assembler.c

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <stdint.h>
#include <ctype.h>

#define MAX_PROGRAM 256
#define MAX_LABELS  64
#define MAX_LINE    128

/* ===================== DATA STRUCTURES ===================== */

typedef struct {
    char name[32];
    uint8_t address;
} Label;

typedef struct {
    char mnemonic[4];
    uint8_t opcode;
    uint8_t size;
} Instruction;

/* ===================== GLOBALS ===================== */

Label labels[MAX_LABELS];
int label_count = 0;

uint8_t program[MAX_PROGRAM];
uint8_t pc = 0;

/* ===================== OPCODE TABLE ===================== */

Instruction table[] = {
    {"LDA", 0x01, 2},
    {"STA", 0x02, 2},
    {"ADD", 0x03, 2},
    {"SUB", 0x04, 2},
    {"JMP", 0x05, 2},
    {"JZ",  0x06, 2},
    {"OUT", 0x07, 1},
    {"HLT", 0xFF, 1}
};

int table_size = sizeof(table) / sizeof(table[0]);

/* ===================== UTILITIES ===================== */

Instruction* find_instruction(char *name) {
    for (int i = 0; i < table_size; i++)
        if (strcmp(table[i].mnemonic, name) == 0)
            return &table[i];
    return NULL;
}

int find_label(char *name) {
    for (int i = 0; i < label_count; i++)
        if (strcmp(labels[i].name, name) == 0)
            return labels[i].address;
    return -1;
}

void trim_comment(char *line) {
    for (int i = 0; line[i]; i++) {
        if (line[i] == ';') {
            line[i] = 0;
            return;
        }
    }
}

/* ===================== PASS 1 ===================== */
/* Find label addresses */

void pass1(FILE *file) {
    char line[MAX_LINE];
    pc = 0;

    while (fgets(line, sizeof(line), file)) {

        trim_comment(line);

        char *token = strtok(line, " \t\n");
        if (!token) continue;

        /* label */
        if (token[strlen(token) - 1] == ':') {
            token[strlen(token) - 1] = 0;
            strcpy(labels[label_count].name, token);
            labels[label_count].address = pc;
            label_count++;
            continue;
        }

        Instruction *inst = find_instruction(token);
        if (inst)
            pc += inst->size;
    }
}

/* ===================== PASS 2 ===================== */
/* Generate machine code */

void pass2(FILE *file) {
    char line[MAX_LINE];
    pc = 0;

    while (fgets(line, sizeof(line), file)) {

        trim_comment(line);

        char *token = strtok(line, " \t\n");
        if (!token) continue;

        if (token[strlen(token) - 1] == ':')
            continue;

        Instruction *inst = find_instruction(token);
        if (!inst) continue;

        program[pc++] = inst->opcode;

        if (inst->size == 2) {
            char *op = strtok(NULL, " \t\n");
            int value;

            if (isalpha(op[0])) {
                value = find_label(op);
                if (value < 0) {
                    printf("Unknown label: %s\n", op);
                    exit(1);
                }
            } else {
                value = strtol(op, NULL, 0);
            }

            program[pc++] = (uint8_t)value;
        }
    }
}

/* ===================== BINARY DISPLAY ===================== */

void print_binary(uint8_t byte) {
    for (int i = 7; i >= 0; i--)
        printf("%d", (byte >> i) & 1);
}

/* ===================== MAIN ===================== */

int main() {

    FILE *file = fopen("program.asm", "r");
    if (!file) {
        printf("ERROR: program.asm not found\n");
        return 1;
    }

    pass1(file);
    rewind(file);
    pass2(file);
    fclose(file);

    printf("\n===== MACHINE CODE OUTPUT =====\n\n");

    for (int i = 0; i < pc; i++) {
        printf("0x%02X  ", i);
        print_binary(program[i]);
        printf("\n");
    }

    printf("\nProgram size: %d bytes\n", pc);

    return 0;
}