add C# demo and encoding check.

Author: lixungeng

.gitignore

@@ -13,3 +13,6 @@
 
 .vs/
 /demo/
+Properties/
+bin/
+obj/

c_sharp/Program.cs

@@ -0,0 +1,49 @@
+using System;
+using System.Runtime.InteropServices;
+using System.Text;
+
+class Program
+{
+    // 注意要把生成的64位dll拷到 c_sharp\bin\x64\Debug\net8.0 目录下
+    // 声明外部函数
+    [DllImport("wcocr.dll", CallingConvention = CallingConvention.Cdecl)]
+    public static extern bool wechat_ocr(
+        [MarshalAs(UnmanagedType.LPWStr)] string ocr_exe,
+        [MarshalAs(UnmanagedType.LPWStr)] string wechat_dir,
+        [MarshalAs(UnmanagedType.LPStr)] string imgfn,
+        SetResultDelegate set_res);
+
+    // 定义委托类型
+    public delegate void SetResultDelegate(IntPtr result);
+
+    public class StringResult
+    {
+        public string result_ = "";
+        public void SetResult(IntPtr dt)
+        {
+            int length = 0;
+            while (Marshal.ReadByte(dt, length) != 0) length++;
+            byte[] byteArray = new byte[length];
+            Marshal.Copy(dt, byteArray, 0, length);
+            result_ = Encoding.UTF8.GetString(byteArray);
+        }
+        public string GetResult()
+        {
+            return result_;
+        }
+    };
+
+    static void Main(string[] args)
+    {
+        if (args.Length != 3)
+        {
+            Console.WriteLine("Usage: wcocr.exe ocr_exe wechat_dir imgfn");
+            return;
+        }
+
+        StringResult stringResult = new StringResult();
+        SetResultDelegate setRes = new SetResultDelegate(stringResult.SetResult);
+        bool success = wechat_ocr(args[0], args[1], args[2], setRes);
+        Console.WriteLine($"OCR Success: {success} res:{stringResult.GetResult()}");
+    }
+}

c_sharp/test_cs.csproj

@@ -0,0 +1,11 @@
+<Project Sdk="Microsoft.NET.Sdk">
+
+  <PropertyGroup>
+    <OutputType>Exe</OutputType>
+    <TargetFramework>net8.0</TargetFramework>
+    <ImplicitUsings>enable</ImplicitUsings>
+    <Nullable>enable</Nullable>
+    <Platforms>AnyCPU;x64</Platforms>
+  </PropertyGroup>
+
+</Project>

c_sharp/test_cs.sln

@@ -0,0 +1,31 @@
+
+Microsoft Visual Studio Solution File, Format Version 12.00
+# Visual Studio Version 17
+VisualStudioVersion = 17.9.34728.123
+MinimumVisualStudioVersion = 10.0.40219.1
+Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "test_cs", "test_cs.csproj", "{0F79C844-AF8C-44CA-9D3C-27F1E47E4805}"
+EndProject
+Global
+	GlobalSection(SolutionConfigurationPlatforms) = preSolution
+		Debug|Any CPU = Debug|Any CPU
+		Debug|x64 = Debug|x64
+		Release|Any CPU = Release|Any CPU
+		Release|x64 = Release|x64
+	EndGlobalSection
+	GlobalSection(ProjectConfigurationPlatforms) = postSolution
+		{0F79C844-AF8C-44CA-9D3C-27F1E47E4805}.Debug|Any CPU.ActiveCfg = Debug|Any CPU
+		{0F79C844-AF8C-44CA-9D3C-27F1E47E4805}.Debug|Any CPU.Build.0 = Debug|Any CPU
+		{0F79C844-AF8C-44CA-9D3C-27F1E47E4805}.Debug|x64.ActiveCfg = Debug|x64
+		{0F79C844-AF8C-44CA-9D3C-27F1E47E4805}.Debug|x64.Build.0 = Debug|x64
+		{0F79C844-AF8C-44CA-9D3C-27F1E47E4805}.Release|Any CPU.ActiveCfg = Release|Any CPU
+		{0F79C844-AF8C-44CA-9D3C-27F1E47E4805}.Release|Any CPU.Build.0 = Release|Any CPU
+		{0F79C844-AF8C-44CA-9D3C-27F1E47E4805}.Release|x64.ActiveCfg = Release|x64
+		{0F79C844-AF8C-44CA-9D3C-27F1E47E4805}.Release|x64.Build.0 = Release|x64
+	EndGlobalSection
+	GlobalSection(SolutionProperties) = preSolution
+		HideSolutionNode = FALSE
+	EndGlobalSection
+	GlobalSection(ExtensibilityGlobals) = postSolution
+		SolutionGuid = {61ED6E08-2B59-4CDB-8534-BC52560BC05E}
+	EndGlobalSection
+EndGlobal

pb/ocr_protobuf.proto

@@ -48,7 +48,7 @@ message OcrRespond {
 message OcrRequest {
 	int32 unknow = 1;								//必定为0
 	// 核查了好几次，在腾讯proto文件中，这个task_id确实是64位的。但在，在执行过程中，高32位会被丢弃，且第32位为1会出错。
-	// 也就是协议上是有64位的uint64，实际上只有31位。必须是>0的整形数字，范围是[12147483647]
+	// 也就是协议上是有64位的uint64，实际上只有31位。必须是>0的整形数字，范围是[1,2147483647]
 	uint64 task_id = 2;
 
 	message OcrInputBuffer {

src/main.cpp

@@ -75,7 +75,6 @@ HRESULT DllRegisterServer(void)
 		return E_FAIL;
 	}
 	wechat_ocr.doOCR(getenv("TEST_PNG"), nullptr);
-	// MessageBoxW(NULL, L"注册成功", L"提示", MB_OK);
 	wechat_ocr.wait_done(-1);
 	return S_OK;
 }

src/wechatocr.cpp

@@ -3,10 +3,55 @@
 #include "ocr_protobuf.pb.h"
 #include "mmmojo.h"
 
+namespace {
+	bool is_text_utf8(const char* sin, size_t len) {
+		const unsigned char* s = (const unsigned char*)sin;
+		const unsigned char* end = s + len;
+		while (s < end) {
+			if (*s < 0x80) {
+				++s;
+			} else if (*s < 0xC0) {
+				return false;
+			} else if (*s < 0xE0) {
+				if (s + 1 >= end || (s[1] & 0xC0) != 0x80)
+					return false;
+				s += 2;
+			} else if (*s < 0xF0) {
+				if (s + 2 >= end || (s[1] & 0xC0) != 0x80 || (s[2] & 0xC0) != 0x80)
+					return false;
+				s += 3;
+			} else if (*s < 0xF8) {
+				if (s + 3 >= end || (s[1] & 0xC0) != 0x80 || (s[2] & 0xC0) != 0x80 || (s[3] & 0xC0) != 0x80)
+					return false;
+				s += 4;
+			} else {
+				return false;
+			}
+		}
+		return true;
+	}
+
+}
+
 CWeChatOCR::CWeChatOCR(LPCWSTR exe, LPCWSTR wcdir)
 {
 	m_exe = exe;
 	m_wcdir = wcdir;
+	DWORD attr1 = GetFileAttributesW(exe);
+	if (attr1 == INVALID_FILE_ATTRIBUTES || (attr1 & FILE_ATTRIBUTE_DIRECTORY) != 0)
+	{
+		// 传入的ocr.exe路径无效
+		m_state = STATE_INVALID;
+		return;
+	}
+	DWORD attr2 = GetFileAttributesW(wcdir);
+	if (attr2 == INVALID_FILE_ATTRIBUTES || (attr2 & FILE_ATTRIBUTE_DIRECTORY) == 0)
+	{
+		// 传入的微信目录无效
+		m_state = STATE_INVALID;
+		return;
+	}
+
 	if (Init(m_wcdir.c_str()))
 	{
 		m_args["user-lib-dir"] = m_wcdir;
@@ -22,17 +67,35 @@ CWeChatOCR::CWeChatOCR(LPCWSTR exe, LPCWSTR wcdir)
 }
 
 #define OCR_MAX_TASK_ID INT_MAX
-bool CWeChatOCR::doOCR(crefstr imgpath, result_t* res)
+bool CWeChatOCR::doOCR(crefstr imgpath0, result_t* res)
 {
 	if (m_state == STATE_INVALID || (m_state == STATE_PENDING && !wait_connection(2000)))
 		return false;
 
+	// 为了更好的中文支持，这里检查imgpath是否是utf8编码的，如果是GBK，需要转换为utf8编码。
+	// TODO: 其实不应该写在这里，应该是调用者确保传入的imgpath是utf8编码的。
+	string tmp;
+	const string * imgpath = &imgpath0;
+	if (!is_text_utf8(imgpath0.c_str(), imgpath0.length())) {
+		wstring wimgpath;
+		wimgpath.resize(imgpath0.size() + 10);
+		int len = MultiByteToWideChar(CP_ACP, 0, imgpath0.c_str(), (int)imgpath0.length(), &wimgpath[0], (int)wimgpath.size());
+		if (len > 0) {
+			tmp.resize(len * 3 + 10);
+			len = WideCharToMultiByte(CP_UTF8, 0, wimgpath.c_str(), len, &tmp[0], (int)tmp.size(), 0, 0);
+			if (len > 0) {
+				tmp.resize(len);
+				imgpath = &tmp;
+			}
+		}
+	}
+
 	int found_id = -1;
 	m_mutex.lock();
 	// TODO: task_id本身可以是任何正整形数，但是这里要不要限制同时并发任务数呢？
 	for (uint32_t i = 1; i <= OCR_MAX_TASK_ID; ++i)
 	{
-		if (m_idpath.insert(std::make_pair(i, std::pair<string,result_t*>(imgpath,res))).second)
+		if (m_idpath.insert(std::make_pair(i, std::pair<string,result_t*>(*imgpath,res))).second)
 		{
 			found_id = i;
 			break;
@@ -47,7 +110,7 @@ bool CWeChatOCR::doOCR(crefstr imgpath, result_t* res)
 	ocr_request.set_unknow(0);
 	ocr_request.set_task_id(found_id);
 	auto pp = new ocr_protobuf::OcrRequest::OcrInputBuffer;
-	pp->set_pic_path(imgpath);
+	pp->set_pic_path(*imgpath);
 	ocr_request.set_allocated_input(pp);
 
 	std::string data_;