Awesome

gdx-jnigen

The gdx-jnigen tool can be used with or without libGDX to allow C/C++ code to be written inline with Java source code.

This increases the locality of code that conceptually belongs together (the Java native class methods and the actual implementation) and makes refactoring a lot easier compared to the usual JNI workflow. Arrays and direct buffers are converted for you, further reducing boilerplate. Building the natives for Windows, Linux, macOS, and Android and iOS is handled for you. jnigen also provides a mechanism for loading native libraries from a JAR at runtime, which avoids "java.library.path" troubles.

See the libGDX Wiki for usage: https://libgdx.com/wiki/utils/jnigen

gdx-jnigen-gradle quickstart

We recommend you look at some existing projects for examples:

• gdx (Uses jnigen 2.x)
• gdx-freetype (Uses jnigen 2.x)
• gdx-bullet (Uses jnigen 2.x)
• gdx-video-desktop (Uses jnigen 2.x)
• Jamepad (Uses jnigen 2.x)

Configuring

// Apply jnigen plugin
plugins {
    id "com.badlogicgames.jnigen.jnigen-gradle"
}

// ...

// Define jnigen extension
jnigen {
    // Your shared library name
    sharedLibName = "example"
    
    //Enable multi threading compilation
    multiThreadedCompile = true

    // Shared configuration for all BuildTargets. Executed first
    // See all BuildTarget options here: https://github.com/libgdx/gdx-jnigen/blob/master/gdx-jnigen/src/main/java/com/badlogic/gdx/jnigen/BuildTarget.java
    // Most paths are relative to the $jniDir directory
    // String options can be replaced by using `x = "value"` or appended to with `x += "extravalue"`
    // String[] options can be replaced by using `x = ["value"]` or appended to with `x += "extravalue"` or `x += ["extravalue", "extravalue2"]`
    all {
        // Add extra flags passed to the C compiler
        cFlags += " -fvisibility=hidden "
        // Add extra flags passed to the C++ compiler
        cppFlags += " -std=c++11 -fvisibility=hidden "
        // Add extra flags passed to the linker
        linkerFlags += " -fvisibility=hidden "
    }

    // Configure robovm.xml for IOS builds, most simple libraries will not need to do this
    robovm {
        // Use preexisting robovm.xml, cannot be combined with other options.
        //manualFile file("robovm.xml")
        
        // Add extra patterns to forceLinkClasses
        //forceLinkClasses "test", "test2"
        //forceLinkClasses "pattern3"
        // Add extra library "test.a" with variant "device"
        //extraLib "test.a", "device"
        //extraXCFramework "libs/test.xcframework"
    }

    // Add BuildTargets
    // All BuildTarget options can be further customized in an OS+Arch specific manner within a {} block

    // Add windows 32-bit BuildTarget and customize it
    addWindows(x32, x86) {
        //cFlags += " -fextraflag=fake "
        //compilerPrefix = "someprefix-";
        //cIncludes += "windowsspecificdir/*.c"
    }
    
    //Add windows 64 bit, x86, MSVC toolchain 
    addWindows(x64, x86, MSVC) {
        msvcPreLinkerFlags += "/MD"
    }
    
    addWindows(x64, x86)    
    addLinux(x32, x86)
    addLinux(x64, x86)
    addLinux(x32, ARM)
    addLinux(x64, ARM)
    addMac(x64, x86)
    addMac(x64, ARM)
    
    //Auto add all possible ABIs 
    addAndroid() {
        // Add extra content to the generated Application.mk file
        //androidApplicationMk += ["APP_STL := c++_static"]
    }
    
    //Add specific android ABI
    addAndroid(AndroidABI.ABI_ARM64_V8A)
    
    //Auto add all possible iOS targets, including sim 
    addIOS() {
        // Define ios framework bundle identifier
        // xcframeworkBundleIdentifier = "com.badlogic.gdx.JniGen
        // Deinfe minimum supported iOS version
        // minIOSVersion = "11.0"
    }
 
    
    //Add specific ios device target
    addIOS(x64, ARM, DEVICE)
    
    //Add ios 64 bit x86 simulator target
    addIOS(x64, x86, SIMULATOR)


    // Customize each BuildTarget that matches the condition
    each({ it.os != Android && !it.isARM }) {
        //cppFlags += " -march=nocona "
    }
    // Customize everything again, can be used for conditional changes
    each({ true }) {
        //if(!it.cppCompiler.contains("clang")) {
        //    it.cFlags += " -flto "
        //    it.cppFlags += " -flto "
        //    it.linkerFlags += " -flto "
        //}

        //if(it.cppCompiler.contains("clang"))
        //    it.linkerFlags += " -Wl,-dead_strip -Wl,-s "
        //else
        //    it.linkerFlags += " -Wl,--gc-sections "
    }
}

Building

After the Gradle plugin is configured, jnigen will automatically generate and register appropriate tasks for building.

Generation

jnigen Required to be run before any other jnigen tasks, must be run any time source is changed or config is changed

e.g. ./gradlew jnigen

Compilation

jnigenBuildXXX Execute the native compilation for the targets you want. Replace XXX with the target name.

e.g. ./gradlew jnigenBuildAllWindows Build all windows targets

e.g. ./gradlew jnigenBuildLinuxArm_64 Build the 64-bit ARM Linux target only

Packaging

jnigenPackageXXX Package the native libraries into a jar. Replace XXX with the target name.

e.g. ./gradlew jnigenPackageAll Package all targets into their jars

e.g. ./gradlew jnigenPackageAllIOS Package the iOS targets into their jar

e.g. ./gradlew jnigenPackageAllAndroid Package the Android targets into their jars

Publishing

Jnigen automatically generates publishing tasks that integrate with the maven publishing plugin. If you are setup to publish to a maven repository, you can simply run:

./gradlew publish

Your artifacts will be published using groupid:artifactid-platform:version:classifierXXX as the coordinates.

replace classifierXXX with the target name.

e.g. natives-desktop, natives-ios, natives-android-arm64-v8a