Fixed gambling caused by incorrect marshal, updated assembly version

[Akarinnnnn]

Gambling is caused by [MarshalAs(UnmanagedType.ByValTStr...
Changed it to [MarshalAs(UnmanagedType.ByValArray... and added a property to operate it

[Akarinnnnn]

This .NET Core project will also pack a Nuget package while building

[rlabrecque]

Hey, thanks, I'll get to this soon! Definitely a good improvement, though I'll probably redo the marshalling part in Steamworks.NET-CodeGen to automatically do stub out the string <-> byte[] conversion code

[rlabrecque]

Could you check this out?

#347

We can't use Span because Unity doesn't support it yet. :(

Maybe #if Unity use this, otherwise use Span?

[Akarinnnnn]

Could you check this out?
#347
We can't use Span because Unity doesn't support it yet. :(
Maybe #if Unity use this, otherwise use Span?

We don't need Span now.
By the way

--- a/Plugins/Steamworks.NET/InteropHelp.cs                                                                             
+++ b/Plugins/Steamworks.NET/InteropHelp.cs
@@ -74,7 +74,7 @@ namespace Steamworks {

                public static void StringToByteArrayUTF8(string str, byte[] outArrayBuffer, int outArrayBufferSize)
                {
-                       outArrayBuffer = new byte[outArrayBufferSize];
+                       //outArrayBuffer = new byte[outArrayBufferSize];
                        int length = Encoding.UTF8.GetBytes(str, 0, str.Length, outArrayBuffer, 0);
                        outArrayBuffer[length] = 0;
                }

outArrayBuffer should not be overriden, or our data won't written to correct buffer.

[rlabrecque]

I'm worried about two things: The \0, and also StringToByteArrayUTF8 writing more than outArrayBufferSize maybe?

[Akarinnnnn]

I'm worried about two things: The \0, and also StringToByteArrayUTF8 writing more than outArrayBufferSize maybe?

> string str = "简体中文";
> str.Length
4
> byte[] buffer = new byte[20];
> buffer
byte[20] { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }
> StringToByteArrayUTF8(str, buffer, 0);
> buffer
byte[20] { 231, 174, 128, 228, 189, 147, 228, 184, 173, 230, 150, 135, 0, 0, 0, 0, 0, 0, 0, 0 }
> buffer[13] = 231;
> buffer[14] = 174;
> buffer[15] = 128;
> ByteArrayToStringUTF8(buffer)
"简体中文"
> 

\0 is there, nothing to worry about.

> byte[] buff2 = new byte[1];
> StringToByteArrayUTF8(str, buff2, 0);
System.ArgumentException: 输出字节缓冲区太小，无法包含编码后的数据，编码“Unicode (UTF-8)”的操作回退“System.Text.EncoderReplacementFallback”。
参数名: bytes
  + System.Text.Encoding.ThrowBytesOverflow()
  + System.Text.Encoding.ThrowBytesOverflow(System.Text.EncoderNLS, bool)
  + System.Text.UTF8Encoding.GetBytes(System.Char*, int, System.Byte*, int, System.Text.EncoderNLS)
  + System.Text.UTF8Encoding.GetBytes(string, int, int, byte[], int)
  + Submission#0.StringToByteArrayUTF8(string, byte[], int)

This shown us if you try to write to a buffer that not large enough, UTF8Encoding.GetBytes will throw an System.ArgumentException.

[yaakov-h]

You definitely need to write \0 or otherwise zero out the array when writing, otherwise you can land up with garbage from previous values.

> var buffer = new byte[20];
> var str = "12345";
> Encoding.UTF8.GetBytes(str, 0, str.Length, buffer, 0);
> Encoding.UTF8.GetString(buffer)
"12345\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"
> str = "543";
> Encoding.UTF8.GetBytes(str, 0, str.Length, buffer, 0);
> Encoding.UTF8.GetString(buffer)
"54345\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"
>

[Akarinnnnn]

You definitely need to write \0 or otherwise zero out the array when writing, otherwise you can land up with garbage from previous values.

var buffer = new byte[20];
var str = "12345";
Encoding.UTF8.GetBytes(str, 0, str.Length, buffer, 0);
Encoding.UTF8.GetString(buffer)
"12345\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"
str = "543";
Encoding.UTF8.GetBytes(str, 0, str.Length, buffer, 0);
Encoding.UTF8.GetString(buffer)
"54345\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"

Sloved, see My work.

Author: Fa鸽 <1207247008@qq.com>  2020-04-13 14:35:06
Committer: Fa鸽 <1207247008@qq.com>  2020-04-13 14:35:06
Parent: d804b61dc2a226d53a83f67ef808ea2a6054d998 (Fixed a bug)
Branches: remotes/origin/u8str-b, u8str-b
Follows: 13.0.0
Precedes: 

    Restrict \0

-------------------- Plugins/Steamworks.NET/InteropHelp.cs --------------------
index d3dcb71..e5d0365 100644
@@ -76,7 +76,13 @@ namespace Steamworks {
 		{
 			//outArrayBuffer = new byte[outArrayBufferSize];
 			int length = Encoding.UTF8.GetBytes(str, 0, str.Length, outArrayBuffer, 0);
-			outArrayBuffer[length] = 0;
+#if NETSTANDARD2_1
+			System.Index index = length >= outArrayBuffer.Length ? ^1 : length;
+#else
+			int index = length >= outArrayBuffer.Length ? outArrayBuffer.Length - 1 : length;
+#endif
+			outArrayBuffer[index] = 0;
+			//outArrayBuffer[] = 0;
 		}
 
 		// This is for 'const char *' arguments which we need to ensure do not get GC'd while Steam is using them.

[rlabrecque]

Yeah I think that's looking better.

		public static void StringToByteArrayUTF8(string str, byte[] outArrayBuffer, int outArrayBufferSize)
		{
			int length = Encoding.UTF8.GetBytes(str, 0, str.Length, outArrayBuffer, 0);
#if NETSTANDARD2_1
			System.Index index = length >= outArrayBuffer.Length ? ^1 : length;
#else
			int index = length >= outArrayBuffer.Length ? outArrayBuffer.Length - 1 : length;
#endif
			outArrayBuffer[index] = 0;
		}

@yaakov-h ? ^