PROJECT_SOURCE_DIR}/../../../distribution/${ANDROID_ABI})

include_directories(${CMAKE_CURRENT_SOURCE_DIR}/include)

add_library(ncnn STATIC IMPORTED)
set_target_properties(ncnn PROPERTIES IMPORTED_LOCATION ${PROJECT_SOURCE_DIR}/../jniLibs/${ANDROID_ABI}/libncnn.a)

add_library(opencv_core STATIC IMPORTED)
set_target_properties(opencv_core PROPERTIES IMPORTED_LOCATION ${PROJECT_SOURCE_DIR}/../jniLibs/${ANDROID_ABI}/libopencv_core.so)

add_library(opencv_imgproc STATIC IMPORTED)
set_target_properties(opencv_imgproc PROPERTIES IMPORTED_LOCATION ${PROJECT_SOURCE_DIR}/../jniLibs/${ANDROID_ABI}/libopencv_imgproc.so)

add_library(opencv_codecs STATIC IMPORTED)
set_target_properties(opencv_codecs PROPERTIES IMPORTED_LOCATION ${PROJECT_SOURCE_DIR}/../jniLibs/${ANDROID_ABI}/libopencv_imgcodecs.so)


file(GLOB jni_srcs
        "*.cpp"
        "detection/*.cpp"
        "live/*.cpp"
    )

add_library(engine SHARED ${jni_srcs})

target_link_libraries(engine log jnigraphics ncnn opencv_core opencv_imgproc opencv_codecs android -static-openmp -fopenmp)

engine模块build.gradle配置 #

apply plugin: 'com.android.library'
apply plugin: 'kotlin-android'
apply plugin: 'kotlin-android-extensions'

android {
    compileSdkVersion 30
    buildToolsVersion "29.0.3"

    defaultConfig {
        minSdkVersion 28
        targetSdkVersion 30
        versionCode 1
        versionName "1.0"

        testInstrumentationRunner "androidx.test.runner.AndroidJUnitRunner"
        consumerProguardFiles 'consumer-rules.pro'

        externalNativeBuild {
            cmake {
                cppFlags "-std=c++11 -frtti -fexceptions"
                abiFilters 'arm64-v8a'
                arguments "-DANDROID_ARM_NEON=TRUE"
                arguments "-DANDROID_TOOLCHAIN=clang"
            }
        }
    }

    packagingOptions {
        pickFirst '**/*.so'
    }

    externalNativeBuild {
        cmake {
            version = "3.6.0"
            path "src/main/cpp/CMakeLists.txt"
        }
    }

    buildTypes {
        release {
            minifyEnabled true
            proguardFiles getDefaultProguardFile('proguard-android-optimize.txt'), 'proguard-rules.pro'
        }
    }
    ndkVersion '22.1.7171670'
}

dependencies {
    implementation fileTree(dir: 'libs', include: ['*.jar'])
    implementation "org.jetbrains.kotlin:kotlin-stdlib-jdk7:$kotlin_version"
    implementation 'androidx.appcompat:appcompat:1.2.0'
    implementation 'androidx.core:core-ktx:1.3.2'
    testImplementation 'junit:4.13'
    androidTestImplementation 'androidx.test.ext:junit:1.1.2'
    androidTestImplementation 'androidx.test.espresso:espresso-core:3.3.0'
}

该报错核心由构建缓存异常、库配置错误、构建规则冲突三类问题触发，按以下顺序排查即可：

1. 清理全量构建缓存
   先删除项目根目录下的.cxx、.gradle、build文件夹，同时删除engine模块下的build文件夹，执行Android Studio的Invalidate Caches / Restart后重新同步Gradle。超六成CXX1410类报错是旧缓存残留导致构建元数据解析异常引发的。另外模拟器通常为x86_64架构，当前配置仅过滤了arm64-v8aABI，模拟器能运行是调用了历史x86架构构建缓存，真机为arm64架构才会触发arm64-v8a的CMake配置任务报错，和“模拟器正常、真机无法安装”的现象完全对应。
2. 修正CMake导入库的类型声明错误
   当前配置存在明显类型不匹配问题：opencv_core、opencv_imgproc、opencv_codecs三个库的实际文件为.so动态库，但代码中用STATIC关键字声明为静态导入库，会导致Gradle生成构建元数据时找不到对应静态库的构建命令组件。将三个库的声明从add_library(xxx STATIC IMPORTED)修改为add_library(xxx SHARED IMPORTED)即可。
3. 移除CMake中硬编码的构建类型设置
   CMakeLists.txt中写死了set(CMAKE_BUILD_TYPE "release")，但当前执行的是configureCMakeDebug任务，构建类型冲突会导致Gradle生成debug变体构建配置时，去找硬编码指定的release变体下ncnn库的构建命令组件，自然匹配失败。直接删除这行硬编码配置，让Gradle根据当前构建变体自动传入对应CMAKE_BUILD_TYPE参数即可。
4. 校验预编译库路径与ABI匹配性
   检查src/main/jniLibs/arm64-v8a/目录下，是否确实存在声明的libncnn.a、三个opencv的so文件，避免因文件缺失导致配置生成失败。
5. 调整CMake版本适配
   当前指定的CMake版本为3.6.0，和NDK 22.1.7171670的默认适配版本不匹配，建议将CMake版本修改为3.18.1，和NDK r22兼容度最高，避免版本不匹配导致的元数据生成异常。

内容的提问来源于stack exchange，提问作者CastleMaster