I'm trying to load an .md2 model in Java.
There's a structure like this:
struct md2_triangle_t
{
unsigned short vertex[3];
unsigned short st[3];
};
Will there be byte packing? Where and how many bytes do I skip? Or can I read the 2-byte unsigned shorts one after the other?
Take a look at the DataInputStream, there are methods to read ushorts etc. Padding etc. are often memory/compiler options, I would think, that the above structure is written 1:1, therefor try something like this:
// fast pseudo code, needs some checking and cleaning up
class md2_triangle_t
{
public int vertex[3];
public int st[3];
public void readFromStream(DataInputStream _inputStream)
{
vertex[0] = _inputStream.readUnsignedShort();
vertex[1] = _inputStream.readUnsignedShort();
vertex[2] = _inputStream.readUnsignedShort();
st[0] = _inputStream.readUnsignedShort();
st[1] = _inputStream.readUnsignedShort();
st[2] = _inputStream.readUnsignedShort();
}
}
They're little-endian ushorts so
// http://darksleep.com/player/JavaAndUnsignedTypes.html
static char ReadUShort(InputStream iS, int offset)
{
byte[] bucket = ReadBytes(iS, offset, 2);
// switched bytes for little-endian:
int firstByte = (0x000000FF & ((int)bucket[1]));
int secondByte = (0x000000FF & ((int)bucket[0]));
return (char) (firstByte << 8 | secondByte);
}
static byte[] ReadBytes(InputStream iS, int offset, int readlen)
{
byte[] bucket = new byte[offset+readlen];
try
{
iS.read(bucket, offset, readlen);
}
catch (IOException e)
{
e.printStackTrace();
}
byte subbucket[] = new byte[readlen];
for(int i=0; i<readlen; i++)
subbucket = bucket[offset+i];
return subbucket;
}
