// A Small Compiler for the WHILE Language

//

//  - includes arrays and a small parser for

//    WHILE programs

//

//  - transpiles BF programs into WHILE programs

//    and compiles and runs them

//

// Call with

//

// scala compile_arr.scala

// Mandelbrot

// mand.j size       236241

// mand.class size   33296

// running time      21 secs

// the abstract syntax trees

abstract class Stmt

abstract class AExp

abstract class BExp 

type Block = List[Stmt]

// statements

case object Skip extends Stmt

case class ArrayDef(s: String, n: Int) extends Stmt

case class If(a: BExp, bl1: Block, bl2: Block) extends Stmt

case class While(b: BExp, bl: Block) extends Stmt

case class Assign(s: String, a: AExp) extends Stmt             // var := exp

case class AssignA(s: String, a1: AExp, a2: AExp) extends Stmt // arr[exp1] := exp2

case class Write(s: String) extends Stmt

case class Read(s: String) extends Stmt

// arithmetic expressions

case class Var(s: String) extends AExp

case class Num(i: Int) extends AExp

case class Aop(o: String, a1: AExp, a2: AExp) extends AExp

case class Ref(s: String, a: AExp) extends AExp

// boolean expressions

case object True extends BExp

case object False extends BExp

case class Bop(o: String, a1: AExp, a2: AExp) extends BExp

// compiler headers needed for the JVM

// (contains an init method, as well as methods for read and write)

val beginning = """

.class public XXX.XXX

.super java/lang/Object

.method public static write(I)V 

    .limit locals 1 

    .limit stack 2 

    getstatic java/lang/System/out Ljava/io/PrintStream; 

    iload 0

    i2c       ; Int => Char

    invokevirtual java/io/PrintStream/print(C)V   ; println(I)V => print(C)V    

    return 

.end method

.method public static main([Ljava/lang/String;)V

   .limit locals 200

   .limit stack 200

; COMPILED CODE STARTS   

"""

val ending = """

; COMPILED CODE ENDS

   return

.end method

"""

// for generating new labels

var counter = -1

def Fresh(x: String) = {

  counter += 1

  x ++ "_" ++ counter.toString()

}

// environments and instructions

type Env = Map[String, Int]

// convenient string interpolations 

// for instructions and labels

import scala.language.implicitConversions

import scala.language.reflectiveCalls

implicit def sring_inters(sc: StringContext) = new {

    def i(args: Any*): String = "   " ++ sc.s(args:_*) ++ "\n"

    def l(args: Any*): String = sc.s(args:_*) ++ ":\n"

}

def compile_op(op: String) = op match {

  case "+" => i"iadd"

  case "-" => i"isub"

  case "*" => i"imul"

}

// arithmetic expression compilation

def compile_aexp(a: AExp, env : Env) : String = a match {

  case Num(i) => i"ldc $i"

  case Var(s) => i"iload ${env(s)}"

  case Aop(op, a1, a2) => 

    compile_aexp(a1, env) ++ compile_aexp(a2, env) ++ compile_op(op)

  case Ref(s, a) =>

    i"aload ${env(s)}" ++ compile_aexp(a, env) ++  i"iaload"

}

// boolean expression compilation

def compile_bexp(b: BExp, env : Env, jmp: String) : String = b match {

  case True => ""

  case False => i"goto $jmp"

  case Bop("==", a1, a2) => 

    compile_aexp(a1, env) ++ compile_aexp(a2, env) ++ i"if_icmpne $jmp"

  case Bop("!=", a1, a2) => 

    compile_aexp(a1, env) ++ compile_aexp(a2, env) ++ i"if_icmpeq $jmp"

  case Bop("<", a1, a2) => 

    compile_aexp(a1, env) ++ compile_aexp(a2, env) ++ i"if_icmpge $jmp"

}

// statement compilation

def compile_stmt(s: Stmt, env: Env) : (String, Env) = s match {

  case Skip => ("", env)

  case Assign(x, a) => {

     val index = env.getOrElse(x, env.keys.size)

    (compile_aexp(a, env) ++ i"istore $index \t\t; $x", env + (x -> index)) 

  } 

  case If(b, bl1, bl2) => {

    val if_else = Fresh("If_else")

    val if_end = Fresh("If_end")

    val (instrs1, env1) = compile_block(bl1, env)

    val (instrs2, env2) = compile_block(bl2, env1)

    (compile_bexp(b, env, if_else) ++

     instrs1 ++

     i"goto $if_end" ++

     l"$if_else" ++

     instrs2 ++

     l"$if_end", env2)

  }

  case While(b, bl) => {

    val loop_begin = Fresh("Loop_begin")

    val loop_end = Fresh("Loop_end")

    val (instrs1, env1) = compile_block(bl, env)

    (l"$loop_begin" ++

     compile_bexp(b, env, loop_end) ++

     instrs1 ++

     i"goto $loop_begin" ++

     l"$loop_end", env1)

  }

  case Write(x) => 

    (i"iload ${env(x)} \t\t; $x" ++ 

     i"invokestatic XXX/XXX/write(I)V", env)

  case Read(x) => {

    val index = env.getOrElse(x, env.keys.size) 

    (i"invokestatic XXX/XXX/read()I" ++ 

     i"istore $index \t\t; $x", env + (x -> index))

  }

  case ArrayDef(s: String, n: Int) => {

    val index = if (env.isDefinedAt(s)) throw new Exception("array def error") else 

                    env.keys.size

    (i"ldc $n" ++

     i"newarray int" ++

     i"astore $index", env + (s -> index))

  }

  case AssignA(s, a1, a2) => {

    val index = if (env.isDefinedAt(s)) env(s) else 

                    throw new Exception("array not defined")

    (i"aload ${env(s)}" ++

     compile_aexp(a1, env) ++

     compile_aexp(a2, env) ++

     i"iastore", env)

  } 

}

// compilation of a block (i.e. list of statements)

def compile_block(bl: Block, env: Env) : (String, Env) = bl match {

  case Nil => ("", env)

  case s::bl => {

    val (instrs1, env1) = compile_stmt(s, env)

    val (instrs2, env2) = compile_block(bl, env1)

    (instrs1 ++ instrs2, env2)

  }

}

// main compilation function for blocks

def compile(bl: Block, class_name: String) : String = {

  val instructions = compile_block(bl, Map())._1

  (beginning ++ instructions ++ ending).replaceAllLiterally("XXX", class_name)

}

// Fibonacci numbers as a test-case

val fib_test = 

  List(Read("n"),                       //  read n;                     

       Assign("minus1",Num(0)),         //  minus1 := 0;

       Assign("minus2",Num(1)),         //  minus2 := 1;

       Assign("temp",Num(0)),           //  temp := 0;

       While(Bop("<",Num(0),Var("n")),  //  while n > 0 do  {

          List(Assign("temp",Var("minus2")),    //  temp := minus2;

               Assign("minus2",Aop("+",Var("minus1"),Var("minus2"))), 

                                        //  minus2 := minus1 + minus2;

               Assign("minus1",Var("temp")), //  minus1 := temp;

               Assign("n",Aop("-",Var("n"),Num(1))))), //  n := n - 1 };

       Write("minus1"))                 //  write minus1

// prints out the JVM-assembly instructions for fib

//

// println(compile(fib_test, "fib"))

//

// can be assembled by hand with 

//

//    java -jar jvm/jasmin-2.4/jasmin.jar fib.j

//

// and started with

//

//    java fib/fib

import scala.util._

import scala.sys.process._

def time_needed[T](i: Int, code: => T) = {

  val start = System.nanoTime()

  for (j <- 2 to i) code

  val result = code

  val end = System.nanoTime()

  ((end - start) / (i * 1.0e9), result)

}

def compile_to_file(bl: Block, class_name: String) : Unit = 

  Using(new java.io.FileWriter(class_name + ".j")) {

    _.write(compile(bl, class_name))

  }

def compile_and_run(bl: Block, class_name: String) : Unit = {

  println(s"Start compilation of $class_name")

  compile_to_file(bl, class_name)

  println("generated .j file")

  (s"java -jar jvm/jasmin-2.4/jasmin.jar ${class_name}.j").!!

  println("generated .class file ")

  println("Time: " + time_needed(3, (s"java ${class_name}/${class_name}").!)._1)

}

val arr_test = 

  List(ArrayDef("a", 10),

       ArrayDef("b", 2),

       AssignA("a", Num(0), Num(10)),

       Assign("x", Ref("a", Num(0))),

       Write("x"),

       AssignA("b", Num(1), Num(5)),

       Assign("x", Ref("b", Num(1))),

       Write("x"))       

//compile_and_run(arr_test, "a")

//====================

// Parser Combinators

//====================

import scala.language.implicitConversions

import scala.language.reflectiveCalls

// more convenience for the semantic actions later on

case class ~[+A, +B](_1: A, _2: B)

type IsSeq[A] = A => Seq[_]

abstract class Parser[I : IsSeq, T] {

  def parse(ts: I): Set[(T, I)]

  def parse_all(ts: I) : Set[T] =

    for ((head, tail) <- parse(ts); if (tail.isEmpty)) yield head

}

class SeqParser[I : IsSeq, T, S](p: => Parser[I, T], q: => Parser[I, S]) extends Parser[I, ~[T, S]] {

  def parse(sb: I) = 

    for ((head1, tail1) <- p.parse(sb); 

         (head2, tail2) <- q.parse(tail1)) yield (new ~(head1, head2), tail2)

}

class AltParser[I : IsSeq, T](p: => Parser[I, T], q: => Parser[I, T]) extends Parser[I, T] {

  def parse(sb: I) = p.parse(sb) ++ q.parse(sb)   

}

class FunParser[I : IsSeq, T, S](p: => Parser[I, T], f: T => S) extends Parser[I, S] {

  def parse(sb: I) = 

    for ((head, tail) <- p.parse(sb)) yield (f(head), tail)

}

import scala.util.matching.Regex

case class RegexParser(reg: Regex) extends Parser[String, String] {

  def parse(sb: String) = reg.findPrefixMatchOf(sb) match {

    case None => Set()

    case Some(m) => Set((m.matched, m.after.toString))  

  }

}

def StringParser(s: String) = RegexParser(Regex.quote(s).r)

implicit def string2parser(s : String) = StringParser(s)

implicit def ParserOps[I : IsSeq, T](p: Parser[I, T]) = new {

  def | (q : => Parser[I, T]) = new AltParser[I, T](p, q)

  def ==>[S] (f: => T => S) = new FunParser[I, T, S](p, f)

  def ~[S] (q : => Parser[I, S]) = new SeqParser[I, T, S](p, q)

}

implicit def StringOps(s: String) = new {

  def | (q : => Parser[String, String]) = new AltParser[String, String](s, q)

  def | (r: String) = new AltParser[String, String](s, r)

  def ==>[S] (f: => String => S) = new FunParser[String, String, S](s, f)

  def ~[S](q : => Parser[String, S]) = 

    new SeqParser[String, String, S](s, q)

  def ~ (r: String) = 

    new SeqParser[String, String, String](s, r)

}

val NumParser = RegexParser("[0-9]+".r) ==> (s => s.toInt : Int)

val IdParser = RegexParser("[a-z][a-z,0-9]*".r)

// Grammar Rules for WHILE with arrays

lazy val AExp: Parser[String, AExp] = 

  (Te ~ "+" ~ AExp) ==> { case x ~ _ ~ z => Aop("+", x, z):AExp } |

  (Te ~ "-" ~ AExp) ==> { case x ~ _ ~ z => Aop("-", x, z):AExp } | Te

lazy val Te: Parser[String, AExp] = 

  (Fa ~ "*" ~ Te) ==> { case x ~ _ ~ z => Aop("*", x, z):AExp } | Fa

lazy val Fa: Parser[String, AExp] = 

   ("(" ~ AExp ~ ")") ==> { case _ ~ y ~ _ => y } | 

   (IdParser ~ "[" ~ AExp ~ "]") ==> { case x ~ _ ~ y ~ _ => Ref(x, y) } |

   IdParser ==> Var | 

   NumParser ==> Num

// boolean expressions

lazy val BExp: Parser[String, BExp] = 

   (AExp ~ "=" ~ AExp) ==> { case x ~ y ~ z => Bop("=", x, z):BExp } | 

   (AExp ~ "!=" ~ AExp) ==> { case x ~ y ~ z => Bop("!=", x, z):BExp } | 

   (AExp ~ "<" ~ AExp) ==> { case x ~ y ~ z => Bop("<", x, z):BExp } | 

   (AExp ~ ">" ~ AExp) ==> { case x ~ y ~ z => Bop("<", z, x):BExp } | 

   ("true" ==> ((_) => True:BExp )) | 

   ("false" ==> ((_) => False:BExp )) |

   ("(" ~ BExp ~ ")") ==> { case x ~ y ~ z => y}

lazy val Stmt: Parser[String, Stmt] =

   ("skip" ==> (_ => Skip: Stmt)) |

   (IdParser ~ ":=" ~ AExp) ==> { case x ~ y ~z => Assign(x, z): Stmt } |

   (IdParser ~ "[" ~ AExp ~ "]" ~ ":=" ~ AExp) ==> { 

     case x ~ _ ~ z ~ _ ~ _ ~ u => AssignA(x, z, u): Stmt } |

   ("if" ~ BExp ~ "then" ~ Block ~ "else" ~ Block) ==>

    { case _ ~ y ~ _ ~ u ~ _ ~w => If(y, u, w): Stmt } |

   ("while" ~ BExp ~ "do" ~ Block) ==> { case _ ~ y ~ _ ~ w => While(y, w) } |

   ("new(" ~ IdParser ~ "[" ~ NumParser ~ "])") ==> { 

    case _ ~ y ~ _ ~ u ~ _ => ArrayDef(y, u) } |

   ("write(" ~ IdParser ~ ")") ==> { case _ ~ y ~ _ => Write(y) } 

lazy val Stmts: Parser[String, Block] =

  (Stmt ~ ";" ~ Stmts) ==> { case x ~ _ ~ z => x :: z : Block } |

  (Stmt ==> ((s) => List(s) : Block)) 

lazy val Block: Parser[String, Block] =

   ("{" ~ Stmts ~ "}") ==> { case _ ~ y ~ _ => y} | 

   (Stmt ==> (s => List(s)))

//Stmts.parse_all("x2:=5+a")

//Stmts.parse_all("x2:=5+a[3+a]")

//Stmts.parse_all("a[x2+3]:=5+a[3+a]")

//Block.parse_all("{x:=5;y:=8}")

//Block.parse_all("if(false)then{x:=5}else{x:=10}")

val fib = """

   n := 10;

   minus1 := 0;

   minus2 := 1;

   temp:=0;

   while (n > 0) do {

     temp := minus2; 

     minus2 := minus1 + minus2;

     minus1 := temp;

     n := n - 1};

   result := minus2;

   write(result)

   """.replaceAll("\\s+", "")

val fib_prog = Stmts.parse_all(fib).toList

//compile_and_run(fib_prog.head, "fib")

//======================================

// BF transpiler into WHILE with arrays

//======================================

// simple BF instructions translation

def instr(c: Char) : String = c match {

  case '>' => "ptr := ptr + 1;"

  case '<' => "ptr := ptr - 1;"

  case '+' => "mem[ptr] := mem[ptr] + 1;"

  case '-' => "mem[ptr] := mem[ptr] - 1;"

  case '.' => "x := mem[ptr]; write x;"

  //case ',' => "XXX" // "ptr = getchar();\n"

  case '['  => "while (mem[ptr] != 0) do {"

  case ']'  => "skip};"

  case _ => ""

}

def instrs(prog: String) : String =

  prog.toList.map(instr).mkString

// Note: the mandelbrot.bf program is so large that

// it does not fit inside the limitations of what the

// JVM imposes on methods (only 64K of instructions).

// Therefore some optimisations are first applied to 

// BF programs before WHILE programs are created. The

// optimisations are 

//  

//  - replacing BF-loops of the form [-] 

//  - combining single increment/decrement instructions

//

// The size of the resulting .j-file is 270K.

def splice(cs: List[Char], acc: List[(Char, Int)]) : List[(Char, Int)] = (cs, acc) match {

  case (Nil, acc) => acc

  case (c :: cs, Nil) => splice(cs, List((c, 1)))

  case (c :: cs, (d, n) :: acc) => 

    if (c == d) splice(cs, (c, n + 1) :: acc)

    else splice(cs, (c, 1) :: (d, n) :: acc)

}

def spl(s: String) = splice(s.toList, Nil).reverse

def instr2(c: Char, n: Int) : String = c match {

  case '>' => s"ptr := ptr + $n;"

  case '<' => s"ptr := ptr - $n;"

  case '0' => s"mem[ptr] := 0;"

  case '+' => s"mem[ptr] := mem[ptr] + $n;"

  case '-' => s"mem[ptr] := mem[ptr] - $n;"

  case '.' => s"x := mem[ptr]; write(x);" 

  case '['  => "while (mem[ptr] != 0) do {" * n 

  case ']'  => "skip};" * n

  case _ => ""

}

def instrs2(prog: String) : String =

  spl(prog.replaceAll("""\[-\]""", "0")).map{ case (c, n) => instr2(c, n) }.mkString

def bf_str(prog: String) : String = {

  "new(mem[30000]);" ++

  "ptr := 15000;" ++

  instrs2(prog) ++

  "skip"

}

def bf_run(prog: String, name: String) = {

  println(s"BF pre-processing of $name")

  val bf_string = bf_str(prog).replaceAll("\\s", "")

  println(s"BF parsing (program length ${bf_string.length} characters)")

  val (time, bf_prog) = time_needed(1, Stmts.parse_all(bf_string).toList.head)

  println(s"BF generated WHILE program (needed $time for parsing)")

  compile_and_run(bf_prog, name)

}

// a benchmark program (counts down from 'Z' to 'A')

val bf0 = """>++[<+++++++++++++>-]<[[>+>+<<-]>[<+>-]++++++++

            [>++++++++<-]>.[-]<<>++++++++++[>++++++++++[>++

            ++++++++[>++++++++++[>++++++++++[>++++++++++[>+

            +++++++++[-]<-]<-]<-]<-]<-]<-]<-]++++++++++."""

bf_run(bf0, "bench")

// Sierpinski triangle

val bf1 = """++++++++[>+>++++<<-]>++>>+<[-[>>+<<-]+>>]>+[-<<<[

      ->[+[-]+>++>>>-<<]<[<]>>++++++[<<+++++>>-]+<<++.[-]<<

      ]>.>+[>>]>+]"""

bf_run(bf1, "sier")