ue4同c#通信时的中文乱码问题

转：https://blog.csdn.net/XIAOZHI0999/article/details/80785570

本文讨论C#同ue4进行通信，出现的中文乱码情况，其他语言类似。

本文分两种情况讨论，

1.C#向ue4发送string

在C#发送string时，要先区分发送的字符串中是否包含中文，判断方法很简单，如下：

private bool IsPureAnsi(string str)
{
for (int i = 0; i < str.Length; ++i)
{
if ((int)str[i] > 127)
{
return false;
}
}

return true;
}
当string中不包含中文时，每个字符占1个字节，string前要先发送string的长度，占4个字节，string结束要加'\0'结尾（c传统）
当string中包含中文时，每个字节占用2个字节，格式同上（注意string长度无需*2，而且string长度要取反，string结尾需要两个字节的'\0'）

具体代码如下：

public byte[] StringToBytes(string str)
{
Stream stream = new MemoryStream();

if (this.IsPureAnsi(str))
{
byte[] strLenBytes = System.BitConverter.GetBytes(str.Length + 1);
stream.Write(strLenBytes, 0, strLenBytes.Length);

byte[] strBytes = Encoding.ASCII.GetBytes(str);
stream.Write(strBytes, 0, strBytes.Length);
stream.WriteByte(0);
}
else
{
int strLen = str.Length + 1;
byte[] strLenBytes = System.BitConverter.GetBytes(-strLen);
stream.Write(strLenBytes, 0, strLenBytes.Length);

byte[] strBytes = Encoding.Unicode.GetBytes(str);
stream.Write(strBytes, 0, strBytes.Length);
stream.WriteByte(0);
stream.WriteByte(0);
}

return StreamToBytes(stream);
}
2.ue4向C#发送string
这种情况，ue4端照常发送，c#解析的时候需要做一定处理

c#在接收到string数据时，先取出string长度信息，如果string长度>=0,则说明c#接收到的string是ascii编码，如果小于0，说明接收到的string是unicode编码

具体代码如下

public String GetString(byte[] data, int index, out int outIndex)
{
int strLen = System.BitConverter.ToInt32(data, index);
bool Ascii = strLen >= 0;
if (Ascii)
{
outIndex = index + 4 + strLen;

int begin = index + 4;
int end = outIndex;
byte[] byteStr = data.Skip(begin).Take(end - begin - 1).ToArray();

string str = Encoding.Default.GetString(byteStr);
return str;
}
else
{
strLen = -strLen;
outIndex = index + 4 + strLen*2;

int begin = index + 4;
int end = outIndex;
byte[] byteStr = data.Skip(begin).Take(end - begin - 2).ToArray();

string str = Encoding.Unicode.GetString(byteStr);
return str;
}
}
以上内容主要参考的是ue4中的源码，具体源码如下：
FArchive& operator<<( FArchive& Ar, FString& A )
{
// > 0 for ANSICHAR, < 0 for UCS2CHAR serialization

if (Ar.IsLoading())
{
int32 SaveNum;
Ar << SaveNum;

bool LoadUCS2Char = SaveNum < 0;
if (LoadUCS2Char)
{
SaveNum = -SaveNum;
}

// If SaveNum is still less than 0, they must have passed in MIN_INT. Archive is corrupted.
if (SaveNum < 0)
{
Ar.ArIsError = 1;
Ar.ArIsCriticalError = 1;
UE_LOG(LogNetSerialization, Error, TEXT("Archive is corrupted"));
return Ar;
}

auto MaxSerializeSize = Ar.GetMaxSerializeSize();
// Protect against network packets allocating too much memory
if ((MaxSerializeSize > 0) && (SaveNum > MaxSerializeSize))
{
Ar.ArIsError = 1;
Ar.ArIsCriticalError = 1;
UE_LOG( LogNetSerialization, Error, TEXT( "String is too large" ) );
return Ar;
}

// Resize the array only if it passes the above tests to prevent rogue packets from crashing
A.Data.Empty (SaveNum);
A.Data.AddUninitialized(SaveNum);

if (SaveNum)
{
if (LoadUCS2Char)
{
// read in the unicode string and byteswap it, etc
auto Passthru = StringMemoryPassthru<UCS2CHAR>(A.Data.GetData(), SaveNum, SaveNum);
Ar.Serialize(Passthru.Get(), SaveNum * sizeof(UCS2CHAR));
// Ensure the string has a null terminator
Passthru.Get()[SaveNum-1] = '\0';
Passthru.Apply();

INTEL_ORDER_TCHARARRAY(A.Data.GetData())

// Since Microsoft's vsnwprintf implementation raises an invalid parameter warning
// with a character of 0xffff, scan for it and terminate the string there.
// 0xffff isn't an actual Unicode character anyway.
int Index = 0;
if(A.FindChar(0xffff, Index))
{
A[Index] = '\0';
A.TrimToNullTerminator();
}
}
else
{
auto Passthru = StringMemoryPassthru<ANSICHAR>(A.Data.GetData(), SaveNum, SaveNum);
Ar.Serialize(Passthru.Get(), SaveNum * sizeof(ANSICHAR));
// Ensure the string has a null terminator
Passthru.Get()[SaveNum-1] = '\0';
Passthru.Apply();
}

// Throw away empty string.
if (SaveNum == 1)
{
A.Data.Empty();
}
}
}
else
{
bool SaveUCS2Char = Ar.IsForcingUnicode() || !FCString::IsPureAnsi(*A);
int32 Num = A.Data.Num();
int32 SaveNum = SaveUCS2Char ? -Num : Num;

Ar << SaveNum;

A.Data.CountBytes( Ar );

if (SaveNum)
{
if (SaveUCS2Char)
{
// TODO - This is creating a temporary in order to byte-swap. Need to think about how to make this not necessary.
#if !PLATFORM_LITTLE_ENDIAN
FString ATemp = A;
FString& A = ATemp;
INTEL_ORDER_TCHARARRAY(A.Data.GetData());
#endif

Ar.Serialize((void*)StringCast<UCS2CHAR>(A.Data.GetData(), Num).Get(), sizeof(UCS2CHAR)* Num);
}
else
{
Ar.Serialize((void*)StringCast<ANSICHAR>(A.Data.GetData(), Num).Get(), sizeof(ANSICHAR)* Num);
}
}
}

return Ar;
}
注意，ue4,发送和接收数据都是用这段代码，Ar.IsLoading()为true代表接收数据，否则为发送数据
---------------------
作者：--小智--
来源：CSDN
原文：https://blog.csdn.net/XIAOZHI0999/article/details/80785570
版权声明：本文为博主原创文章，转载请附上博文链接！