// I used: http://devernay.free.fr/hacks/chip8/C8TECH10.HTM#00E0
// and https://en.wikipedia.org/wiki/CHIP-8#Opcode_table
// for information about how to implement a CHIP-8 Emulator
use std::fs::File;
use std::io::{self, Read};
use std::path::Path;
use std::time::Duration;

const SCREEN_SIZE: usize = 64 * 32;

#[derive(Copy, Clone)]
struct Chip8 {
    opcode: u16,
    i: u16,
    pc: u16,
    sp: u8,
    key: Option<u8>,
    v: [u8; 16],
    stack: [u16; 16],
    memory: [u8; 4096],
    delay: Timer,
    sound: Timer,
    display: Display,
}

impl Default for Chip8 {
    fn default() -> Self {
        Chip8 {
            opcode: 0,
            i: 0,
            pc: 0x200, // Progrm counter starts at 0x200
            sp: 0,
            key: None,
            v: [0; 16],
            stack: [0; 16],
            memory: [0; 4096],
            delay: Default::default(),
            sound: Default::default(),
            display: Display::default(),
        }
    }
}

impl Chip8 {
    pub fn execute_cycle(&mut self) {
        self.opcode = self.get_opcode();
        println!("{:#x}", self.opcode);
        self.handle_opcode();

        self.delay.tick();
        self.sound.tick();
    }

    pub fn load_rom<P: AsRef<Path>>(&mut self, path: P) -> Result<(), io::Error> {
        let mut file = File::open(path.as_ref())?;
        let mut rom_buf: Vec<u8> = vec![];
        file.read_to_end(&mut rom_buf)?;

        for (i, byte) in rom_buf.iter().enumerate() {
            self.memory[i + 0x200] = *byte;
        }

        Ok(())
    }

    fn get_opcode(&self) -> u16 {
        let pc = self.pc as usize;

        ((self.memory[pc] as u16) << 8) | self.memory[pc + 1] as u16
    }

    fn handle_opcode(&mut self) {
        // Given: 0xA2F0
        let nib_1 = (self.opcode & 0xF000) >> 12; //0xA
        let nib_2 = (self.opcode & 0x0F00) >> 8; // 0x2
        let nib_3 = (self.opcode & 0x00F0) >> 4; // 0xF
        let nib_4 = self.opcode & 0x000F; // 0x0

        // nib_ns are u16s so we waste 4 bytes here.

        match (nib_1, nib_2, nib_3, nib_4) {
            // CLS
            (0x0, 0x0, 0xE, 0x0) => self.cls(),
            // 00EE
            (0x0, 0x0, 0xE, 0xE) => self.ret(),
            // 1NNN
            (0x1, _, _, _) => self.jmp_addr(nib_2, nib_3, nib_4),
            // 2NNN
            (0x2, _, _, _) => self.call_addr(nib_2, nib_3, nib_4),
            // 3XKK
            (0x3, _, _, _) => self.se_vx_byte(nib_2, nib_3, nib_4),
            // 4xkk
            (0x4, _, _, _) => self.sne_vx_byte(nib_2, nib_3, nib_4),
            // 5xy0
            (0x5, _, _, 0x0) => self.se_vx_vy(nib_2, nib_3),
            // 6xkk
            (0x6, _, _, _) => self.ld_vx_byte(nib_2, nib_3, nib_4),
            // 7xkk
            (0x7, _, _, _) => self.add_vx_byte(nib_2, nib_3, nib_4),
            // 8xy0
            (0x8, _, _, 0x0) => self.ld_vx_vy(nib_2, nib_3),
            // 8xy1
            (0x8, _, _, 0x1) => self.or_vx_vy(nib_2, nib_3),
            // 8xy2
            (0x8, _, _, 0x2) => self.and_vx_vy(nib_2, nib_3),
            // 8xy3
            (0x8, _, _, 0x3) => self.xor_vx_vy(nib_2, nib_3),
            // 8xy4
            (0x8, _, _, 0x4) => self.add_vx_vy(nib_2, nib_3),
            // 8xy5
            (0x8, _, _, 0x5) => self.sub_vx_vy(nib_2, nib_3),
            // 8xy6
            (0x8, _, _, 0x6) => self.shr_vx(nib_2),
            // 8xy7
            (0x8, _, _, 0x7) => self.subn_vx_vy(nib_2, nib_3),
            // 8xyE
            (0x8, _, _, 0xE) => self.shl_vx(nib_2),
            // 9xy0
            (0x9, _, _, 0x0) => self.sne_vx_vy(nib_2, nib_3),
            // Annn
            (0xA, _, _, _) => self.ld_i_addr(nib_2, nib_3, nib_4),
            // Bnnn
            (0xB, _, _, _) => self.jmp_v0_addr(nib_2, nib_3, nib_4),
            // Cxkk
            (0xC, _, _, _) => self.rnd_vx_byte(nib_2, nib_3, nib_4),
            //Dxyn
            (0xD, _, _, _) => self.drw_vx_vy_nib(nib_2, nib_3, nib_4),
            // Ex9E
            (0xE, _, 0x9, 0xE) => self.skp_vx(nib_2),
            // ExA1
            (0xE, _, 0xA, 0x1) => self.sknp_vx(nib_2),
            // Fx07
            (0xF, _, 0x0, 0x7) => self.ld_vx_dt(nib_2),
            // Fx0A
            (0xF, _, 0x0, 0xA) => self.ld_vx_k(nib_2),
            // Fx15
            (0xF, _, 0x1, 0x15) => self.ld_dt_vx(nib_2),
            // Fx18
            (0xF, _, 0x1, 0x8) => self.ld_st_vx(nib_2),
            // Fx1E
            (0xF, _, 0x1, 0xE) => self.add_i_vx(nib_2),
            // Fx29
            (0xF, _, 0x2, 0x9) => self.ld_f_vx(nib_2),
            // Fx33
            (0xF, _, 0x3, 0x3) => self.ld_b_vx(nib_2),
            // Fx55
            (0xF, _, 0x5, 0x5) => self.ld_i_vx(nib_2),
            // Fx65
            (0xF, _, 0x6, 0x5) => self.ld_vx_i(nib_2),
            // General Case
            _ => println!("UNIMPLEMENTED OPCODE: {:#x}", self.opcode),
        }
    }

    fn cls(&mut self) {
        // Clear the display
        self.display.clear();
        self.pc += 1;
    }

    fn jmp_addr(&mut self, n_1: u16, n_2: u16, n_3: u16) {
        // sets the program counter to addr (nnn)
        self.pc = Self::convert_to_addr(n_1, n_2, n_3);
        self.pc += 1;
    }

    fn ret(&mut self) {
        // sets the program counter to the addres at the top of the stack,
        // then subtracts one (1) from the stack pointer
        self.pc = self.stack[self.sp as usize];
        self.sp -= 1;
        self.pc += 1;
    }

    fn call_addr(&mut self, n_1: u16, n_2: u16, n_3: u16) {
        // increments the stack pointer, then puts current pc on top of stack
        // pc is then set to addr
        self.sp += 1;
        self.stack[self.sp as usize] = self.pc;
        self.pc = Self::convert_to_addr(n_1, n_2, n_3);
        self.pc += 1;
    }

    fn se_vx_byte(&mut self, x: u16, k_1: u16, k_2: u16) {
        // compares Vx to kk. If they are equal, pc is incremented by 2
        if self.v[x as usize] == Self::convert_to_byte(k_1, k_2) {
            self.pc += 2;
        } else {
            self.pc += 1;
        }
    }

    fn sne_vx_byte(&mut self, x: u16, k_1: u16, k_2: u16) {
        // compares Vx to kk. If they are **not** equal, pc is incremented by 2
        if self.v[x as usize] != Self::convert_to_byte(k_1, k_2) {
            self.pc += 2;
        } else {
            self.pc += 1;
        }
    }

    fn se_vx_vy(&mut self, x: u16, y: u16) {
        // compares Vx to Vy. If they are equal, pc is incremented by 2
        if self.v[x as usize] == self.v[y as usize] {
            self.pc += 2;
        } else {
            self.pc += 1;
        }
    }

    fn ld_vx_byte(&mut self, x: u16, k_1: u16, k_2: u16) {
        // put value kk into Vx
        self.v[x as usize] = Self::convert_to_byte(k_1, k_2);
        self.pc += 1;
    }

    fn add_vx_byte(&mut self, x: u16, k_1: u16, k_2: u16) {
        // calculate Vx + kk, then store it in Vx
        let x = x as usize;
        self.v[x] = self.v[x] + Self::convert_to_byte(k_1, k_2);
        self.pc += 1;
    }

    fn ld_vx_vy(&mut self, x: u16, y: u16) {
        // store Vy in Vx
        self.v[x as usize] = self.v[y as usize];
        self.pc += 1;
    }

    fn or_vx_vy(&mut self, x: u16, y: u16) {
        // calc bitwise OR on Vx and Vy, then store in Vx
        let x = x as usize;
        self.v[x] = self.v[x] | self.v[y as usize];
        self.pc += 1;
    }

    fn and_vx_vy(&mut self, x: u16, y: u16) {
        // calc bitwise AND on Vx and Vy, then store in Vx
        let x = x as usize;
        self.v[x] = self.v[x] & self.v[y as usize];
        self.pc += 1;
    }

    fn xor_vx_vy(&mut self, x: u16, y: u16) {
        // calc bitwise XOR on Vx and Vy, then store in Vx
        let x = x as usize;
        self.v[x] = self.v[x] ^ self.v[y as usize];
        self.pc += 1;
    }

    fn add_vx_vy(&mut self, x: u16, y: u16) {
        // add Vx and Vy, if result is greater than 8 bits
        // set VF to 1, otherwise 0
        // only the lowest 8 bits of result are stored in Vx
        let x = x as usize;

        let (res, did_overflow) = self.v[x].overflowing_add(self.v[y as usize]);
        self.v[0xF as usize] = if did_overflow { 1 } else { 0 };
        self.v[x] = res;
        self.pc += 1;
    }

    fn sub_vx_vy(&mut self, x: u16, y: u16) {
        // subtract Vx and Vy, if Vx > Vy VF is set to 1, otherwise 0
        // then set Vx to Vx - Vy
        let vx = self.v[x as usize];
        let vy = self.v[y as usize];

        self.v[0xF as usize] = if vx > vy { 1 } else { 0 };
        self.v[x as usize] = vx - vy;
        self.pc += 1;
    }

    fn shr_vx(&mut self, x: u16) {
        // if LSB is set to 1, set VF to 1, otherwise set VF to 0
        // then shift Vx one to the right
        let x = x as usize;

        self.v[0xF as usize] = if (self.v[x] & 1) == 1 { 1 } else { 0 };
        self.v[x] = self.v[x] >> 1;
        self.pc += 1;
    }

    fn subn_vx_vy(&mut self, x: u16, y: u16) {
        // subtract Vy and Vx, if Vy > Vx VF is set to 1, otherwise 0
        // then set Vx = Vy - Vx
        let vx = self.v[x as usize];
        let vy = self.v[y as usize];

        self.v[0xF as usize] = if vy > vx { 1 } else { 0 };
        self.v[x as usize] = vy - vx;
        self.pc += 1;
    }

    fn shl_vx(&mut self, x: u16) {
        // if MSB is set to 1, set VF to 1, otherwise set VF to 0
        // then shift Vx one to the left;
        let x = x as usize;

        self.v[0xF as usize] = if (self.v[x] & 0x80) == 1 { 1 } else { 0 };
        self.v[x] = self.v[x] << 1;
        self.pc += 1;
    }

    fn sne_vx_vy(&mut self, x: u16, y: u16) {
        // if Vx != vy program counter is increased by 2
        if self.v[x as usize] != self.v[y as usize] {
            self.pc += 2;
        } else {
            self.pc += 1;
        }
    }

    fn ld_i_addr(&mut self, n_1: u16, n_2: u16, n_3: u16) {
        // set i to addr
        self.i = Self::convert_to_addr(n_1, n_2, n_3);
        self.pc += 1;
    }

    fn jmp_v0_addr(&mut self, n_1: u16, n_2: u16, n_3: u16) {
        // set program counter to addr + V0
        self.pc = Self::convert_to_addr(n_1, n_2, n_3) + self.v[0 as usize] as u16;
        self.pc += 1;
    }

    fn rnd_vx_byte(&mut self, x: u16, k_1: u16, k_2: u16) {
        // generate a random number from 0 to 255
        // AND with the value of kk, then store in Vx

        self.v[x as usize] = rand::random::<u8>() & Self::convert_to_byte(k_1, k_2);
        self.pc += 1;
    }

    fn drw_vx_vy_nib(&mut self, x: u16, y: u16, nib: u16) {
        // read n bytes from memory starting from self.i
        // then display them starting at (vx, vy)
        let i = self.i as usize;
        let _disp_x = self.v[x as usize];
        let _disp_y = self.v[y as usize];

        // There's Probably an off-by-one error here.
        let _sprite_buf = &self.memory[i..=(i + nib as usize)];
        let _display_buf = self.display.buf;

        todo!("Finish implementing DRW");
    }

    fn skp_vx(&mut self, x: u16) {
        // if current key is the same as the on in Vx
        // program counter is increased by 2
        if let Some(key) = self.key {
            if self.v[x as usize] == key {
                self.pc += 2;
            } else {
                self.pc += 1;
            }
        } else {
            self.pc += 1;
        }
    }

    fn sknp_vx(&mut self, x: u16) {
        // if current key is not the sameas the one in Vx
        // increment the program counter by 2
        match self.key {
            Some(key) => {
                if key != self.v[x as usize] {
                    self.pc += 2;
                } else {
                    self.pc += 1;
                }
            }
            None => self.pc += 2,
        }
    }

    fn ld_vx_dt(&mut self, x: u16) {
        // set Vx to be the value of the delay timer
        self.v[x as usize] = self.delay.get();
        self.pc += 1;
    }

    fn ld_vx_k(&mut self, x: u16) {
        // wait (blocking) until a key is pressed
        // once pressed, store in Vx
        loop {
            if let Some(key) = self.key {
                self.v[x as usize] = key;
                break;
            }
        }
        self.pc += 1;
    }

    fn ld_dt_vx(&mut self, x: u16) {
        // set delay timer to be value of Vx
        self.delay.set(self.v[x as usize]);
        self.pc += 1;
    }

    fn ld_st_vx(&mut self, x: u16) {
        // set sound timer to be value of Vx
        self.delay.set(self.v[x as usize]);
        self.pc += 1;
    }

    fn add_i_vx(&mut self, x: u16) {
        // set I to be I + Vx
        self.i = self.i + self.v[x as usize] as u16;
        self.pc += 1;
    }

    fn ld_f_vx(&mut self, _x: u16) {
        todo!("Implement 0xF_29");
    }

    fn ld_b_vx(&mut self, x: u16) {
        // take hundreds digit and place it at I
        // take tens digit and place it at I + 1
        // take ones digit and place it at I + 2
        let i = self.i as usize;
        let mut iter = Self::digits(self.v[x as usize]);

        let ones = iter.next().unwrap(); // Ther has to at least be a ones lol
        let tens = iter.next().unwrap_or(0);
        let hundreds = iter.next().unwrap_or(0);

        self.memory[i] = hundreds;
        self.memory[i + 1] = tens;
        self.memory[i + 2] = ones;
        self.pc += 1;
    }

    fn ld_i_vx(&mut self, x: u16) {
        // copy values v0 -> Vx to memory starting at i

        for n in 0..=(x as usize) {
            self.memory[self.i as usize + n] = self.v[n];
        }
        self.pc += 1;
    }

    fn ld_vx_i(&mut self, x: u16) {
        // read what will be values of v0 -> vx from memory starting at i
        let x = x as usize;

        for n in 0..=x {
            self.v[n] = self.memory[self.i as usize + n];
        }
        self.pc += 1;
    }

    fn convert_to_byte(k_1: u16, k_2: u16) -> u8 {
        ((k_1 as u8) << 4) | (k_2 as u8)
    }

    fn convert_to_addr(n_1: u16, n_2: u16, n_3: u16) -> u16 {
        (n_1 << 8) | (n_2 << 4) | n_3
    }

    // https://stackoverflow.com/questions/41536479/how-do-i-split-an-integer-into-individual-digits
    fn digits(mut num: u8) -> impl Iterator<Item = u8> {
        let mut divisor = 1;
        while num >= divisor * 10 {
            divisor *= 10;
        }

        std::iter::from_fn(move || {
            if divisor == 0 {
                None
            } else {
                let v = num / divisor;
                num %= divisor;
                divisor /= 10;
                Some(v)
            }
        })
    }
}

#[derive(Copy, Clone)]
struct Display {
    buf: [u8; SCREEN_SIZE],
}

impl Display {
    pub fn clear(&mut self) {
        self.buf = [0; SCREEN_SIZE]
    }
}

impl Default for Display {
    fn default() -> Self {
        Display {
            buf: [0; SCREEN_SIZE],
        }
    }
}

#[derive(Debug, Copy, Clone)]
struct Timer {
    remaining: u8,
    enabled: bool,
}

impl Default for Timer {
    fn default() -> Self {
        Timer {
            remaining: 0,
            enabled: false,
        }
    }
}

impl Timer {
    pub fn set(&mut self, secs: u8) {
        self.remaining = secs;
        self.enabled = true;
    }

    pub fn get(&self) -> u8 {
        self.remaining
    }

    pub fn tick(&mut self) {
        if self.enabled {
            self.remaining -= 1;

            if self.remaining == 0 {
                println!("Beep!");
                self.enabled = false;
            }
        }
    }
}

/// Remember that the digits are yielded in the oposite order
#[derive(Debug, Copy, Clone)]
pub struct Digits {
    num: u8,
    flag: bool,
}

impl From<u8> for Digits {
    fn from(num: u8) -> Self {
        Digits { num, flag: true }
    }
}

impl Iterator for Digits {
    type Item = u8;

    fn next(&mut self) -> Option<u8> {
        if self.num >= 10 {
            // There is something next
            let tmp = self.num;
            self.num = self.num / 10;

            Some(tmp % 10)
        } else {
            if self.flag {
                self.flag = false;
                Some(self.num % 10)
            } else {
                None
            }
        }
    }
}

fn main() {
    let mut chip8: Chip8 = Default::default();

    chip8
        .load_rom(Path::new("C:\\Users\\paoda\\Documents\\chip8_roms\\TICTAC"))
        .expect("Unable to load ROM");

    loop {
        chip8.execute_cycle();

        std::thread::sleep(Duration::from_millis(1000));
    }
}