lonely-tiger/tokenizer/Tokenizer.c

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

#include "tokenizer/File.h"
#include "tokenizer/Tokenizer.h"
#include "utils/Utils.h"

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

static jmp_buf errorJump;
static char tokenBuffer[TOKEN_BUFFER_LENGTH];
static int writeIndex = 0;
static int readIndex = 0;
static int16 line = 1;
static char error[ERROR_LENGTH] = {'\0'};

static void tError(const char* format, ...) {
    va_list args;
    va_start(args, format);
    vsnprintf(error, ERROR_LENGTH, format, args);
    va_end(args);
    longjmp(errorJump, 0);
}

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

static void tAddToken(Token token) {
    unsigned char c = token;
    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 void tParseLiteral(int c) {
    int index = 1;
    char buffer[64];
    buffer[0] = c;
    while(isAllowedInName(fPeek())) {
        if(index >= 63) {
            tError("literal is too long");
        }
        buffer[index++] = fRead();
    }
    buffer[index] = '\0';
    Token t = tFromName(buffer);
    if(t != T_END) {
        tAddToken(t);
    } else {
        tAddToken(T_LITERAL);
        tAdd(buffer, index + 1);
    }
}

static bool tParseInt(char* s, int length, long long* l) {
    *l = 0;
    for(int i = 0; i < length; i++) {
        if(*l > (LLONG_MAX / 10)) {
            return true;
        }
        *l *= 10;
        int digit = s[i] - '0';
        if(*l > LLONG_MAX - digit) {
            return true;
        }
        *l += digit;
    }
    return false;
}

static void tParseNumber(int c) {
    int index = 1;
    char buffer[64];
    buffer[0] = c;
    bool point = false;
    while(true) {
        int c = fPeek();
        if(c == '.') {
            point = true;
        } else if(!isNumber(c)) {
            break;
        } else if(index >= 63) {
            tError("number is too long");
        }
        buffer[index++] = fRead();
    }
    buffer[index] = '\0';
    if(fPeek() == 'L' || fPeek() == 'l') {
        fRead();
        if(point) {
            tError("invalid mix of long and float", line);
        }
        long long l;
        if(tParseInt(buffer, index, &l) || l > INT64_MAX) {
            tError("invalid long on line %d", line);
        }
        int64 i = l;
        tAddToken(T_CONST_INT64);
        tAdd(&i, sizeof(int64));
    } else if(point) {
        char* end = NULL;
        float f = strtof(buffer, &end);
        if(end[0] != '\0') {
            tError("invalid float on line %d", line);
        }
        tAddToken(T_CONST_FLOAT);
        tAdd(&f, sizeof(float));
    } else {
        long long l;
        if(tParseInt(buffer, index, &l) || l > INT32_MAX) {
            tError("invalid int on line %d", line);
        }
        int32 i = l;
        tAddToken(T_CONST_INT32);
        tAdd(&i, sizeof(int32));
    }
}

static void tAddString() {
    tAddToken(T_TEXT);
    while(true) {
        int c = fRead();
        if(c == '"') {
            break;
        } else if(c == '\\') {
            switch(fRead()) {
                case '"': c = '"'; break;
                case '\\': c = '\\'; break;
                default: tError("unknown escaped character at line %d", line);
            }
        } else if(c == EOF) {
            tError("unclosed string starting at line %d", line);
        }
        tAdd(&c, 1);
    }
    char c = '\0';
    tAdd(&c, 1);
}

static void tAddUnicode() {
    int32 c = fRead();
    if((c & 0xE0) == 0xC0) {
        c = ((c & 0x1F) << 6) | (fRead() & 0x3F);
    } else if((c & 0xF0) == 0xE0) {
        c = ((c & 0xF) << 12) | ((fRead() & 0x3F) << 6);
        c |= fRead() & 0x3F;
    } else if((c & 0xF8) == 0xF0) {
        c = ((c & 0x7) << 18) | ((fRead() & 0x3F) << 12);
        c |= (fRead() & 0x3F) << 6;
        c |= fRead() & 0x3F;
    }
    tAddToken(T_CONST_INT32);
    tAdd(&c, sizeof(int32));
    if(fRead() != '\'') {
        tError("expecting unicode end");
    }
}

static void tAddToken2(Token te, Token t) {
    if(fReadIf('=')) {
        tAddToken(te);
    } else {
        tAddToken(t);
    }
}

static void tAddToken3(int c, Token tc, Token te, Token t) {
    if(fReadIf(c)) {
        tAddToken(tc);
    } else {
        tAddToken2(te, t);
    }
}

static void tAddToken4(int c, Token tce, Token tc, Token te, Token t) {
    if(fReadIf(c)) {
        tAddToken2(tce, tc);
    } else {
        tAddToken2(te, t);
    }
}

static void tAddTokenMinus() {
    tAddToken3('-', T_DECREMENT, T_SUB_SET, fReadIf('>') ? T_ARROW : T_SUB);
}

static void tLineComment() {
    while(true) {
        int c = fRead();
        if(c == EOF || c == '\n') {
            line++;
            return;
        }
    }
}

static void tMultipleLineComment() {
    while(true) {
        int c = fRead();
        if(c == EOF) {
            tError("unclosed comment at line %d", line);
        } else if(c == '\n') {
            line++;
        } else if(c == '*' && fReadIf('/')) {
            return;
        }
    }
}

static void tSlash() {
    if(fReadIf('/')) {
        tLineComment();
    } else if(fReadIf('*')) {
        tMultipleLineComment();
    } else {
        tAddToken2(T_DIV_SET, T_DIV);
    }
}

static void tParseToken(int c) {
    if(isLetter(c)) {
        tParseLiteral(c);
        return;
    } else if(isNumber(c)) {
        tParseNumber(c);
        return;
    }
    switch(c) {
        case ' ': return;
        case '\n': line++; return;
        case '+': tAddToken3('+', T_INCREMENT, T_ADD_SET, T_ADD); return;
        case '-': tAddTokenMinus(); return;
        case '*': tAddToken2(T_MUL_SET, T_MUL); return;
        case '/': tSlash(); return;
        case '%': tAddToken2(T_MOD_SET, T_MOD); return;
        case '<':
            tAddToken4('<', T_LEFT_SHIFT_SET, T_LEFT_SHIFT, T_LESS_EQUAL,
                       T_LESS);
            return;
        case '>':
            tAddToken4('>', T_RIGHT_SHIFT_SET, T_RIGHT_SHIFT, T_GREATER_EQUAL,
                       T_GREATER);
            return;
        case '=': tAddToken2(T_EQUAL, T_SET); return;
        case '!': tAddToken2(T_NOT_EQUAL, T_NOT); return;
        case '&': tAddToken3('&', T_AND, T_BIT_AND_SET, T_BIT_AND); return;
        case '|': tAddToken3('|', T_OR, T_BIT_OR_SET, T_BIT_OR); return;
        case '~': tAddToken(T_BIT_NOT); return;
        case '^': tAddToken2(T_BIT_XOR_SET, T_BIT_XOR); return;
        case ',': tAddToken(T_COMMA); return;
        case ';': tAddToken(T_SEMICOLON); return;
        case '(': tAddToken(T_OPEN_BRACKET); return;
        case ')': tAddToken(T_CLOSE_BRACKET); return;
        case '{': tAddToken(T_OPEN_CURVED_BRACKET); return;
        case '}': tAddToken(T_CLOSE_CURVED_BRACKET); return;
        case '"': tAddString(); return;
        case '\'': tAddUnicode(); return;
        case '.': tAddToken(T_POINT); return;
        case '[': tAddToken(T_OPEN_SQUARE_BRACKET); return;
        case ']': tAddToken(T_CLOSE_SQUARE_BRACKET); return;
    }
    tError("unknown character on line %d: %c", line, c);
}

static void tParseFile() {
    readIndex = 0;
    writeIndex = 0;
    line = 1;
    error[0] = '\0';
    while(true) {
        int c = fRead();
        if(c == EOF) {
            return;
        }
        tParseToken(c);
    }
}

bool tTokenize(const char* path) {
    if(!setjmp(errorJump)) {
        fOpen(path, tError);
        tParseFile();
    }
    fClose();
    return error[0] != '\0';
}

const char* tGetError() {
    return error;
}

int tGetLine() {
    return line;
}

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 tReadInt16(int16* i) {
    return tReadTokens(i, sizeof(int16));
}

bool tReadInt32(int32* i) {
    return tReadTokens(i, sizeof(int32));
}

bool tReadInt64(int64* i) {
    return tReadTokens(i, sizeof(int64));
}

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

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

int tGetMarker() {
    return readIndex;
}

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