1 // A tokeniser for the fun language

     2 

     3 

     4 object Fun_Tokens {

     5 

     6 import scala.language.implicitConversions    

     7 import scala.language.reflectiveCalls 

     8 

     9 abstract class Rexp 

    10 case object ZERO extends Rexp

    11 case object ONE extends Rexp

    12 case class CHAR(c: Char) extends Rexp

    13 case class ALT(r1: Rexp, r2: Rexp) extends Rexp 

    14 case class SEQ(r1: Rexp, r2: Rexp) extends Rexp 

    15 case class STAR(r: Rexp) extends Rexp 

    16 case class RECD(x: String, r: Rexp) extends Rexp

    17   

    18 abstract class Val

    19 case object Empty extends Val

    20 case class Chr(c: Char) extends Val

    21 case class Sequ(v1: Val, v2: Val) extends Val

    22 case class Left(v: Val) extends Val

    23 case class Right(v: Val) extends Val

    24 case class Stars(vs: List[Val]) extends Val

    25 case class Rec(x: String, v: Val) extends Val

    26    

    27 // some convenience for typing in regular expressions

    28 def charlist2rexp(s : List[Char]): Rexp = s match {

    29   case Nil => ONE

    30   case c::Nil => CHAR(c)

    31   case c::s => SEQ(CHAR(c), charlist2rexp(s))

    32 }

    33 implicit def string2rexp(s : String) : Rexp = 

    34   charlist2rexp(s.toList)

    35 

    36 implicit def RexpOps(r: Rexp) = new {

    37   def | (s: Rexp) = ALT(r, s)

    38   def % = STAR(r)

    39   def ~ (s: Rexp) = SEQ(r, s)

    40 }

    41 

    42 implicit def stringOps(s: String) = new {

    43   def | (r: Rexp) = ALT(s, r)

    44   def | (r: String) = ALT(s, r)

    45   def % = STAR(s)

    46   def ~ (r: Rexp) = SEQ(s, r)

    47   def ~ (r: String) = SEQ(s, r)

    48   def $ (r: Rexp) = RECD(s, r)

    49 }

    50 

    51 def nullable (r: Rexp) : Boolean = r match {

    52   case ZERO => false

    53   case ONE => true

    54   case CHAR(_) => false

    55   case ALT(r1, r2) => nullable(r1) || nullable(r2)

    56   case SEQ(r1, r2) => nullable(r1) && nullable(r2)

    57   case STAR(_) => true

    58   case RECD(_, r1) => nullable(r1)

    59 }

    60 

    61 def der (c: Char, r: Rexp) : Rexp = r match {

    62   case ZERO => ZERO

    63   case ONE => ZERO

    64   case CHAR(d) => if (c == d) ONE else ZERO

    65   case ALT(r1, r2) => ALT(der(c, r1), der(c, r2))

    66   case SEQ(r1, r2) => 

    67     if (nullable(r1)) ALT(SEQ(der(c, r1), r2), der(c, r2))

    68     else SEQ(der(c, r1), r2)

    69   case STAR(r) => SEQ(der(c, r), STAR(r))

    70   case RECD(_, r1) => der(c, r1)

    71 }

    72 

    73 

    74 // extracts a string from value

    75 def flatten(v: Val) : String = v match {

    76   case Empty => ""

    77   case Chr(c) => c.toString

    78   case Left(v) => flatten(v)

    79   case Right(v) => flatten(v)

    80   case Sequ(v1, v2) => flatten(v1) + flatten(v2)

    81   case Stars(vs) => vs.map(flatten).mkString

    82   case Rec(_, v) => flatten(v)

    83 }

    84 

    85 // extracts an environment from a value;

    86 // used for tokenise a string

    87 def env(v: Val) : List[(String, String)] = v match {

    88   case Empty => Nil

    89   case Chr(c) => Nil

    90   case Left(v) => env(v)

    91   case Right(v) => env(v)

    92   case Sequ(v1, v2) => env(v1) ::: env(v2)

    93   case Stars(vs) => vs.flatMap(env)

    94   case Rec(x, v) => (x, flatten(v))::env(v)

    95 }

    96 

    97 // The Injection Part of the lexer

    98 

    99 def mkeps(r: Rexp) : Val = r match {

   100   case ONE => Empty

   101   case ALT(r1, r2) => 

   102     if (nullable(r1)) Left(mkeps(r1)) else Right(mkeps(r2))

   103   case SEQ(r1, r2) => Sequ(mkeps(r1), mkeps(r2))

   104   case STAR(r) => Stars(Nil)

   105   case RECD(x, r) => Rec(x, mkeps(r))

   106 }

   107 

   108 def inj(r: Rexp, c: Char, v: Val) : Val = (r, v) match {

   109   case (STAR(r), Sequ(v1, Stars(vs))) => Stars(inj(r, c, v1)::vs)

   110   case (SEQ(r1, r2), Sequ(v1, v2)) => Sequ(inj(r1, c, v1), v2)

   111   case (SEQ(r1, r2), Left(Sequ(v1, v2))) => Sequ(inj(r1, c, v1), v2)

   112   case (SEQ(r1, r2), Right(v2)) => Sequ(mkeps(r1), inj(r2, c, v2))

   113   case (ALT(r1, r2), Left(v1)) => Left(inj(r1, c, v1))

   114   case (ALT(r1, r2), Right(v2)) => Right(inj(r2, c, v2))

   115   case (CHAR(d), Empty) => Chr(c) 

   116   case (RECD(x, r1), _) => Rec(x, inj(r1, c, v))

   117   case _ => { println ("Injection error") ; sys.exit(-1) } 

   118 }

   119 

   120 // some "rectification" functions for simplification

   121 def F_ID(v: Val): Val = v

   122 def F_RIGHT(f: Val => Val) = (v:Val) => Right(f(v))

   123 def F_LEFT(f: Val => Val) = (v:Val) => Left(f(v))

   124 def F_ALT(f1: Val => Val, f2: Val => Val) = (v:Val) => v match {

   125   case Right(v) => Right(f2(v))

   126   case Left(v) => Left(f1(v))

   127 }

   128 def F_SEQ(f1: Val => Val, f2: Val => Val) = (v:Val) => v match {

   129   case Sequ(v1, v2) => Sequ(f1(v1), f2(v2))

   130 }

   131 def F_SEQ_Empty1(f1: Val => Val, f2: Val => Val) = 

   132   (v:Val) => Sequ(f1(Empty), f2(v))

   133 def F_SEQ_Empty2(f1: Val => Val, f2: Val => Val) = 

   134   (v:Val) => Sequ(f1(v), f2(Empty))

   135 def F_RECD(f: Val => Val) = (v:Val) => v match {

   136   case Rec(x, v) => Rec(x, f(v))

   137 }

   138 def F_ERROR(v: Val): Val = throw new Exception("error")

   139 

   140 def simp(r: Rexp): (Rexp, Val => Val) = r match {

   141   case ALT(r1, r2) => {

   142     val (r1s, f1s) = simp(r1)

   143     val (r2s, f2s) = simp(r2)

   144     (r1s, r2s) match {

   145       case (ZERO, _) => (r2s, F_RIGHT(f2s))

   146       case (_, ZERO) => (r1s, F_LEFT(f1s))

   147       case _ => if (r1s == r2s) (r1s, F_LEFT(f1s))

   148                 else (ALT (r1s, r2s), F_ALT(f1s, f2s)) 

   149     }

   150   }

   151   case SEQ(r1, r2) => {

   152     val (r1s, f1s) = simp(r1)

   153     val (r2s, f2s) = simp(r2)

   154     (r1s, r2s) match {

   155       case (ZERO, _) => (ZERO, F_ERROR)

   156       case (_, ZERO) => (ZERO, F_ERROR)

   157       case (ONE, _) => (r2s, F_SEQ_Empty1(f1s, f2s))

   158       case (_, ONE) => (r1s, F_SEQ_Empty2(f1s, f2s))

   159       case _ => (SEQ(r1s,r2s), F_SEQ(f1s, f2s))

   160     }

   161   }

   162   case RECD(x, r1) => {

   163     val (r1s, f1s) = simp(r1)

   164     (RECD(x, r1s), F_RECD(f1s))

   165   }

   166   case r => (r, F_ID)

   167 }

   168 

   169 // lexing functions including simplification

   170 def lex_simp(r: Rexp, s: List[Char]) : Val = s match {

   171   case Nil => if (nullable(r)) mkeps(r) else { println ("Lexing Error") ; sys.exit(-1) } 

   172   case c::cs => {

   173     val (r_simp, f_simp) = simp(der(c, r))

   174     inj(r, c, f_simp(lex_simp(r_simp, cs)))

   175   }

   176 }

   177 

   178 def lexing_simp(r: Rexp, s: String) = env(lex_simp(r, s.toList))

   179 

   180 

   181 // The Lexing Rules for the Fun Language

   182 

   183 def PLUS(r: Rexp) = r ~ r.%

   184 

   185 val SYM = "a" | "b" | "c" | "d" | "e" | "f" | "g" | "h" | "i" | "j" | "k" | 

   186           "l" | "m" | "n" | "o" | "p" | "q" | "r" | "s" | "t" | "u" | "v" | 

   187           "w" | "x" | "y" | "z" | "T" | "_"

   188 val DIGIT = "0" | "1" | "2" | "3" | "4" | "5" | "6" | "7" | "8" | "9"

   189 val ID = SYM ~ (SYM | DIGIT).% 

   190 val NUM = PLUS(DIGIT)

   191 val KEYWORD : Rexp = "if" | "then" | "else" | "write" | "def"

   192 val SEMI: Rexp = ";"

   193 val OP: Rexp = "=" | "==" | "-" | "+" | "*" | "!=" | "<" | ">" | "<=" | ">=" | "%" | "/"

   194 val WHITESPACE = PLUS(" " | "\n" | "\t")

   195 val RPAREN: Rexp = ")"

   196 val LPAREN: Rexp = "("

   197 val COMMA: Rexp = ","

   198 val ALL = SYM | DIGIT | OP | " " | ":" | ";" | "\"" | "=" | "," | "(" | ")"

   199 val ALL2 = ALL | "\n"

   200 val COMMENT = ("/*" ~ ALL2.% ~ "*/") | ("//" ~ ALL.% ~ "\n")

   201 

   202 

   203 val WHILE_REGS = (("k" $ KEYWORD) | 

   204                   ("i" $ ID) | 

   205                   ("o" $ OP) | 

   206                   ("n" $ NUM) | 

   207                   ("s" $ SEMI) | 

   208                   ("c" $ COMMA) |

   209                   ("pl" $ LPAREN) |

   210                   ("pr" $ RPAREN) |

   211                   ("w" $ (WHITESPACE | COMMENT))).%

   212 

   213 

   214 

   215 // The tokens for the Fun language

   216 

   217 import java.io._

   218 

   219 abstract class Token extends Serializable 

   220 case object T_SEMI extends Token

   221 case object T_COMMA extends Token

   222 case object T_LPAREN extends Token

   223 case object T_RPAREN extends Token

   224 case class T_ID(s: String) extends Token

   225 case class T_OP(s: String) extends Token

   226 case class T_NUM(n: Int) extends Token

   227 case class T_KWD(s: String) extends Token

   228 

   229 val token : PartialFunction[(String, String), Token] = {

   230   case ("k", s) => T_KWD(s)

   231   case ("i", s) => T_ID(s)

   232   case ("o", s) => T_OP(s)

   233   case ("n", s) => T_NUM(s.toInt)

   234   case ("s", _) => T_SEMI

   235   case ("c", _) => T_COMMA

   236   case ("pl", _) => T_LPAREN

   237   case ("pr", _) => T_RPAREN

   238 }

   239 

   240 

   241 def tokenise(s: String) : List[Token] = 

   242   lexing_simp(WHILE_REGS, s).collect(token)

   243 

   244 def serialise[T](fname: String, data: T) = {

   245   val out = new ObjectOutputStream(new FileOutputStream(fname))

   246   out.writeObject(data)

   247   out.close

   248 }

   249 

   250 def main(args: Array[String]) = {

   251   val fname = args(0)

   252   val file = io.Source.fromFile(fname).mkString

   253   val tks = fname.stripSuffix(".fun") ++ ".tks"

   254   serialise(tks, tokenise(file))

   255 }

   256 

   257 

   258 }