cardiograph-computer/sketches/assembler.js

// Syntax:
// ADD $01    ; comments follow a `;`
// ADD $FF    ; this is direct addressing
// ADD ($CC)  ; this is indirect addressing
// END        ; END, CFC, and CHP don't require arguments
//            ; (a default value of 0 will be used as their operand)
// 
// @label     ; create a label
// JMP @label ; reference a label

const printMemory = require('./print-memory.js');

const { debug } = require("console");

// 0 = silent
// 1 = verbose
// 2 = what i'm currently focusing on
// 3 = always print
const DEBUG = 2;

const mnemonicsWithOptionalArgs = ['end', 'cfc', 'chp'];
const mnemonics2opcodes = {
  end: { direct: 0, indirect: 0 },
  sto: { direct: 1, indirect: 2 },
  lda: { direct: 3, indirect: 4 },
  add: { direct: 5, indirect: 6 },
  sub: { direct: 7, indirect: 8 },
  hop: { direct: 9, indirect: 10 },
  jmp: { direct: 11, indirect: 12 },
  cfc: { direct: 13, indirect: 13 },
  chp: { direct: 14, indirect: 14 },
};

function decodeInstructions(str) {
  let lines = str.split(/\n/); // returns an array of lines
  let machineCode = [];
  let labels = {};
  let byteCursor = 0;
  for (let i = 0; i < lines.length; i++) {
    console.log();
    console.group(`Input line ${i}, cursor ${byteCursor}`);
    dbg(3, `> ${lines[i]}`);
    let line = stripWhitespaceFromEnds(stripComments(lines[i]));
    // console.log();
    // console.log(`> ${line}`);

    // Handle blank lines
    if (line.length === 0) {
      dbg(3, `cursor: ${byteCursor}, new code: none`);
      dbg(1, 'blank');
      console.groupEnd('Input line');
      continue;
    }

    // Handle labels -- anchors
    if (line.startsWith('@')) {
      // TODO: validate label
      // validateLabel(line);
      label = line.substring(1); // strip '@'

      if (label in labels) {
        labels[label].pointsToByte = byteCursor;
      } else {
        labels[label] = {
          pointsToByte: byteCursor,
          bytesToReplace: [],
        };
      }

      dbg(2, '');
      dbg(2, `@label anchor: ${label}`);
      dbg(2, labels);
      dbg(2, '');
      dbg(3, `cursor: ${byteCursor}, new code: none`);
      console.groupEnd('Input line');
      continue;
    }

    let op_arg_array = line.split(" "); // split line into an array of [op, arg]
    let opName = op_arg_array[0].toLowerCase();
    let addressingMode = 'direct'; // Must be "direct" or "indirect"
    let arg_str = op_arg_array[1];
    let arg_num = null;

    if (typeof arg_str === 'undefined') {
      // Handle mnemonics without arguments (eg END) ...
      if (mnemonicsWithOptionalArgs.indexOf(opName) < 0) {
        console.error(`Missing opcode: ${line}`);
        throw new Error("Missing opcode");
      }
      arg_num = 0;
    } else if (arg_str.startsWith('@')) {
      // Handle mnemonics with pointers to labels
      // TODO: validate label // validateLabel(line);
      label = arg_str.substring(1); // strip '@'
      arg_num = 0;
      dbg(2, '');
      dbg(2, `@label reference: ${label}`);

      if (label in labels) {
        dbg(2, '');
        dbg(1, `'${label}' already in labels object`);
        labels[label].bytesToReplace.push(byteCursor + 1);
      } else {
        dbg(1, `'${label}' NOT in labels object`);
        labels[label] = {
          pointsToByte: byteCursor,
          bytesToReplace: [byteCursor + 1],
        };
      }
      // dbg(2, labels);
      dbg(2, `label pointsToByte: ${labels[label].pointsToByte}`);
      dbg(2, `label bytesToReplace: ${labels[label].bytesToReplace}`);

      dbg(2, '');
    } else if (arg_str.startsWith("(")) {
      // Handle indirect expressions
      addressingMode = "indirect";
      arg_str = arg_str.replace("(", "");
      arg_str = arg_str.replace(")", "");
      arg_num = parseInt(arg_str);
    } else if (arg_str.startsWith("$")) {
      // Handle direct expressions
      arg_str = arg_str.replace("$", "");
      arg_num = hex2num(arg_str);
    } else {
      // Accept decimal i guess
      arg_num = parseInt(arg_str);
    }

    // Decode!
    op = mnemonics2opcodes[opName][addressingMode];

    machineCode.push(op);
    machineCode.push(arg_num);
    byteCursor += 2;
    dbg(3, `cursor: ${byteCursor}, new code: ${op}, ${arg_num}`);
    console.groupEnd('Input line');
  };

  printMemory.printTable(machineCode);

  // Backfill label pointers
  for (let k of Object.keys(labels)) {
    let label = labels[k];
    dbg(2, label);
    for (let j = 0; j < label.bytesToReplace.length; j++) {
      machineCode[label.bytesToReplace[j]] = label.pointsToByte;
    }
  }

  return new Uint8Array(machineCode);
}


function stripComments(line) {
  return line.replace(/;.+/,"");
}

function stripWhitespaceFromEnds(line) {
  line = line.replace(/^\s+/,"");
  line = line.replace(/\s+$/,"");
  return line;
}

function hex2num(hex) { return parseInt(hex, 16) };

function dbg(debugLevel, string) {
  if (debugLevel >= DEBUG) console.log(string);
}

// RUN IT

exports.assemble = (str) => {
  return decodeInstructions(str);
}