Question

我是新手JNA开发人员所以如果这个问题很简单，请原谅我，但是我正在努力反对处理C-lang结构对象，包括带有JNA的float数组。目前我遇到了编译错误。

我认为问题是初始化并将FloatByReference实例（mys.Weights）分配给mystruct类实例（mys）的错误方法，并且认为不应修改下面的Java代码，因为JNAerator会自动生成它

我的C代码：

typedef struct
{
  int n;
  float *Weights;
} mystruct;

void testfunc(mystruct* mys){
  printf(mys->Weights[0]);
}

public interface CLibrary extends Library {
    public static final String JNA_LIBRARY_NAME = "test";
    public static final NativeLibrary JNA_NATIVE_LIB = NativeLibrary.getInstance(CLibrary.JNA_LIBRARY_NAME);
    public static final CLibrary INSTANCE = (CLibrary)Native.loadLibrary(CLibrary.JNA_LIBRARY_NAME, CLibrary.class);

和Java：

public interface CLibrary extends Library {
    public static final String JNA_LIBRARY_NAME = "test";
    public static final NativeLibrary JNA_NATIVE_LIB = NativeLibrary.getInstance(CLibrary.JNA_LIBRARY_NAME);
    public static final CLibrary INSTANCE = (CLibrary)Native.loadLibrary(CLibrary.JNA_LIBRARY_NAME, CLibrary.class);


    public static class mystruct extends Structure {
        public int n;
        public FloatByReference Weights;
        protected List<? > getFieldOrder() {
            return Arrays.asList("n", "Weights");
        }
        public mystruct(int n, JnaeratorTestLibrary.feature_t.ByReference Features, FloatByReference Weights) {
            super();
            this.n = n;
            this.Features = Features;
            this.Weights = Weights;
        }
        public mystruct(Pointer peer) {
            super(peer);
        }
        public static class ByReference extends mystruct implements Structure.ByReference {

        };
        public static class ByValue extends mystruct implements Structure.ByValue {

        };
    };
    void testfunc(CLibrary.mystruct mys);
}

public static void main(String[] args) {
    mystruct mys = new mystruct();
    mys.n = 10;
    mys.Weights = new FloatByReference();
    // Compile error raises.
    mys.Weights = new float[4]; mys.Weights[0] = 1.0;

    testfunc(mys);
}

编译错误消息：

类型不匹配：无法从float []转换为FloatByReference

Answer 1

使用Pointer，然后使用Pointer.getFloatArray(0, n)提取所需的数据。

在将内存分配给结构字段之前，您需要分配Memory并使用Memory.write()将Java原始浮点数组写入本机内存。

FloatByReference基本上是指向一个分配的float的指针。 JNAerator根本无法区分指向单个float的指针和指向float的缓冲区的指针（没有一些额外的上下文的人也不能）。

如何在JNA中传递包含基本类型数组的结构？

1 个答案: