New lexer 2 — Electric Boogaloo

CMakeLists.txt

@@ -46,7 +46,7 @@ if(MSVC)
   add_definitions(/D_CRT_SECURE_NO_WARNINGS)
 else()
   if(DEVELOP)
-    add_compile_options(-Werror -Wall -Wextra -pedantic
+    add_compile_options(-Werror -Wall -Wextra -pedantic -Wno-type-limits
                         -Wno-sign-compare -Wformat -Wformat-security -Wformat-overflow=2
                         -Wformat-truncation=1 -Wformat-y2k -Wswitch-enum -Wunused
                         -Wuninitialized -Wunknown-pragmas -Wstrict-overflow=5

Makefile

@@ -56,7 +56,6 @@ rgbasm_obj := \
 	src/asm/asmy.o \
 	src/asm/charmap.o \
 	src/asm/fstack.o \
-	src/asm/globlex.o \
 	src/asm/lexer.o \
 	src/asm/macro.o \
 	src/asm/main.o \
@@ -73,7 +72,7 @@ rgbasm_obj := \
 	src/hashmap.o \
 	src/linkdefs.o
 
-src/asm/globlex.o src/asm/lexer.o src/asm/constexpr.o: src/asm/asmy.h
+src/asm/lexer.o src/asm/main.o: src/asm/asmy.h
 
 rgblink_obj := \
 	src/link/assign.o \
@@ -187,7 +186,7 @@ checkpatch:
 # compilation and make the continous integration infrastructure return failure.
 
 develop:
-	$Qenv $(MAKE) -j WARNFLAGS="-Werror -Wall -Wextra -Wpedantic \
+	$Qenv $(MAKE) -j WARNFLAGS="-Werror -Wall -Wextra -Wpedantic -Wno-type-limits \
 		-Wno-sign-compare -Wformat -Wformat-security -Wformat-overflow=2 \
 		-Wformat-truncation=1 -Wformat-y2k -Wswitch-enum -Wunused \
 		-Wuninitialized -Wunknown-pragmas -Wstrict-overflow=5 \
@@ -199,7 +198,7 @@ develop:
 		-fsanitize=unreachable -fsanitize=vla-bound \
 		-fsanitize=signed-integer-overflow -fsanitize=bounds \
 		-fsanitize=object-size -fsanitize=bool -fsanitize=enum \
-		-fsanitize=alignment -fsanitize=null -DDEVELOP" CFLAGS="-g -O0"
+		-fsanitize=alignment -fsanitize=null -DDEVELOP" CFLAGS="-ggdb3 -O0"
 
 # Targets for the project maintainer to easily create Windows exes.
 # This is not for Windows users!

include/asm/asm.h

@@ -24,15 +24,8 @@
 #define MAXMACROARGS	99999
 #define MAXINCPATHS	128
 
-extern int32_t nLineNo;
 extern uint32_t nTotalLines;
 extern uint32_t nIFDepth;
-extern bool skipElif;
-extern char tzCurrentFileName[_MAX_PATH + 1];
 extern struct Section *pCurrentSection;
-extern bool oDontExpandStrings;
-
-size_t symvaluetostring(char *dest, size_t maxLength, char *sym,
-			const char *mode);
 
 #endif /* RGBDS_ASM_ASM_H */

include/asm/fstack.h

@@ -21,36 +21,59 @@
 
 #include "types.h"
 
-struct MacroArgs;
+struct FileStackNode {
+	struct FileStackNode *parent; /* Pointer to parent node, for error reporting */
+	/* Line at which the parent context was exited; meaningless for the root level */
+	uint32_t lineNo;
+
+	struct FileStackNode *next; /* Next node in the output linked list */
+	bool referenced; /* If referenced, don't free! */
+	uint32_t ID; /* Set only if referenced: ID within the object file, -1 if not output yet */
+
+	enum {
+		NODE_REPT,
+		NODE_FILE,
+		NODE_MACRO,
+	} type;
+};
 
-struct sContext {
-	YY_BUFFER_STATE FlexHandle;
-	struct Symbol const *pMacro;
-	struct sContext *next;
-	char tzFileName[_MAX_PATH + 1];
-	struct MacroArgs *macroArgs;
-	uint32_t uniqueID;
-	int32_t nLine;
-	uint32_t nStatus;
-	FILE *pFile;
-	char *pREPTBlock;
-	uint32_t nREPTBlockCount;
-	uint32_t nREPTBlockSize;
-	int32_t nREPTBodyFirstLine;
-	int32_t nREPTBodyLastLine;
+struct FileStackReptNode { /* NODE_REPT */
+	struct FileStackNode node;
+	uint32_t reptDepth;
+	/* WARNING: if changing this type, change overflow check in `fstk_Init` */
+	uint32_t iters[]; /* REPT iteration counts since last named node, in reverse depth order */
 };
 
-extern unsigned int nMaxRecursionDepth;
-
-void fstk_RunInclude(char *tzFileName);
-void fstk_Init(char *s);
-void fstk_Dump(void);
-void fstk_DumpToStr(char *buf, size_t len);
-void fstk_DumpStringExpansions(void);
-void fstk_AddIncludePath(char *s);
-void fstk_RunMacro(char *s, struct MacroArgs *args);
-void fstk_RunRept(uint32_t count, int32_t nReptLineNo);
-FILE *fstk_FindFile(char const *fname, char **incPathUsed);
-int32_t fstk_GetLine(void);
+struct FileStackNamedNode { /* NODE_FILE, NODE_MACRO */
+	struct FileStackNode node;
+	char name[]; /* File name for files, file::macro name for macros */
+};
+
+extern size_t nMaxRecursionDepth;
+
+struct MacroArgs;
+
+void fstk_Dump(struct FileStackNode const *node, uint32_t lineNo);
+void fstk_DumpCurrent(void);
+struct FileStackNode *fstk_GetFileStack(void);
+/* The lifetime of the returned chars is until reaching the end of that file */
+char const *fstk_GetFileName(void);
+
+void fstk_AddIncludePath(char const *s);
+/**
+ * @param path The user-provided file name
+ * @param fullPath The address of a pointer, which will be made to point at the full path
+ *                 The pointer's value must be a valid argument to `realloc`, including NULL
+ * @param size Current size of the buffer, or 0 if the pointer is NULL
+ * @return True if the file was found, false if no path worked
+ */
+bool fstk_FindFile(char const *path, char **fullPath, size_t *size);
+
+bool yywrap(void);
+void fstk_RunInclude(char const *path);
+void fstk_RunMacro(char const *macroName, struct MacroArgs *args);
+void fstk_RunRept(uint32_t count, int32_t nReptLineNo, char *body, size_t size);
+
+void fstk_Init(char const *mainPath, size_t maxRecursionDepth);
 
 #endif /* RGBDS_ASM_FSTACK_H */

include/asm/lexer.h

@@ -9,78 +9,65 @@
 #ifndef RGBDS_ASM_LEXER_H
 #define RGBDS_ASM_LEXER_H
 
-#include <stdint.h>
-#include <stdio.h>
-
-#define LEXHASHSIZE	(1 << 11)
 #define MAXSTRLEN	255
 
-struct sLexInitString {
-	char *tzName;
-	uint32_t nToken;
-};
+struct LexerState;
+extern struct LexerState *lexerState;
+extern struct LexerState *lexerStateEOL;
 
-struct sLexFloat {
-	uint32_t (*Callback)(char *s, uint32_t size);
-	uint32_t nToken;
-};
+static inline struct LexerState *lexer_GetState(void)
+{
+	return lexerState;
+}
 
-struct yy_buffer_state {
-	/* Actual starting address */
-	char *pBufferRealStart;
-	/* Address where the data is initially written after a safety margin */
-	char *pBufferStart;
-	char *pBuffer;
-	size_t nBufferSize;
-	uint32_t oAtLineStart;
-};
+static inline void lexer_SetState(struct LexerState *state)
+{
+	lexerState = state;
+}
 
-enum eLexerState {
-	LEX_STATE_NORMAL,
-	LEX_STATE_MACROARGS
-};
+static inline void lexer_SetStateAtEOL(struct LexerState *state)
+{
+	lexerStateEOL = state;
+}
 
-struct sStringExpansionPos {
-	char *tzName;
-	char *pBuffer;
-	char *pBufferPos;
-	struct sStringExpansionPos *pParent;
-};
+extern char const *binDigits;
+extern char const *gfxDigits;
 
-#define INITIAL		0
-#define macroarg	3
+static inline void lexer_SetBinDigits(char const *digits)
+{
+	binDigits = digits;
+}
 
-typedef struct yy_buffer_state *YY_BUFFER_STATE;
+static inline void lexer_SetGfxDigits(char const *digits)
+{
+	gfxDigits = digits;
+}
 
-void setup_lexer(void);
+/*
+ * `path` is referenced, but not held onto..!
+ */
+struct LexerState *lexer_OpenFile(char const *path);
+struct LexerState *lexer_OpenFileView(char *buf, size_t size, uint32_t lineNo);
+void lexer_RestartRept(uint32_t lineNo);
+void lexer_DeleteState(struct LexerState *state);
+void lexer_Init(void);
 
-void yy_set_state(enum eLexerState i);
-YY_BUFFER_STATE yy_create_buffer(FILE *f);
-YY_BUFFER_STATE yy_scan_bytes(char const *mem, uint32_t size);
-void yy_delete_buffer(YY_BUFFER_STATE buf);
-void yy_switch_to_buffer(YY_BUFFER_STATE buf);
-uint32_t lex_FloatAlloc(const struct sLexFloat *tok);
-void lex_FloatAddRange(uint32_t id, uint16_t start, uint16_t end);
-void lex_FloatDeleteRange(uint32_t id, uint16_t start, uint16_t end);
-void lex_FloatAddFirstRange(uint32_t id, uint16_t start, uint16_t end);
-void lex_FloatDeleteFirstRange(uint32_t id, uint16_t start, uint16_t end);
-void lex_FloatAddSecondRange(uint32_t id, uint16_t start, uint16_t end);
-void lex_FloatDeleteSecondRange(uint32_t id, uint16_t start, uint16_t end);
-void lex_Init(void);
-void lex_AddStrings(const struct sLexInitString *lex);
-void lex_SetBuffer(char *buffer, uint32_t len);
-void lex_BeginStringExpansion(const char *tzName);
-int yywrap(void);
-int yylex(void);
-void yyunput(char c);
-void yyunputstr(const char *s);
-void yyskipbytes(uint32_t count);
-void yyunputbytes(uint32_t count);
+enum LexerMode {
+	LEXER_NORMAL,
+	LEXER_RAW,
+	LEXER_SKIP_TO_ELIF,
+	LEXER_SKIP_TO_ENDC
+};
 
-extern YY_BUFFER_STATE pCurrentBuffer;
-extern struct sStringExpansionPos *pCurrentStringExpansion;
+void lexer_SetMode(enum LexerMode mode);
+void lexer_ToggleStringExpansion(bool enable);
 
-void upperstring(char *s);
-void lowerstring(char *s);
+char const *lexer_GetFileName(void);
+uint32_t lexer_GetLineNo(void);
+uint32_t lexer_GetColNo(void);
+void lexer_DumpStringExpansions(void);
+int yylex(void);
+void lexer_CaptureRept(char **capture, size_t *size);
+void lexer_CaptureMacroBody(char **capture, size_t *size);
 
 #endif /* RGBDS_ASM_LEXER_H */

include/asm/macro.h

@@ -28,6 +28,7 @@ char const *macro_GetArg(uint32_t i);
 uint32_t macro_GetUniqueID(void);
 char const *macro_GetUniqueIDStr(void);
 void macro_SetUniqueID(uint32_t id);
+uint32_t macro_UseNewUniqueID(void);
 void macro_ShiftCurrentArgs(void);
 uint32_t macro_NbArgs(void);
 

include/asm/main.h

@@ -43,6 +43,10 @@ void opt_Push(void);
 void opt_Pop(void);
 void opt_Parse(char *s);
 
+void upperstring(char *s);
+void lowerstring(char *s);
+
+/* TODO: are these really needed? */
 #define YY_FATAL_ERROR fatalerror
 
 #ifdef YYLMAX

include/asm/output.h

@@ -18,6 +18,8 @@ struct Expression;
 extern char *tzObjectname;
 extern struct Section *pSectionList, *pCurrentSection;
 
+void out_RegisterNode(struct FileStackNode *node);
+void out_ReplaceNode(struct FileStackNode *node);
 void out_SetFileName(char *s);
 void out_CreatePatch(uint32_t type, struct Expression const *expr,
 		     uint32_t ofs);

include/asm/symbol.h

@@ -35,18 +35,21 @@ struct Symbol {
 	bool isExported; /* Whether the symbol is to be exported */
 	bool isBuiltin;  /* Whether the symbol is a built-in */
 	struct Section *section;
-	char fileName[_MAX_PATH + 1]; /* File where the symbol was defined. */
-	uint32_t fileLine; /* Line where the symbol was defined. */
+	struct FileStackNode *src; /* Where the symbol was defined */
+	uint32_t fileLine; /* Line where the symbol was defined */
 
+	bool hasCallback;
 	union {
-		struct { /* If sym_IsNumeric */
-			int32_t value;
-			int32_t (*callback)(void);
-		};
-		struct { /* For SYM_MACRO */
-			uint32_t macroSize;
+		/* If sym_IsNumeric */
+		int32_t value;
+		int32_t (*numCallback)(void);
+		/* For SYM_MACRO */
+		struct {
+			size_t macroSize;
 			char *macro;
 		};
+		/* For SYM_EQUS, TODO: separate "base" fields from SYM_MACRO */
+		char const *(*strCallback)(void); /* For SYM_EQUS */
 	};
 
 	uint32_t ID; /* ID of the symbol in the object file (-1 if none) */
@@ -101,6 +104,8 @@ static inline bool sym_IsExported(struct Symbol const *sym)
  */
 static inline char const *sym_GetStringValue(struct Symbol const *sym)
 {
+	if (sym->hasCallback)
+		return sym->strCallback();
 	return sym->macro;
 }
 
@@ -114,9 +119,10 @@ void sym_Export(char const *symName);
 struct Symbol *sym_AddEqu(char const *symName, int32_t value);
 struct Symbol *sym_AddSet(char const *symName, int32_t value);
 uint32_t sym_GetPCValue(void);
+uint32_t sym_GetConstantSymValue(struct Symbol const *sym);
 uint32_t sym_GetConstantValue(char const *s);
 struct Symbol *sym_FindSymbol(char const *symName);
-struct Symbol *sym_AddMacro(char const *symName, int32_t defLineNo);
+struct Symbol *sym_AddMacro(char const *symName, int32_t defLineNo, char *body, size_t size);
 struct Symbol *sym_Ref(char const *symName);
 struct Symbol *sym_AddString(char const *symName, char const *value);
 uint32_t sym_GetDefinedValue(char const *s);

include/asm/util.h

@@ -12,6 +12,7 @@
 #include <stdint.h>
 
 uint32_t calchash(const char *s);
+char const *print(int c);
 size_t readUTF8Char(uint8_t *dest, char const *src);
 
 #endif /* RGBDS_UTIL_H */