2011.5.18 Converting C++ Unsigned Numbers to Java

Converting C++ Unsigned Numbers to Java

原址：http://jwinblad.com/resources/unsignedtojava.html

 

Dealing with unsigned values from C++ can seem like a headache sometimes, if you don't understand what's going on "under the hood". Luckily you don't have to understand everything under the hood to read a binary file written by a C++ app in Java or vice versa, just have a vague understanding of the issues involved..

The major differences that can cause challenges are:
1) Figuring out which data types are "equivalent"
2) Java doesn't have unsigned types, but C++ types can be signed or unsigned.
3) Some types just don't have a direct equivalent (eg: C++ float has no exact match in Java)

Here's a table showing how internally C++ (on a 32-bit version of Windows) and Java store integer types:

Size	1 byte	2 bytes	4 bytes	8 bytes	16 bytes
	8 bits	16 bits	32 bits	64 bits	128 bits
C++ (on Win32)	byte	short	int/long	long long	__int128
Java	byte	short	int	long	BigInteger

If you are using a 64 bit operating system or a 16 bit operating system, the C++ sizes such as how many bytes an int is will vary. Microsoft invented some funky looking type names like __int32 for those times when you want to tell the compiler "when I said I wanted 4 bytes for my integer, I meant 4 bytes", but that may only work with Microsoft compilers.

Here are some other binary data types that are frequently encountered:

Size	1 byte	2 bytes	4 bytes	8 bytes	16 bytes
	8 bits	16 bits	32 bits	64 bits	128 bits
C++ (on Win32)	char/bool		float	double/long double
Java	boolean	char		float	double

Going from C++ to Java:

C++ Type	Java Type	Read from File **	Then translate the value
unsigned byte	(signed) short	byte a = in.readByte();	short b = (short)(a & 0xff);
unsigned short	(signed) int	short a = in.readShort();	int b = a & 0xffff;
unsigned int/long	(signed) long	int a = in.readInt();	long b = a & 0xffffffffL;
float*	float	int a = in.readInt();	float b = Float.intBitsToFloat(a);
double	float	float a = in.readFloat();	//doing nothing extra

*note: because Java does not have a 4 byte float, converting 4 byte c++ float to 8 byte java float and then back to 4 byte c++ float may result in rounding errors. Use Float.floatBitsToInt(floatValue); to convert in the other direction.
** in refers to a binary input stream type of your choice.

To convert back and forth, you just cast in the to C++ direction. Easy.

Code example: (int or short could be used here just as well)

public static short toUnsigned(byte b) { 
    return (short)(b & 0xff);
}
public static byte toSigned(short i) { 
     return (byte) i; 
}

*C++ type sizes are assuming you are running a 32 bit version of Windows. You can always explicitly set your sizing by declaring the types as __int8 (unsigned __int8)

 

For a C++ unsigned byte (value 0 to 255), mask with 0xff and store the result in a short:
byte a = inputStream.readByte(); // assuming you have your stream already set up
short b = (short)(a & 0xff);

short c = inputStream.readShort();

 

 

 

另：

如:C++中的unsigned long,无符号的long型,如果是VC的话,应该是4个字节,只不过由于它是无符号的,所以它的补码就是原码,表示范围是0-2^32-1. 那么Java程序接收时,也会接收到一个4字节的东西.但是,由于这个是四字节,无符号.所以很有可能,它的第31位(第0位是最低位)是1,而Java接收到这个,如果给int型的话,Java的int就是有符号的,最高位是1,表示是负数,这样就产生了问题(当然,如果最高位不是1,就没有问题).现在假设这个值已经给一个int了
int a=...;//Java的a中已经接收了这个unsigned long,并且最高位是1,所以变成了负数
long b=a; //将a转换为long型,由于它是负数,所以它的符号位会填充到long的高32位中.你现在只需要做一件事情况:
b=(0xffffffff&b);//哈哈,这样就可以了