Question

我正在尝试修补OBJ加载类，因为它只支持短号，并且我有一个超过短数限制的大OBJ。

我必须更改代码以使用整数向量（而不是短向），并将ShortBuffer更改为IntBuffer。

现在，基本的OBJ（10,000行）曾经在3秒内加载，根本不再加载并给我一个BufferOverflowException。

以下是有问题的代码块：

ByteBuffer fBuf = ByteBuffer.allocateDirect(faces.size() * 2);
fBuf.order(ByteOrder.nativeOrder());
faceBuffer = fBuf.asIntBuffer();
faceBuffer.put(toPrimitiveArrayS(faces)); // this gives the exception
faceBuffer.position(0);

下面是函数 toPrimitiveArrayS

private static int[] toPrimitiveArrayS(Vector<Integer> vector){
    int[] s;
    s = new int[vector.size()];
    for (int i=0; i<vector.size(); i++){
        s[i] = vector.get(i);
    }
    return s;
}

奇怪的是，如果我在allocateDirect中将2更改为4，那么模型会加载，但会被错误地渲染（以前，它加载时该数字为2，所以我只是将其更改为看）

以下完整课程代码：

public class TDModelPart {
Vector<Integer> faces;
Vector<Integer> vtPointer;
Vector<Integer> vnPointer;
Material material;
private FloatBuffer normalBuffer;
IntBuffer faceBuffer;

public TDModelPart(Vector<Integer> faces, Vector<Integer> vtPointer, Vector<Integer> vnPointer, Material material, Vector<Float> vn) {
    super();

    this.faces = faces;
    this.vtPointer = vtPointer;
    this.vnPointer = vnPointer;
    this.material = material;

    ByteBuffer byteBuf = ByteBuffer.allocateDirect(vnPointer.size() * 4 * 3);
    byteBuf.order(ByteOrder.nativeOrder());
    normalBuffer = byteBuf.asFloatBuffer();

    for(int i=0; i<vnPointer.size(); i++){
        float x=vn.get(vnPointer.get(i)*3);
        float y=vn.get(vnPointer.get(i)*3+1);
        float z=vn.get(vnPointer.get(i)*3+2);
        normalBuffer.put(x);
        normalBuffer.put(y);
        normalBuffer.put(z);
    }

    normalBuffer.position(0);

    ByteBuffer fBuf = ByteBuffer.allocateDirect(faces.size() * 2);
    fBuf.order(ByteOrder.nativeOrder());
    faceBuffer = fBuf.asIntBuffer();
    faceBuffer.put(toPrimitiveArrayS(faces));
    faceBuffer.position(0);
}

public String toString(){
    String str=new String();
    if(material!=null)
        str+="Material name:"+material.getName();
    else
        str+="Material not defined!";
    str+="\nNumber of faces:"+faces.size();
    str+="\nNumber of vnPointers:"+vnPointer.size();
    str+="\nNumber of vtPointers:"+vtPointer.size();
    return str;
}

public IntBuffer getFaceBuffer(){
    return faceBuffer;
}

public FloatBuffer getNormalBuffer(){
    return normalBuffer;
}

private static int[] toPrimitiveArrayS(Vector<Integer> vector){
    int[] s;
    s = new int[vector.size()];
    for (int i=0; i<vector.size(); i++){
        s[i] = vector.get(i);
    }
    return s;
}

public int getFacesCount(){
    return faces.size();
}

public Material getMaterial(){
    return material;
}
}

Answer 1

你为每个面部索引分配两个字节（16位），你应该分配四个，因为你每个索引使用一个int（一个int是32位整数，四个字节），你试图分配双倍数量你的faceBuffer可以处理的字节数

将ShortBuffer更改为IntBuffer时的BufferOverFlowException

1 个答案: