// A Small Compiler for the WHILE Language
// (stub for including arrays)
// 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 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
case class AssignA(s: String, a1: AExp, a2: AExp) extends Stmt
case class Write(s: String) extends Stmt // writes out a variable
case class Array(s: String, n: Int) 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, a1: 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 <init>()V
aload_0
invokenonvirtual java/lang/Object/<init>()V
return
.end method
.method public static write(I)V
.limit locals 1
.limit stack 2
getstatic java/lang/System/out Ljava/io/PrintStream;
iload 0
i2c
invokevirtual java/io/PrintStream/print(C)V
return
.end method
.method public static main([Ljava/lang/String;)V
.limit locals 200
.limit stack 200
"""
val ending = """
return
.end method
"""
println("Start compilation")
// for generating new labels
var counter = -1
def Fresh(x: String) = {
counter += 1
x ++ "_" ++ counter.toString()
}
// environments and instructions
type Env = Map[String, String]
type Instrs = List[String]
// arithmetic expression compilation
def compile_aexp(a: AExp, env : Env) : Instrs = a match {
case Num(i) => List("ldc " + i.toString + "\n")
case Var(s) => List("iload " + env(s) + "\n")
case Aop("+", a1, a2) =>
compile_aexp(a1, env) ++
compile_aexp(a2, env) ++ List("iadd\n")
case Aop("-", a1, a2) =>
compile_aexp(a1, env) ++ compile_aexp(a2, env) ++ List("isub\n")
case Aop("*", a1, a2) =>
compile_aexp(a1, env) ++ compile_aexp(a2, env) ++ List("imul\n")
case Ref(s, a1) =>
List("aload " + env(s) + "\n") ++ compile_aexp(a1, env) ++ List("iaload \n")
}
// boolean expression compilation
def compile_bexp(b: BExp, env : Env, jmp: String) : Instrs = b match {
case True => Nil
case False => List("goto " + jmp + "\n")
case Bop("=", a1, a2) =>
compile_aexp(a1, env) ++ compile_aexp(a2, env) ++
List("if_icmpne " + jmp + "\n")
case Bop("!=", a1, a2) =>
compile_aexp(a1, env) ++ compile_aexp(a2, env) ++
List("if_icmpeq " + jmp + "\n")
case Bop("<", a1, a2) =>
compile_aexp(a1, env) ++ compile_aexp(a2, env) ++
List("if_icmpge " + jmp + "\n")
}
// statement compilation
def compile_stmt(s: Stmt, env: Env) : (Instrs, Env) = s match {
case Skip => (Nil, env)
case Assign(x, a) => {
val index = if (env.isDefinedAt(x)) env(x) else
env.keys.size.toString
(compile_aexp(a, env) ++
List("istore " + index + "\n"), 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 ++
List("goto " + if_end + "\n") ++
List("\n" + if_else + ":\n\n") ++
instrs2 ++
List("\n" + if_end + ":\n\n"), env2)
}
case While(b, bl) => {
val loop_begin = Fresh("Loop_begin")
val loop_end = Fresh("Loop_end")
val (instrs1, env1) = compile_block(bl, env)
(List("\n" + loop_begin + ":\n\n") ++
compile_bexp(b, env, loop_end) ++
instrs1 ++
List("goto " + loop_begin + "\n") ++
List("\n" + loop_end + ":\n\n"), env1)
}
case Write(x) =>
(List("iload " + env(x) + "\n" +
"invokestatic XXX/XXX/write(I)V\n"), env)
case Array(s, n) => {
val index = if (env.isDefinedAt(s)) throw new Exception("Array already defined") else
env.keys.size.toString
(List("ldc " ++ n.toString ++ "\n",
"newarray int \n",
"astore " ++ index ++ "\n"), env + (s -> index))
}
case AssignA(s, a1, a2) => {
val index = if (env.isDefinedAt(s)) env(s) else
throw new Exception("Array not yet defined")
(List("aload " + index + "\n") ++
compile_aexp(a1, env) ++
compile_aexp(a2, env) ++
List("iastore \n"), env)
}
}
// compilation of a block (i.e. list of instructions)
def compile_block(bl: Block, env: Env) : (Instrs, Env) = bl match {
case Nil => (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.empty)._1
(beginning ++ instructions.mkString ++ ending).replaceAllLiterally("XXX", class_name)
}
// compiling and running files
//
// JVM files can be assembled with
//
// java -jar jvm/jasmin-2.4/jasmin.jar fib.j
//
// and started with
//
// java fib/fib
import scala.util._
import scala.sys.process._
import scala.io
def compile_tofile(bl: Block, class_name: String) = {
val output = compile(bl, class_name)
val fw = new java.io.FileWriter(class_name + ".j")
fw.write(output)
fw.close()
}
def compile_all(bl: Block, class_name: String) : Unit = {
compile_tofile(bl, class_name)
println("compiled ")
val test = ("java -jar jvm/jasmin-2.4/jasmin.jar " + class_name + ".j").!!
println("assembled ")
}
def time_needed[T](i: Int, code: => T) = {
val start = System.nanoTime()
for (j <- 1 to i) code
val end = System.nanoTime()
(end - start)/(i * 1.0e9)
}
def compile_run(bl: Block, class_name: String) : Unit = {
println("Start compilation")
compile_all(bl, class_name)
println("Time: " + time_needed(1, ("java " + class_name + "/" + class_name).!))
}
// Fibonacci numbers as a test-case
val fib_test =
List(Assign("n", Num(10)), // n := 10;
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
compile_run(fib_test, "fib")
val arr_test =
List(Array("a", 10), // a[10]
Array("b", 2), // b[2]
AssignA("a", Num(0), Num(10)), // a[0] := 10
Assign("x", Ref("a", Num(0))), // x := a[0]
Write("x"), // write x
AssignA("b", Num(1), Num(5)), // b[1] := 5
Assign("x", Ref("b", Num(1))), // x := b[1]
Write("x")) // write x
compile_run(arr_test, "a")
//====================
// Parser Combinators
//====================
import scala.language.implicitConversions
import scala.language.reflectiveCalls
abstract class Parser[I <% Seq[_], 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 <% Seq[_], 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 ((head1, head2), tail2)
}
class AltParser[I <% Seq[_], 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 <% Seq[_], 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<% Seq[_], 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
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, y), z), v), w), u) => AssignA(x, z, u): Stmt } |
("if" ~ BExp ~ "then" ~ Block ~ "else" ~ Block) ==>
{ case (((((x,y),z),u),v),w) => If(y, u, w): Stmt } |
("while" ~ BExp ~ "do" ~ Block) ==> { case (((x, y), z), w) => While(y, w) } |
("new" ~ IdParser ~ "[" ~ NumParser ~ "]") ==> { case ((((x, y), z), u), v) => Array(y, u) } |
("write" ~ IdParser) ==> { case (_, y) => Write(y) }
lazy val Stmts: Parser[String, Block] =
(Stmt ~ ";" ~ Stmts) ==> { case ((x, y), z) => x :: z : Block } |
(Stmt ==> ((s) => List(s) : Block))
lazy val Block: Parser[String, Block] =
("{" ~ Stmts ~ "}") ==> { case ((x, y), z) => 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_run(fib_prog.head, "fib")
// BF
// simple instructions
def instr(c: Char) : String = c match {
case '>' => "ptr := ptr + 1;"
case '<' => "ptr := ptr - 1;"
case '+' => "field[ptr] := field[ptr] + 1;"
case '-' => "field[ptr] := field[ptr] - 1;"
case '.' => "x := field[ptr]; write x;"
//case ',' => "XXX" // "ptr = getchar();\n"
case '[' => "while (field[ptr] != 0) do {"
case ']' => "skip};"
case _ => ""
}
def instrs(prog: String) : String =
prog.toList.map(instr).mkString
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 '>' => "ptr := ptr + " + n.toString + ";"
case '<' => "ptr := ptr - " + n.toString + ";"
case '+' => "field[ptr] := field[ptr] + " + n.toString + ";"
case '-' => "field[ptr] := field[ptr] - " + n.toString + ";"
case '.' => "x := field[ptr]; write x;" //* n
//case ',' => "*ptr = getchar();\n" * n
case '[' => "while (field[ptr] != 0) do {" * n
case ']' => "skip};" * n
case _ => ""
}
def instrs2(prog: String) : String =
spl(prog).map{ case (c, n) => instr2(c, n) }.mkString
def bf_str(prog: String) : String = {
"\n" ++
//"new field[30000];\n" ++
"ptr := 15000;" ++
instrs2(prog) ++
"skip"
}
def bf_run(prog: String, name: String) = {
println("BF processing start")
val bf_string = bf_str(prog).replaceAll("\\s", "")
println(s"BF parsing start (string length ${bf_string.length})")
val bf_prog = Stmts.parse_all(bf_string).toList.head
println("BF Compile start")
compile_run(Array("field", 30000) :: bf_prog, name)
}
val bf1 = """++++++++[>+>++++<<-]>++>>+<[-[>>+<<-]+>>]>+[-<<<[
->[+[-]+>++>>>-<<]<[<]>>++++++[<<+++++>>-]+<<++.[-]<<
]>.>+[>>]>+]"""
bf_run(bf1, "sier")
bf_run("""++++++++[>++++[>++>+++>+++>+<<<<-]>+>+>->>+[<]<-]>>.>---.+++++++
..+++.>>.<-.<.+++.------.--------.>>+.>++.""", "hello")
bf_run("""+++++++++++
>+>>>>++++++++++++++++++++++++++++++++++++++++++++
>++++++++++++++++++++++++++++++++<<<<<<[>[>>>>>>+>
+<<<<<<<-]>>>>>>>[<<<<<<<+>>>>>>>-]<[>++++++++++[-
<-[>>+>+<<<-]>>>[<<<+>>>-]+<[>[-]<[-]]>[<<[>>>+<<<
-]>>[-]]<<]>>>[>>+>+<<<-]>>>[<<<+>>>-]+<[>[-]<[-]]
>[<<+>>[-]]<<<<<<<]>>>>>[+++++++++++++++++++++++++
+++++++++++++++++++++++.[-]]++++++++++<[->-<]>++++
++++++++++++++++++++++++++++++++++++++++++++.[-]<<
<<<<<<<<<<[>>>+>+<<<<-]>>>>[<<<<+>>>>-]<-[>>.>.<<<
[-]]<<[>>+>+<<<-]>>>[<<<+>>>-]<<[<+>-]>[<+>-]<<<-]""", "fibs")