1 files changed, 109 insertions, 146 deletions

diff --git a/buildtools/codegen/main.cpp b/buildtools/codegen/main.cpp
index cf31bd8..9f89191 100644
--- a/buildtools/codegen/main.cpp
+++ b/buildtools/codegen/main.cpp
@@ -1,4 +1,5 @@
 #include "CodegenLookupTable.h"
+#include "CodegenOutput.hpp"
 #include "Macros.hpp"
 #include "ScopeGuard.hpp"
 #include "Utils.hpp"
@@ -65,124 +66,10 @@ struct InputDefinitionEnum {
 	EnumUnderlyingType underlyingType;
 };
 
-enum LexedTokenType {
-	// stb_c_lexer token types, ported over
-	LTT_Identifier,
-	LTT_IntLiteral,
-	LTT_FloatLiteral,
-	LTT_DqString,
-	LTT_SqString,
-	LTT_CharLiteral,
-	LTT_OperEquals,
-	LTT_OperNotEquals,
-	LTT_OperLessOrEqual,
-	LTT_OperGreaterOrEqual,
-	LTT_OperAndAnd,
-	LTT_OperOrOr,
-	LTT_OperShiftLeft,
-	LTT_OperShiftRight,
-	LTT_OperIncrement,
-	LTT_OperDecrement,
-	LTT_OperAddAssign,
-	LTT_OperSubAssign,
-	LTT_OperMulAssign,
-	LTT_OperDivAssign,
-	LTT_OperModAssign,
-	LTT_OperAndAssign,
-	LTT_OperOrAssign,
-	LTT_OperXorAssign,
-	LTT_OperArrow,
-	LTT_OperEqualArrow,
-	LTT_OperShiftLeflAssign,
-	LTT_OperShiftRightAssign,
-
-	// Custom token types
-	LTT_OperAdd,
-	LTT_OperSub,
-	LTT_OperMul,
-	LTT_OperDiv,
-	LTT_OperMod,
-	LTT_ParenOpen,
-	LTT_ParenClose,
-	LTT_BracketOpen,
-	LTT_BracketClose,
-	LTT_BraceOpen,
-	LTT_BraceClose,
-
-	LTT_COUNT,
-};
-
-// NOTE: maintain with CLEX_* defined in stb_c_lexer.h
-LUT_DECL_VAR(gClexTokens, int, CLEX_first_unused_token, LexedTokenType, LTT_COUNT) {
-	LUT_MAP_FOR(gClexTokens);
-	LUT_MAP(CLEX_id, LTT_Identifier);
-	LUT_MAP(CLEX_intlit, LTT_IntLiteral);
-	LUT_MAP(CLEX_floatlit, LTT_FloatLiteral);
-	LUT_MAP(CLEX_dqstring, LTT_DqString);
-	LUT_MAP(CLEX_sqstring, LTT_SqString);
-	LUT_MAP(CLEX_charlit, LTT_CharLiteral);
-	LUT_MAP(CLEX_eq, LTT_OperEquals);
-	LUT_MAP(CLEX_noteq, LTT_OperNotEquals);
-	LUT_MAP(CLEX_lesseq, LTT_OperLessOrEqual);
-	LUT_MAP(CLEX_greatereq, LTT_OperGreaterOrEqual);
-	LUT_MAP(CLEX_andand, LTT_OperAndAnd);
-	LUT_MAP(CLEX_oror, LTT_OperOrOr);
-	LUT_MAP(CLEX_shl, LTT_OperShiftLeft);
-	LUT_MAP(CLEX_shr, LTT_OperShiftRight);
-	LUT_MAP(CLEX_plusplus, LTT_OperIncrement);
-	LUT_MAP(CLEX_minusminus, LTT_OperDecrement);
-	LUT_MAP(CLEX_pluseq, LTT_OperAddAssign);
-	LUT_MAP(CLEX_minuseq, LTT_OperSubAssign);
-	LUT_MAP(CLEX_muleq, LTT_OperMulAssign);
-	LUT_MAP(CLEX_diveq, LTT_OperDivAssign);
-	LUT_MAP(CLEX_modeq, LTT_OperModAssign);
-	LUT_MAP(CLEX_andeq, LTT_OperAndAssign);
-	LUT_MAP(CLEX_oreq, LTT_OperOrAssign);
-	LUT_MAP(CLEX_xoreq, LTT_OperXorAssign);
-	LUT_MAP(CLEX_arrow, LTT_OperArrow);
-	LUT_MAP(CLEX_eqarrow, LTT_OperEqualArrow);
-	LUT_MAP(CLEX_shleq, LTT_OperShiftLeflAssign);
-	LUT_MAP(CLEX_shreq, LTT_OperShiftRightAssign);
-}
-
-LUT_DECL_VAR(gSingleCharTokens, char, std::numeric_limits<char>::max() + 1, LexedTokenType, LTT_COUNT) {
-	LUT_MAP_FOR(gSingleCharTokens);
-	LUT_MAP('+', LTT_OperAdd);
-	LUT_MAP('-', LTT_OperSub);
-	LUT_MAP('*', LTT_OperMul);
-	LUT_MAP('/', LTT_OperDiv);
-	LUT_MAP('%', LTT_OperMod);
-	LUT_MAP('(', LTT_ParenOpen);
-	LUT_MAP(')', LTT_ParenClose);
-	LUT_MAP('[', LTT_BracketOpen);
-	LUT_MAP(']', LTT_BracketClose);
-	LUT_MAP('{', LTT_BraceOpen);
-	LUT_MAP('}', LTT_BraceClose);
-}
-
-// See stb_c_lexer.h's comments, here are a few additinos that aren't made clear in the file:
-// - `lexer->token` (noted as "token" below) after calling stb_c_lexer_get_token() contains either:
-//     1. 0 <= token < 256:          an ASCII character (more precisely a single char that the lexer ate; technically can be an incomplete code unit)
-//     2. token < 0:                 an unknown token
-//     3. One of the `CLEX_*` enums: a special, recognized token such as an operator
-LexedTokenType MapFromStb(const stb_lexer& lexer) {
-	if (lexer.token >= 0 && lexer.token < 256) {
-		// Single char token
-		char c = lexer.token;
-		return LUT_LOOKUP(gSingleCharTokens, lexer.token);
-	}
-
-	return LUT_LOOKUP(gClexTokens, lexer.token);
-}
-int MapToStb(LexedTokenType token) {
-	// TODO
-
-	return LUT_REV_LOOKUP(gClexTokens, token);
-}
-
 struct StbLexerToken {
 	std::string text;
-	LexedTokenType type;
+	// Can either be CLEX_* values, or just chars for single character tokens
+	int type;
 };
 
 void CheckBraceDepth(int braceDpeth) {
@@ -200,17 +87,37 @@ void HandleInputFile(std::string_view source) {
 
 	std::vector<StbLexerToken> tokens;
 	std::vector<InputDefinitionStruct> foundStructs;
+	InputDefinitionStruct currStruct;
 	std::vector<InputDefinitionEnum> foundEnums;
+	InputDefinitionEnum currEnum;
+
+	auto PushFoundStruct = [&]() {
+		foundStructs.push_back(std::move(currStruct));
+		currStruct = {};
+	};
+	auto PushFoundEnum = [&]() {
+		foundEnums.push_back(std::move(currEnum));
+		currEnum = {};
+	};
 
 	enum NextMatchingConstruct {
 		NMC_None,
 		NMC_Enum,
 		NMC_StructClass,
 	} matchingConstruct = NMC_None;
+	bool matchingConstructInBody = false;
+
+	bool matchingDirectiveParams = false;
 
 	int bracePairDepth = 0;
 
 	while (true) {
+		// See stb_c_lexer.h's comments, here are a few additinos that aren't made clear in the file:
+		// - `lexer->token` (noted as "token" below) after calling stb_c_lexer_get_token() contains either:
+		//     1. 0 <= token < 256:          an ASCII character (more precisely a single char that the lexer ate; technically can be an incomplete code unit)
+		//     2. token < 0:                 an unknown token
+		//     3. One of the `CLEX_*` enums: a special, recognized token such as an operator
+
 		int stbToken = stb_c_lexer_get_token(&lexer);
 		if (stbToken == 0) {
 			// EOF
@@ -225,13 +132,53 @@ void HandleInputFile(std::string_view source) {
 
 		switch (lexer.token) {
 			case CLEX_id: {
-				std::string_view idenText(lexer.string, lexer.string_len);
-				if (idenText == "struct"sv || idenText == "class"sv) {
-					// TODO
-					matchingConstruct = NMC_StructClass;
-				} else if (idenText == "enum"sv) {
-					// TODO
-					matchingConstruct = NMC_Enum;
+				// WORKAROUND: stb_c_lexer doens't set string_len properly when parsing identifiers
+				std::string_view idenText(lexer.string);
+				// std::string_view idenText(lexer.string, lexer.string_len);
+				switch (matchingConstruct) {
+					case NMC_StructClass: {
+						if (matchingConstructInBody) {
+							// TODO
+						}
+					} break;
+
+					case NMC_Enum: {
+						if (matchingConstructInBody) {
+							printf("[DEBUG] found enum element '%s'\n", lexer.string);
+							currEnum.elements.push_back(InputDefinitionEnumElement{
+								.name = std::string(idenText),
+								.value = 0, // TODO parse
+							});
+						} else {
+							currEnum.name = std::string(idenText);
+						}
+					} break;
+
+					default: {
+						if (idenText == "struct"sv || idenText == "class"sv) {
+							printf("[DEBUG] found struct named\n");
+							matchingConstruct = NMC_StructClass;
+						} else if (idenText == "enum"sv) {
+							printf("[DEBUG] found enum\n");
+							matchingConstruct = NMC_Enum;
+						} else if (idenText == "BRUSSEL_CLASS"sv) {
+							// TODO
+							printf("[DEBUG] found BRUSSEL_CLASS\n");
+						} else if (idenText == "BRUSSEL_ENUM"sv) {
+							matchingDirectiveParams = true;
+							printf("[DEBUG] found BRUSSEL_ENUM\n");
+						}
+
+						if (matchingDirectiveParams) {
+							for (auto& foundEnum : foundEnums) {
+								if (foundEnum.name == idenText) {
+									// TODO generate data
+									break;
+								}
+							}
+							matchingDirectiveParams = false;
+						}
+					} break;
 				}
 			} break;
 
@@ -265,30 +212,49 @@ void HandleInputFile(std::string_view source) {
 
 			} break;
 
-			case CLEX_parse_error: {
-				fprintf(stderr, "[ERROR] stb_c_lexer countered a parse error.");
-				// TODO how to handle?
+			case '{': {
+				bracePairDepth++;
+				CheckBraceDepth(bracePairDepth);
+
+				switch (matchingConstruct) {
+					case NMC_StructClass:
+					case NMC_Enum: {
+						matchingConstructInBody = true;
+					} break;
+
+					default: break;
+				}
 			} break;
 
-			default: {
-				if (lexer.token >= 0 && lexer.token < 256) {
-					// Single char token
-					char c = lexer.token;
-					switch (c) {
-						case '{': {
-							bracePairDepth++;
-							CheckBraceDepth(bracePairDepth);
-						} break;
+			case '}': {
+				bracePairDepth--;
+				CheckBraceDepth(bracePairDepth);
+
+				switch (matchingConstruct) {
+					case NMC_StructClass: {
+						matchingConstruct = NMC_None;
+						matchingConstructInBody = false;
+					} break;
 
-						case '}': {
-							bracePairDepth--;
-							CheckBraceDepth(bracePairDepth);
-						} break;
-					}
-				} else {
-					fprintf(stderr, "[ERROR] Encountered unknown token %ld.", lexer.token);
+					case NMC_Enum: {
+						printf("[DEBUG] committed enum '%s'\n", currEnum.name.c_str());
+						for (auto& elm : currEnum.elements) {
+							printf("        - element %s = %" PRId64 "\n", elm.name.c_str(), elm.value);
+						}
+
+						matchingConstruct = NMC_None;
+						matchingConstructInBody = false;
+						PushFoundEnum();
+					} break;
+
+					default: break;
 				}
 			} break;
+
+			case CLEX_parse_error: {
+				fprintf(stderr, "[ERROR] stb_c_lexer countered a parse error.");
+				// TODO how to handle?
+			} break;
 		}
 	}
 
@@ -345,9 +311,6 @@ void HandleArgument(InputOpcode opcode, std::string_view operand) {
 }
 
 int main(int argc, char* argv[]) {
-	LUT_INIT(gClexTokens);
-	LUT_INIT(gSingleCharTokens);
-
 	// TODO better arg parser
 	//      option 1: use cxxopts and positional arguments
 	//      option 1: take one argument only, being a json objecet
@@ -359,7 +322,7 @@ int main(int argc, char* argv[]) {
 		auto separatorLoc = arg.find(':');
 		if (separatorLoc != std::string_view::npos) {
 			auto opcode = ParseInputOpcode(arg.substr(0, separatorLoc));
-			auto operand = arg.substr(separatorLoc);
+			auto operand = arg.substr(separatorLoc + 1);
 			HandleArgument(opcode, operand);
 		}
 	}