lonely-tiger/Tokenizer.c

#include <limits.h>
#include <stdarg.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#include "Tokenizer.h"
#include "Utils.h"

#define TOKEN_BUFFER_LENGTH (1024 * 1024)
#define ERROR_LENGTH 256

static char tokenBuffer[TOKEN_BUFFER_LENGTH];
static int writeIndex = 0;
static int readIndex = 0;
static int16 line = 1;
static FILE* file = NULL;
static char error[ERROR_LENGTH] = {'\0'};

typedef struct Literal {
    const char* name;
    Token token;
} Literal;

Literal LITERALS[] = {{"print", T_PRINT},       {"null", T_NULL},     {"true", T_TRUE},  {"false", T_FALSE},
                      {"function", T_FUNCTION}, {"return", T_RETURN}, {"if", T_IF},      {"else", T_ELSE},
                      {"while", T_WHILE},       {"for", T_FOR},       {"break", T_BREAK}};
const int LITERAL_AMOUNT = sizeof(LITERALS) / sizeof(Literal);

static void tError(const char* format, ...) {
    va_list args;
    va_start(args, format);
    vsnprintf(error, ERROR_LENGTH, format, args);
    va_end(args);
}

static bool tAdd(const void* data, int length) {
    if(writeIndex + length > TOKEN_BUFFER_LENGTH) {
        tError("the token buffer is too small");
        return false;
    }
    memcpy(tokenBuffer + writeIndex, data, length);
    writeIndex += length;
    return true;
}

static bool tAddToken(Token token) {
    unsigned char c = token;
    return tAdd(&c, 1) && tAdd(&line, sizeof(line));
}

static bool tReadTokens(void* dest, int length) {
    if(readIndex + length > writeIndex) {
        return false;
    }
    memcpy(dest, tokenBuffer + readIndex, length);
    readIndex += length;
    return true;
}

static int tRead() {
    return fgetc(file);
}

static int tPeek() {
    int c = tRead();
    ungetc(c, file);
    return c;
}

static int tReadIf(int c) {
    if(tPeek() == c) {
        tRead();
        return true;
    }
    return false;
}

static bool tParseLiteral(int c) {
    int index = 1;
    char buffer[64];
    buffer[0] = c;
    while(isLetter(tPeek())) {
        if(index >= 63) {
            tError("literal is too long");
            return false;
        }
        buffer[index++] = tRead();
    }
    buffer[index] = '\0';
    for(int i = 0; i < LITERAL_AMOUNT; i++) {
        if(strcmp(buffer, LITERALS[i].name) == 0) {
            return tAddToken(LITERALS[i].token);
        }
    }
    return tAddToken(T_LITERAL) && tAdd(buffer, index + 1);
}

static bool tParseNumber(int c) {
    int index = 1;
    char buffer[64];
    buffer[0] = c;
    bool point = false;
    while(true) {
        int c = tPeek();
        if(c == '.') {
            point = true;
        } else if(!isNumber(c)) {
            break;
        } else if(index >= 63) {
            tError("number is too long");
            return false;
        }
        buffer[index++] = tRead();
    }
    buffer[index] = '\0';
    if(point) {
        char* end = NULL;
        float f = strtof(buffer, &end);
        if(end[0] != '\0') {
            tError("invalid float on line %d", line);
            return false;
        }
        return tAddToken(T_FLOAT) && tAdd(&f, sizeof(float));
    } else {
        char* end = NULL;
        long l = strtol(buffer, &end, 10);
        if(end[0] != '\0' || l > INT_MAX) {
            tError("invalid int on line %d", line);
            return false;
        }
        int i = l;
        return tAddToken(T_INT) && tAdd(&i, sizeof(int));
    }
}

static bool tAddTokenChecked(int c, Token tc, Token te, Token t) {
    if(tReadIf(c)) {
        return tAddToken(tc);
    } else if(tReadIf('=')) {
        return tAddToken(te);
    }
    return tAddToken(t);
}

static bool tAddLongTokenChecked(int c, Token tce, Token tc, Token te, Token t) {
    if(tReadIf(c)) {
        if(tReadIf('=')) {
            return tAddToken(tce);
        } else {
            return tAddToken(tc);
        }
    } else if(tReadIf('=')) {
        return tAddToken(te);
    }
    return tAddToken(t);
}

static bool tParseToken() {
    int c = tRead();
    if(c == EOF) {
        return false;
    } else if(isLetter(c)) {
        return tParseLiteral(c);
    } else if(isNumber(c)) {
        return tParseNumber(c);
    }
    switch(c) {
        case ' ': return true;
        case '\n': line++; return true;
        case '+': return tAddTokenChecked('+', T_INCREMENT, T_ADD_SET, T_ADD);
        case '-': return tAddTokenChecked('-', T_DECREMENT, T_SUB_SET, T_SUB);
        case '*': return tReadIf('=') ? tAddToken(T_MUL_SET) : tAddToken(T_MUL);
        case '/': return tReadIf('=') ? tAddToken(T_DIV_SET) : tAddToken(T_DIV);
        case '%': return tReadIf('=') ? tAddToken(T_MOD_SET) : tAddToken(T_MOD);
        case '<': return tAddLongTokenChecked('<', T_LEFT_SHIFT_SET, T_LEFT_SHIFT, T_LESS_EQUAL, T_LESS);
        case '>': return tAddLongTokenChecked('>', T_RIGHT_SHIFT_SET, T_RIGHT_SHIFT, T_GREATER_EQUAL, T_GREATER);
        case '=': return tReadIf('=') ? tAddToken(T_EQUAL) : tAddToken(T_SET);
        case '!': return tReadIf('=') ? tAddToken(T_NOT_EQUAL) : tAddToken(T_NOT);
        case '&': return tAddTokenChecked('&', T_AND, T_BIT_AND_SET, T_BIT_AND);
        case '|': return tAddTokenChecked('|', T_OR, T_BIT_OR_SET, T_BIT_OR);
        case '~': return tAddToken(T_BIT_NOT);
        case '^': return tReadIf('=') ? tAddToken(T_BIT_XOR_SET) : tAddToken(T_BIT_XOR);
        case ',': return tAddToken(T_COMMA);
        case ';': return tAddToken(T_SEMICOLON);
        case '(': return tAddToken(T_OPEN_BRACKET);
        case ')': return tAddToken(T_CLOSE_BRACKET);
        case '{': return tAddToken(T_OPEN_CURVED_BRACKET);
        case '}': return tAddToken(T_CLOSE_CURVED_BRACKET);
    }
    tError("unknown character on line %d: %c", line, c);
    return false;
}

static void tParseFile() {
    readIndex = 0;
    writeIndex = 0;
    line = 1;
    error[0] = '\0';
    while(tParseToken()) {
    }
}

bool tTokenize(const char* path) {
    file = fopen(path, "r");
    if(file == NULL) {
        tError("cannot read file '%s'", path);
        return true;
    }
    tParseFile();
    fclose(file);
    return error[0] != '\0';
}

const char* tGetError() {
    return error;
}

void tResetReader() {
    readIndex = 0;
}

Token tPeekToken() {
    if(readIndex >= writeIndex) {
        return T_END;
    }
    return tokenBuffer[readIndex];
}

Token tReadToken() {
    if(readIndex >= writeIndex) {
        return T_END;
    }
    return tokenBuffer[readIndex++];
}

bool tReadInt(int* i) {
    if(tReadTokens(i, sizeof(int))) {
        return true;
    }
    return false;
}

bool tReadInt16(int16* i) {
    if(tReadTokens(i, sizeof(int16))) {
        return true;
    }
    return false;
}

bool tReadFloat(float* f) {
    if(tReadTokens(f, sizeof(float))) {
        return true;
    }
    return false;
}

const char* tReadString() {
    const char* s = tokenBuffer + readIndex;
    while(readIndex <= writeIndex) {
        if(tokenBuffer[readIndex] == '\0') {
            readIndex++;
            return s;
        }
        readIndex++;
    }
    return NULL;
}

const char* tGetTokenName(Token token) {
    switch(token) {
        case T_INT: return "int";
        case T_FLOAT: return "float";
        case T_NULL: return "null";
        case T_TRUE: return "true";
        case T_FALSE: return "false";
        case T_ADD: return "+";
        case T_SUB: return "-";
        case T_MUL: return "*";
        case T_DIV: return "/";
        case T_MOD: return "%";
        case T_LESS: return "<";
        case T_LESS_EQUAL: return "<=";
        case T_GREATER: return ">";
        case T_GREATER_EQUAL: return ">=";
        case T_EQUAL: return "==";
        case T_NOT_EQUAL: return "!=";
        case T_NOT: return "!";
        case T_AND: return "&&";
        case T_OR: return "||";
        case T_BIT_NOT: return "~";
        case T_BIT_AND: return "&";
        case T_BIT_OR: return "|";
        case T_BIT_XOR: return "^";
        case T_LEFT_SHIFT: return "<<";
        case T_RIGHT_SHIFT: return ">>";
        case T_SET: return "=";
        case T_ADD_SET: return "+=";
        case T_SUB_SET: return "-=";
        case T_MUL_SET: return "*=";
        case T_DIV_SET: return "/=";
        case T_MOD_SET: return "%=";
        case T_BIT_AND_SET: return "&=";
        case T_BIT_OR_SET: return "|=";
        case T_BIT_XOR_SET: return "^=";
        case T_LEFT_SHIFT_SET: return "<<=";
        case T_RIGHT_SHIFT_SET: return ">>=";
        case T_INCREMENT: return "++";
        case T_DECREMENT: return "--";
        case T_LITERAL: return "literal";
        case T_PRINT: return "print";
        case T_IF: return "if";
        case T_ELSE: return "else";
        case T_WHILE: return "while";
        case T_FOR: return "for";
        case T_BREAK: return "break";
        case T_FUNCTION: return "function";
        case T_RETURN: return "return";
        case T_COMMA: return ",";
        case T_SEMICOLON: return ";";
        case T_OPEN_BRACKET: return "(";
        case T_CLOSE_BRACKET: return ")";
        case T_OPEN_CURVED_BRACKET: return "{";
        case T_CLOSE_CURVED_BRACKET: return "}";
        case T_END: return "end";
    }
    return "Unknown";
}

int tGetMarker() {
    return readIndex;
}

void tResetToMarker(int marker) {
    readIndex = marker;
}

void tPrint() {
    puts("----------------");
    while(true) {
        Token t = tReadToken();
        if(t == T_END) {
            break;
        }
        int line = 0;
        tReadInt(&line);
        printf("%d: %s\n", line, tGetTokenName(t));
        if(t == T_INT) {
            tReadInt(&line);
        }
    }
    tResetReader();
}