// A Small Compiler for the WHILE Language with Arrays
// (it does not use a parser and lexer)
// the new parts are
// - declaring an array
// - references an array cell
// - assigning an array cell
// 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
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
invokevirtual java/io/PrintStream/println(I)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
"""
// Compiler functions
// for generating new labels
var counter = -1
def Fresh(x: String) = {
counter += 1
x ++ "_" ++ counter.toString()
}
// 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"
}
// environments
type Env = Map[String, String]
// 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("+", a1, a2) =>
compile_aexp(a1, env) ++ compile_aexp(a2, env) ++ i"iadd"
case Aop("-", a1, a2) =>
compile_aexp(a1, env) ++ compile_aexp(a2, env) ++ i"isub"
case Aop("*", a1, a2) =>
compile_aexp(a1, env) ++ compile_aexp(a2, env) ++ i"imul"
case Ref(s, a1) =>
i"aload ${env(s)}" ++ compile_aexp(a1, 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 = if (env.isDefinedAt(x)) env(x) else
env.keys.size.toString
(compile_aexp(a, env) ++ i"istore $index", 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)}" ++
i"invokestatic XXX/XXX/write(I)V", env)
case Array(s, n) => {
val index = if (env.isDefinedAt(s)) throw new Exception("Array already defined") else
env.keys.size.toString
(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 yet defined")
(i"aload $index" ++
compile_aexp(a1, env) ++
compile_aexp(a2, env) ++
i"iastore", env)
}
}
// compilation of a block (i.e. list of instructions)
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.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("running")
println("Time: " + time_needed(1, ("java " + class_name + "/" + class_name).!))
}
// BF Part
// 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 '[' => "while (field[ptr] != 0) do {"
case ']' => "skip};"
case _ => ""
}
def instrs(prog: String) : String =
prog.toList.map(instr).mkString
// compound instructions
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 '+' => s"field[ptr] := field[ptr] + $n;"
case '-' => s"field[ptr] := field[ptr] - $n;"
case '.' => s"x := field[ptr]; write x;"
case '[' => s"while (field[ptr] != 0) do {" * n
case ']' => s"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(bfprog: String, name: String) = {
println("BF processing start")
val bf_string = bf_str(bfprog).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")