implement searching for read codes

GUI/misc.cpp

@@ -18,10 +18,10 @@ namespace
 		case L'S':
 			break;
 		case L'Q':
-			hp.type |= USING_STRING | USING_UNICODE;
+			hp.type |= USING_UNICODE;
 			break;
 		case L'V':
-			hp.type |= USING_STRING | USING_UTF8;
+			hp.type |= USING_UTF8;
 			break;
 		default:
 			return {};
@@ -36,7 +36,7 @@ namespace
 			RCode.remove(0, codepage.captured(0).length());
 		}
 
-		// [*deref_offset|0]
+		// [*deref_offset]
 		if (RCode.at(0).unicode() == L'0') RCode.remove(0, 1); // Legacy
 		QRegularExpressionMatch deref = QRegularExpression("^\\*(\\-?[[:xdigit:]]+)").match(RCode);
 		if (deref.hasMatch())
@@ -53,6 +53,24 @@ namespace
 		return hp;
 	}
 
+	std::optional<HookParam> ParseSCode(QString SCode)
+	{
+		HookParam hp = {};
+		hp.type |= READ_SEARCH;
+
+		// [codepage#]
+		QRegularExpressionMatch codepage = QRegularExpression("^([0-9]+)#").match(SCode);
+		if (codepage.hasMatch())
+		{
+			hp.codepage = codepage.captured(1).toInt();
+			SCode.remove(0, codepage.captured(0).length());
+		}
+
+		wcscpy_s<MAX_MODULE_SIZE>(hp.text, S(SCode).c_str());
+
+		return hp;
+	}
+
 	std::optional<HookParam> ParseHCode(QString HCode)
 	{
 		HookParam hp = {};
@@ -157,7 +175,7 @@ namespace
 
 	QString GenerateRCode(HookParam hp)
 	{
-		QString RCode = "/R";
+		QString RCode = "R";
 		QTextStream codeBuilder(&RCode);
 
 		if (hp.type & USING_UNICODE) codeBuilder << "Q";
@@ -178,7 +196,7 @@ namespace
 
 	QString GenerateHCode(HookParam hp, DWORD processId)
 	{
-		QString HCode = "/H";
+		QString HCode = "H";
 		QTextStream codeBuilder(&HCode);
 
 		if (hp.type & USING_UNICODE)
@@ -229,8 +247,10 @@ namespace
 
 std::optional<HookParam> ParseCode(QString code)
 {
-	if (code.startsWith("/H")) return ParseHCode(code.remove(0, 2));
+	if (code.startsWith("/")) code.remove(0, 1); // legacy/AGTH compatibility
-	else if (code.startsWith("/R")) return ParseRCode(code.remove(0, 2));
+	if (code.startsWith("R")) return ParseRCode(code.remove(0, 1));
+	else if (code.startsWith("S")) return ParseSCode(code.remove(0, 1));
+	else if (code.startsWith("H")) return ParseHCode(code.remove(0, 1));
 	else return {};
 }
 
@@ -242,7 +262,7 @@ QString GenerateCode(HookParam hp, DWORD processId)
 
 TEST(
 	assert(ParseCode("/HQN936#-c*C:C*1C@4AA:gdi.dll:GetTextOutA")),
-	assert(ParseCode("/HB4@0")),
+	assert(ParseCode("HB4@0")),
 	assert(ParseCode("/RS*10@44")),
 	assert(!ParseCode("HQ@4")),
 	assert(!ParseCode("/RW@44")),

include/const.h

@@ -33,8 +33,9 @@ enum HookParamType : unsigned long
 	USING_SPLIT = 0x10, // aware of split time?
 	SPLIT_INDIRECT = 0x20,
 	MODULE_OFFSET = 0x40, // address is relative to module
-	FUNCTION_OFFSET = 0x80,    // address is relative to function
+	FUNCTION_OFFSET = 0x80, // address is relative to function
 	USING_UTF8 = 0x100,
+	READ_SEARCH = 0x200, // unspecified address: search for text instead
 	NO_CONTEXT = 0x400,
 	HOOK_EMPTY = 0x800,
 	FIXING_SPLIT = 0x1000,

include/text.h

@@ -14,10 +14,13 @@ constexpr auto SELECT_PROCESS = u8"Select Process";
 constexpr auto ATTACH_INFO = u8"If you don't see the process you want to attach, try running with admin rights\r\n"
 u8"You can also type in the process id";
 constexpr auto CODE_INFODUMP = u8"Enter hook code\r\n"
-u8"/H{A|B|W|S|Q|V}[N][codepage#]data_offset[*deref_offset1][:split_offset[*deref_offset2]]@addr[:module[:func]]\r\n"
+u8"H{A|B|W|S|Q|V}[N][codepage#]data_offset[*deref_offset1][:split_offset[*deref_offset2]]@addr[:module[:func]]\r\n"
 u8"OR\r\n"
 u8"Enter read code\r\n"
-u8"/R{S|Q|V}[codepage#][*deref_offset|0]@addr\r\n"
+u8"R{S|Q|V}[codepage#][*deref_offset]@addr\r\n"
+u8"OR\r\n"
+u8"Search for read code\r\n"
+u8"S[codepage#]text\r\n"
 u8"All numbers except codepage in hexadecimal\r\n"
 u8"A/B: Shift-JIS char little/big endian\r\n"
 u8"W: UTF-16 char\r\n"
@@ -50,6 +53,7 @@ constexpr auto INSERTING_HOOK = u8"Textractor: inserting hook: %s";
 constexpr auto REMOVING_HOOK = u8"Textractor: removing hook: %s";
 constexpr auto HOOK_FAILED = u8"Textractor: failed to insert hook";
 constexpr auto TOO_MANY_HOOKS = u8"Textractor: too many hooks: can't insert";
+constexpr auto NOT_ENOUGH_TEXT = u8"Textractor: not enough text to search accurately";
 constexpr auto FUNC_MISSING = u8"Textractor: function not present";
 constexpr auto MODULE_MISSING = u8"Textractor: module not present";
 constexpr auto GARBAGE_MEMORY = u8"Textractor: memory constantly changing, useless to read";

include/types.h

@@ -54,7 +54,11 @@ struct HookParam
 		index, // deref_offset1
 		split, // offset of the split character
 		split_index; // deref_offset2
-	wchar_t module[MAX_MODULE_SIZE];
+	union
+	{
+		wchar_t module[MAX_MODULE_SIZE];
+		wchar_t text[MAX_MODULE_SIZE];
+	};
 	char function[MAX_MODULE_SIZE];
 	DWORD type; // flags
 	UINT codepage; // text encoding

vnrhook/main.cc

@@ -9,6 +9,7 @@
 #include "MinHook.h"
 #include "engine/match.h"
 #include "texthook.h"
+#include "util.h"
 
 std::unique_ptr<WinMutex> viewMutex;
 
@@ -133,11 +134,35 @@ BOOL WINAPI DllMain(HINSTANCE hModule, DWORD fdwReason, LPVOID)
 
 void NewHook(HookParam hp, LPCSTR lpname, DWORD flag)
 {
-	if (++currentHook >= MAX_HOOK) return ConsoleOutput(TOO_MANY_HOOKS);
+	if (hp.type & READ_SEARCH)
-	if (lpname && *lpname) strcpy_s<HOOK_NAME_SIZE>(hp.name, lpname);
+	{
-	ConsoleOutput(INSERTING_HOOK, hp.name);
+		char utf8Text[MAX_MODULE_SIZE * 4] = {};
-	RemoveHook(hp.address, 0);
+		WideCharToMultiByte(CP_UTF8, 0, hp.text, MAX_MODULE_SIZE, utf8Text, MAX_MODULE_SIZE * 4, nullptr, nullptr);
-	if (!hooks[currentHook].Insert(hp, flag)) ConsoleOutput(HOOK_FAILED);
+		char codepageText[MAX_MODULE_SIZE * 4] = {};
+		WideCharToMultiByte(hp.codepage ? hp.codepage : SHIFT_JIS, 0, hp.text, MAX_MODULE_SIZE, codepageText, MAX_MODULE_SIZE * 4, nullptr, nullptr);
+		if (strlen(utf8Text) < 8 || strlen(codepageText) < 8 || wcslen(hp.text) < 4) return ConsoleOutput(NOT_ENOUGH_TEXT);
+		for (auto[addrs, type] : Array<std::tuple<std::vector<uint64_t>, HookParamType>>{
+			{ Util::SearchMemory(utf8Text, strlen(utf8Text), PAGE_READWRITE), USING_UTF8 },
+			{ Util::SearchMemory(codepageText, strlen(codepageText), PAGE_READWRITE), (HookParamType)0 },
+			{ Util::SearchMemory(hp.text, wcslen(hp.text) * 2, PAGE_READWRITE), USING_UNICODE }
+		})
+			for (auto addr : addrs)
+			{
+				HookParam h = {};
+				h.type = DIRECT_READ | type;
+				h.address = addr;
+				h.codepage = hp.codepage;
+				NewHook(h, lpname, 0);
+			}
+	}
+	else
+	{
+		if (++currentHook >= MAX_HOOK) return ConsoleOutput(TOO_MANY_HOOKS);
+		if (lpname && *lpname) strcpy_s<HOOK_NAME_SIZE>(hp.name, lpname);
+		ConsoleOutput(INSERTING_HOOK, hp.name);
+		RemoveHook(hp.address, 0);
+		if (!hooks[currentHook].Insert(hp, flag)) ConsoleOutput(HOOK_FAILED);
+	}
 }
 
 void RemoveHook(uint64_t addr, int maxOffset)

vnrhook/util/util.cc

@@ -302,7 +302,7 @@ bool SearchResourceString(LPCWSTR str)
   return false;
 }
 
-std::vector<uint64_t> SearchMemory(const BYTE* bytes, short length, DWORD protect)
+std::vector<uint64_t> SearchMemory(const void* bytes, short length, DWORD protect)
 {
 	std::vector<std::pair<uint64_t, uint64_t>> validMemory;
 	for (BYTE* probe = NULL; (uint64_t)probe < 0x80000000;) // end of user memory space
@@ -323,7 +323,7 @@ std::vector<uint64_t> SearchMemory(const BYTE* bytes, short length, DWORD protec
 	std::vector<uint64_t> ret;
 	for (auto memory : validMemory)
 		for (uint64_t addr = memory.first; true;)
-			if (addr = SafeSearchMemory(addr, memory.first + memory.second, bytes, length))
+			if (addr = SafeSearchMemory(addr, memory.first + memory.second, (const BYTE*)bytes, length))
 				ret.push_back(addr++);
 			else break;
 

vnrhook/util/util.h

@@ -22,7 +22,7 @@ bool CheckFile(LPCWSTR name);
 
 bool SearchResourceString(LPCWSTR str);
 
-std::vector<uint64_t> SearchMemory(const BYTE* bytes, short length, DWORD protect = PAGE_EXECUTE);
+std::vector<uint64_t> SearchMemory(const void* bytes, short length, DWORD protect = PAGE_EXECUTE);
 
 } // namespace Util