Writing an HTML Interpreter

It was quite simple, to begin with. Instead of creating a standard recursive descent parser, I decided to take advantage of BASIC's 'line number-command list' structure, and create a simple regex to parse each line. I added a couple of checks to ensure that the command separator (the colon) wasn't mis-presented within quoted strings.

CodeLine.parse = function(str) {
  var split = str.match(/^\s*(\d+)\s+(.*)\s*$/m);
  if (split) {
    var line = new CodeLine(split[1]);
    var colon = split[2].split(':');
    
    for(var instr=0;instr<colon.length;++instr) {
      var numOfQuotes = (colon[instr].split('"').length - 1);
      if (numOfQuotes & 1 == 1) {  // there's an odd number quote, so we failed to correctly split
        // join this and the next text together to rebuild the string(BUGWARN)
        var cmd = colon[instr] + ":" + colon[instr+1];
        line.addInstruction(CodeInstruction.parse(cmd)).lineNumber = line.lineNumber;
        ++instr;
      } else {
        line.addInstruction(CodeInstruction.parse(colon[instr])).lineNumber = line.lineNumber;
      }
    }

  return line;
  }
  
  return null;
}

From here, I studied the syntax of each BASIC command to create one, or sometimes two, regexes to differentiate between them.

  if (args = str.match(/^goto\s+(\d+)/i)) {
    this.method = new cbGoto(args[1]);
    ...
  } else if (args = str.match(/^for\s+(\w+)\s*=(.*)\s*to\s*(.*)/i)) {
    this.method = new cbForLoop(args[1], args[2], args[3], 1);
  } else if (args = str.match(/^for\s+(\w+)\s*=(.*)\s*to\s*(.*)\s*step\s+(.*)/i)) {
    this.method = new cbForLoop(args[1], args[2], args[3], args[4]);
  ...

When combined, these steps build up a program parse tree, ready for evaluation. When each command is encountered a callback function is invoked with the current program state, and the operation performed. To handle the GOSUB library, I simply added a system-level callback at the appropriate line number to invoke the HTML5-specific code.

Et voila!

Well, nearly! I still need to properly parse the expressions... And that's where I am now. Awaiting some time to write/discover/incorporate a parser.