basic/src/Compiler.c

#include "Compiler.h"

#include <errno.h>
#include <setjmp.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#include "Code.h"
#include "Constants.h"

static Error error = {};
static FILE* file = nullptr;
static size_t lineCounter = 0;
static jmp_buf jumpPosition = {};
static u8 tokens[MAX_TOKENS] = {};
static size_t tokenReadIndex = 0;
static size_t tokenWriteIndex = 0;

typedef enum : u8 { LITERAL, INT64, DOUBLE, STRING, PLUS, NEWLINE, END } Token;

#define THROW_ERROR(...)                                   \
    snprintf(error.text, sizeof(error.text), __VA_ARGS__); \
    longjmp(jumpPosition, 1)

#define CODE(command)                                          \
    if(command) {                                              \
        THROW_ERROR("Code overflow on line %zu", lineCounter); \
    }

static void cleanup() {
    if(file != nullptr) {
        fclose(file);
        file = nullptr;
    }
}

static void reset() {
    error.text[0] = '\0';
    lineCounter = 0;
    tokenReadIndex = 0;
    tokenWriteIndex = 0;
}

static void addTokenN(const void* p, size_t n) {
    if(tokenWriteIndex + n > MAX_TOKENS) {
        THROW_ERROR("Line %zu has too much tokens", lineCounter);
    }
    memcpy(tokens + tokenWriteIndex, p, n);
    tokenWriteIndex += n;
}

static void addToken(Token t) {
    addTokenN(&t, sizeof(t));
}

static void addChar(char c) {
    addTokenN(&c, sizeof(c));
}

static bool isLetter(char c) {
    return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z');
}

static bool isNumber(char c) {
    return c >= '0' && c <= '9';
}

static bool isAlphaNumeric(char c) {
    return isLetter(c) || isNumber(c);
}

[[noreturn]] static void invalidToken(char c, int line) {
    THROW_ERROR("Unexpected token '%c' on line %zu %d", c, lineCounter, line);
}

static bool isTokenEnd(char c) {
    return c == ' ' || c == '\0' || c == '\n';
}

static const char* tokenizeLiteral(const char* s) {
    addToken(LITERAL);
    addChar(*s);
    while(true) {
        char c = *(++s);
        if(isAlphaNumeric(c)) {
            addChar(c);
        } else if(isTokenEnd(c)) {
            break;
        } else {
            invalidToken(c, __LINE__);
        }
    }
    addChar('\0');
    return s;
}

[[noreturn]] static void invalidNumber() {
    THROW_ERROR("Invalid number on line %zu", lineCounter);
}

static const char* tokenizeNumber(const char* s) {
    size_t nIndex = 0;
    char number[64] = {};
    number[nIndex++] = *s;
    while(true) {
        char c = *(++s);
        if(isTokenEnd(c)) {
            break;
        } else if(!isNumber(c) && c != '.') {
            invalidToken(c, __LINE__);
        } else if(nIndex >= sizeof(number) - 1) {
            invalidNumber();
        }
        number[nIndex++] = c;
    }
    char* end = nullptr;
    errno = 0;
    i64 i = strtoll(number, &end, 10);
    if(errno != 0) {
        invalidNumber();
    } else if(*end == '\0') {
        addToken(INT64);
        addTokenN(&i, sizeof(i));
        return s;
    }
    double d = strtod(number, &end);
    if(errno != 0) {
        invalidNumber();
    } else if(*end != '\0') {
        invalidNumber();
    }
    addToken(DOUBLE);
    addTokenN(&d, sizeof(d));
    return s;
}

static const char* tokenizeString(const char* s) {
    addToken(STRING);
    while(true) {
        char c = *(++s);
        if(c == '\0') {
            THROW_ERROR("Unclosed string on line %zu", lineCounter);
        } else if(c == '"') {
            s++;
            break;
        }
        addChar(c);
    }
    addChar('\0');
    return s;
}

static void tokenizeLineString(const char* s) {
    while(true) {
        char c = *s;
        if(isLetter(c)) {
            s = tokenizeLiteral(s);
        } else if(isNumber(c)) {
            s = tokenizeNumber(s);
        } else if(c == '"') {
            s = tokenizeString(s);
        } else if(c == '\n') {
            addToken(NEWLINE);
            break;
        } else if(c == ' ') {
            s++;
        } else if(c == '+') {
            addToken(PLUS);
            s++;
        } else if(c == '\0') {
            break;
        } else {
            invalidToken(c, __LINE__);
        }
    }
}

static void tokenizeLine() {
    tokenReadIndex = 0;
    tokenWriteIndex = 0;
    lineCounter++;
    char line[256] = {};
    if(fgets(line, sizeof(line), file) == nullptr) {
        return;
    }
    char c = line[sizeof(line) - 2];
    if(c != '\n' && c != '\0') {
        THROW_ERROR("Line %zu is too long", lineCounter);
    }
    tokenizeLineString(line);
}

static Token peekToken() {
    if(tokenReadIndex < tokenWriteIndex) {
        return tokens[tokenReadIndex];
    }
    tokenizeLine();
    return tokenReadIndex < tokenWriteIndex ? tokens[tokenReadIndex] : END;
}

static Token nextToken() {
    if(tokenReadIndex < tokenWriteIndex) {
        return tokens[tokenReadIndex++];
    }
    tokenizeLine();
    return tokenReadIndex < tokenWriteIndex ? tokens[tokenReadIndex++] : END;
}

static const char* peekLiteral() {
    Token t = peekToken();
    if(t != LITERAL) {
        return nullptr;
    }
    return (char*)(tokens + tokenReadIndex + sizeof(Token));
}

static const char* readString() {
    if(tokenReadIndex >= tokenWriteIndex) {
        THROW_ERROR("readString on empty buffer, line %zu", lineCounter);
    }
    const char* c = (char*)(tokens + tokenReadIndex);
    while(tokenReadIndex < tokenWriteIndex && tokens[tokenReadIndex] != '\0') {
        tokenReadIndex++;
    }
    tokenReadIndex++;
    return c;
}

static i64 readInt64() {
    if(tokenReadIndex + sizeof(i64) >= tokenWriteIndex) {
        THROW_ERROR("readInt64 on empty buffer, line %zu", lineCounter);
    }
    i64 i = 0;
    memcpy(&i, tokens + tokenReadIndex, sizeof(i));
    tokenReadIndex += sizeof(i);
    return i;
}

static double readDouble() {
    if(tokenReadIndex + sizeof(double) >= tokenWriteIndex) {
        THROW_ERROR("readDouble on empty buffer, line %zu", lineCounter);
    }
    double d = 0;
    memcpy(&d, tokens + tokenReadIndex, sizeof(d));
    tokenReadIndex += sizeof(d);
    return d;
}

[[noreturn]] static void unexpectedToken(Token t) {
    switch(t) {
        case LITERAL:
            THROW_ERROR(
                "Unexpected literal(%s) on line %zu", readString(),
                lineCounter);
            break;
        case INT64:
            THROW_ERROR(
                "Unexpected int(%ld) on line %zu", readInt64(), lineCounter);
            break;
        case DOUBLE:
            THROW_ERROR(
                "Unexpected double(%lf) on line %zu", readDouble(),
                lineCounter);
            break;
        case STRING:
            THROW_ERROR(
                "Unexpected string(%s) on line %zu", readString(), lineCounter);
            break;
        case PLUS:
            THROW_ERROR("Unexpected plus on line %zu", lineCounter);
            break;
        case NEWLINE:
            THROW_ERROR("Unexpected newline on line %zu", lineCounter);
            break;
        case END: THROW_ERROR("Unexpected end on line %zu", lineCounter); break;
    }
    THROW_ERROR("Unexpected unknown token on line %zu", lineCounter);
}

static void consumeToken(Token t) {
    Token actual = nextToken();
    if(t != actual) {
        unexpectedToken(actual);
    }
}

static void consumeLiteral(const char* name) {
    consumeToken(LITERAL);
    const char* actual = readString();
    if(strcmp(actual, name) != 0) {
        THROW_ERROR("Unexpected literal(%s) on line %zu", actual, lineCounter);
    }
}

static void compileConstant() {
    Token t = nextToken();
    if(t == STRING) {
        CODE(codePushInstruction(PUSH_CONSTANT_STRING));
        CODE(codePushConstantString(readString()));
    } else if(t == INT64) {
        CODE(codePushInstruction(PUSH_INT64));
        CODE(codePushI64(readInt64()));
    } else {
        unexpectedToken(t);
    }
}

static void compileAdd() {
    compileConstant();
    while(peekToken() == PLUS) {
        nextToken();
        compileConstant();
        CODE(codePushInstruction(ADD));
    }
}

static void compileExpression() {
    compileAdd();
}

static void compileLine(Token t);

static void compileIf() {
    compileExpression();
    CODE(codePushInstruction(JUMP_ON_0));
    size_t posIndex = codeGetWritePosition();
    CODE(codePushSize(0));
    consumeLiteral("then");
    consumeToken(NEWLINE);
    while(true) {
        const char* s = peekLiteral();
        if(s == nullptr || strcmp(s, "endif") != 0) {
            compileLine(nextToken());
            continue;
        } else {
            break;
        }
    }
    consumeLiteral("endif");
    consumeToken(NEWLINE);
    size_t endIndex = codeGetWritePosition();
    codeSetWritePosition(posIndex);
    CODE(codePushSize(endIndex));
    codeSetWritePosition(endIndex);
}

static void compileLine(Token t) {
    if(t == NEWLINE) {
        return;
    }
    if(t != LITERAL) {
        unexpectedToken(t);
    }
    const char* s = readString();
    if(strcmp(s, "print") == 0) {
        while(peekToken() != NEWLINE) {
            compileExpression();
            CODE(codePushInstruction(PRINT));
        }
        nextToken();
        CODE(codePushInstruction(PRINT_NEWLINE));
    } else if(strcmp(s, "if") == 0) {
        compileIf();
    } else {
        THROW_ERROR("Unexpected literal(%s) on line %zu", s, lineCounter);
    }
}

static void parseTokens() {
    codeReset();
    while(true) {
        Token t = nextToken();
        if(t == END) {
            break;
        }
        compileLine(t);
    }
}

const Error* compileFile(const char* path) {
    reset();
    if(setjmp(jumpPosition)) {
        cleanup();
        return &error;
    }
    file = fopen(path, "r");
    if(file == nullptr) {
        THROW_ERROR("Cannot read file '%s'", path);
    } else {
        parseTokens();
    }
    cleanup();
    return &error;
}