afl-material: comparison progs/lexer.scala

equal deleted inserted replaced

-:89a37a73734a
+:2eecbcbc7c51
-// A Simple Lexer according to Sulzmann & Lu
+// A simple lexer inspired by work of Sulzmann & Lu
+//==================================================
 import scala.language.implicitConversions
 import scala.language.reflectiveCalls
+// regular expressions including records
 abstract class Rexp
 case object ZERO extends Rexp
 case object ONE extends Rexp
 case class CHAR(c: Char) extends Rexp
 case class ALT(r1: Rexp, r2: Rexp) extends Rexp
 case class SEQ(r1: Rexp, r2: Rexp) extends Rexp
 case class STAR(r: Rexp) extends Rexp
 case class RECD(x: String, r: Rexp) extends Rexp
+// values
 abstract class Val
 case object Empty extends Val
 case class Chr(c: Char) extends Val
 case class Sequ(v1: Val, v2: Val) extends Val
 case class Left(v: Val) extends Val
 def ~ (r: Rexp) = SEQ(s, r)
 def ~ (r: String) = SEQ(s, r)
 def $ (r: Rexp) = RECD(s, r)
 }
-// A test for more conveninet syntax
+def nullable(r: Rexp) : Boolean = r match {
-val re : Rexp = ("ab" | "a") ~ ("b" | ONE)
-// the nullable function: tests whether the regular
-// expression can recognise the empty string
-def nullable (r: Rexp) : Boolean = r match {
 case ZERO => false
 case ONE => true
 case CHAR(_) => false
 case ALT(r1, r2) => nullable(r1) || nullable(r2)
 case SEQ(r1, r2) => nullable(r1) && nullable(r2)
 case STAR(_) => true
 case RECD(_, r1) => nullable(r1)
 }
-// the derivative of a regular expression w.r.t. a character
+def der(c: Char, r: Rexp) : Rexp = r match {
-def der (c: Char, r: Rexp) : Rexp = r match {
 case ZERO => ZERO
 case ONE => ZERO
 case CHAR(d) => if (c == d) ONE else ZERO
 case ALT(r1, r2) => ALT(der(c, r1), der(c, r2))
 case SEQ(r1, r2) =>
 else SEQ(der(c, r1), r2)
 case STAR(r) => SEQ(der(c, r), STAR(r))
 case RECD(_, r1) => der(c, r1)
 }
-// the derivative w.r.t. a string (iterates der)
-def ders (s: List[Char], r: Rexp) : Rexp = s match {
-case Nil => r
-case c::s => ders(s, der(c, r))
-}
 // extracts a string from value
 def flatten(v: Val) : String = v match {
 case Empty => ""
 case Chr(c) => c.toString
 case Sequ(v1, v2) => flatten(v1) + flatten(v2)
 case Stars(vs) => vs.map(flatten).mkString
 case Rec(_, v) => flatten(v)
 }
 // extracts an environment from a value;
-// used for lexing a string
+// used for tokenise a string
 def env(v: Val) : List[(String, String)] = v match {
 case Empty => Nil
 case Chr(c) => Nil
 case Left(v) => env(v)
 case Right(v) => env(v)
 case Sequ(v1, v2) => env(v1) ::: env(v2)
 case Stars(vs) => vs.flatMap(env)
 case Rec(x, v) => (x, flatten(v))::env(v)
 }
-// The Injection Part of the Lexer
+// The Injection Part of the lexer
-// calculates a value for how a nullable regex
-// matches the empty string
 def mkeps(r: Rexp) : Val = r match {
 case ONE => Empty
 case ALT(r1, r2) =>
 if (nullable(r1)) Left(mkeps(r1)) else Right(mkeps(r2))
 case SEQ(r1, r2) => Sequ(mkeps(r1), mkeps(r2))
 case STAR(r) => Stars(Nil)
 case RECD(x, r) => Rec(x, mkeps(r))
 }
-// injects back a character into a value
 def inj(r: Rexp, c: Char, v: Val) : Val = (r, v) match {
 case (STAR(r), Sequ(v1, Stars(vs))) => Stars(inj(r, c, v1)::vs)
 case (SEQ(r1, r2), Sequ(v1, v2)) => Sequ(inj(r1, c, v1), v2)
 case (SEQ(r1, r2), Left(Sequ(v1, v2))) => Sequ(inj(r1, c, v1), v2)
 case (SEQ(r1, r2), Right(v2)) => Sequ(mkeps(r1), inj(r2, c, v2))
 case (ALT(r1, r2), Right(v2)) => Right(inj(r2, c, v2))
 case (CHAR(d), Empty) => Chr(c)
 case (RECD(x, r1), _) => Rec(x, inj(r1, c, v))
 }
-// the main lexing function (produces a value)
-def lex(r: Rexp, s: List[Char]) : Val = s match {
-case Nil => if (nullable(r)) mkeps(r)
-else throw new Exception("Not matched")
-case c::cs => inj(r, c, lex(der(c, r), cs))
-}
-def lexing(r: Rexp, s: String) : Val = lex(r, s.toList)
-// a simple test for extracting an environment
-val re1 : Rexp = ("first" $ ("a" | "ab")) ~ ("second" $ ("b" | ONE))
-env(lexing(re1, "ab"))
 // some "rectification" functions for simplification
 def F_ID(v: Val): Val = v
 def F_RIGHT(f: Val => Val) = (v:Val) => Right(f(v))
 def F_LEFT(f: Val => Val) = (v:Val) => Left(f(v))
 def F_ALT(f1: Val => Val, f2: Val => Val) = (v:Val) => v match {
 def F_RECD(f: Val => Val) = (v:Val) => v match {
 case Rec(x, v) => Rec(x, f(v))
 }
 def F_ERROR(v: Val): Val = throw new Exception("error")
-// simplification of regular expressions returns now also
-// an rectification function; no simplification under STAR
 def simp(r: Rexp): (Rexp, Val => Val) = r match {
 case ALT(r1, r2) => {
 val (r1s, f1s) = simp(r1)
 val (r2s, f2s) = simp(r2)
 (r1s, r2s) match {
 case r => (r, F_ID)
 }
 // lexing functions including simplification
 def lex_simp(r: Rexp, s: List[Char]) : Val = s match {
-case Nil => if (nullable(r)) mkeps(r) else throw new Exception("Not matched")
+case Nil => if (nullable(r)) mkeps(r) else
+{ throw new Exception("lexing error") }
 case c::cs => {
 val (r_simp, f_simp) = simp(der(c, r))
 inj(r, c, f_simp(lex_simp(r_simp, cs)))
 }
 }
-def lexing_simp(r: Rexp, s: String) : Val = lex_simp(r, s.toList)
+def lexing_simp(r: Rexp, s: String) =
+env(lex_simp(r, s.toList))
-lexing_simp(("a" | "ab") ~ ("b" | ""), "ab")
-// The Lexing Rules for a Small While Language
+// The Lexing Rules for the Fun Language
 def PLUS(r: Rexp) = r ~ r.%
-val SYM = "a" | "b" | "c" | "d" | "e" | "f" | "g" | "h" | "i" | "j" | "k" | "l" | "m" | "n" | "o" | "p" | "q" | "r" | "s" | "t" | "u" | "v" | "w" | "x" | "y" | "z"
+def Range(s : List[Char]) : Rexp = s match {
-val DIGIT = "0" | "1" | "2" | "3" | "4" | "5" | "6" | "7" | "8" | "9"
+case Nil => ZERO
+case c::Nil => CHAR(c)
+case c::s => ALT(CHAR(c), Range(s))
+}
+def RANGE(s: String) = Range(s.toList)
+val SYM = RANGE("ABCDEFGHIJKLMNOPQRSTUVXYZabcdefghijklmnopqrstuvwxyz_")
+val DIGIT = RANGE("0123456789")
 val ID = SYM ~ (SYM | DIGIT).%
 val NUM = PLUS(DIGIT)
-val KEYWORD : Rexp = "skip" | "while" | "do" | "if" | "then" | "else" | "read" | "write" | "true" | "false"
+val KEYWORD : Rexp = "skip" | "while" | "do" | "if" | "then" | "else" | "read" | "write"
 val SEMI: Rexp = ";"
-val OP: Rexp = ":=" | "==" | "-" | "+" | "*" | "!=" | "<" | ">" | "<=" | ">=" | "%" | "/"
+val OP: Rexp = ":=" | "=" | "-" | "+" | "*" | "!=" | "<" | ">"
 val WHITESPACE = PLUS(" " | "\n" | "\t")
-val RPAREN: Rexp = ")"
+val RPAREN: Rexp = "{"
-val LPAREN: Rexp = "("
+val LPAREN: Rexp = "}"
-val BEGIN: Rexp = "{"
-val END: Rexp = "}"
 val STRING: Rexp = "\"" ~ SYM.% ~ "\""
 val WHILE_REGS = (("k" $ KEYWORD) |
 ("i" $ ID) |
 ("o" $ OP) |
 ("n" $ NUM) |
 ("s" $ SEMI) |
 ("str" $ STRING) |
 ("p" $ (LPAREN | RPAREN)) |
-("b" $ (BEGIN | END)) |
 ("w" $ WHITESPACE)).%
-//   Testing
-//============
-def time[T](code: => T) = {
-val start = System.nanoTime()
-val result = code
-val end = System.nanoTime()
-println((end - start)/1.0e9)
-result
-}
-val r1 = ("a" | "ab") ~ ("bcd" | "c")
-println(lexing(r1, "abcd"))
-val r2 = ("" | "a") ~ ("ab" | "b")
-println(lexing(r2, "ab"))
 // Two Simple While Tests
 //========================
-println("prog0 test")
+println("test: read n")
-val prog0 = """read if"""
-println(env(lexing_simp(WHILE_REGS, prog0)))
+val prog0 = """read n"""
+println(lexing_simp(WHILE_REGS, prog0))
-println("prog1 test")
+println("test: read  n; write n ")
-val prog1 = """read  n; write (n)"""
-println(env(lexing_simp(WHILE_REGS, prog1)))
+val prog1 = """read  n; write n"""
+println(lexing_simp(WHILE_REGS, prog1))
-// Bigger Test
-//=============
+// Bigger Tests
+//==============
+// escapes strings and prints them out as "", "\n" and so on
+def esc(raw: String): String = {
+import scala.reflect.runtime.universe._
+Literal(Constant(raw)).toString
+}
+def escape(tks: List[(String, String)]) =
+tks.map{ case (s1, s2) => (s1, esc(s2))}
 val prog2 = """
-write "fib";
+write "Fib";
 read n;
 minus1 := 0;
 minus2 := 1;
 while n > 0 do {
 temp := minus2;
 minus2 := minus1 + minus2;
 minus1 := temp;
 n := n - 1
 };
-write "result";
-write minus2
-"""
-println("Tokens")
-println(env(lexing_simp(WHILE_REGS, prog2)))
-println(env(lexing_simp(WHILE_REGS, prog2)).filterNot{_._1 == "w"}.mkString("\n"))
-// some more timing tests with
-// i copies of the program
-for (i <- 0 to 20 by 10) {
-print(i.toString + ":  ")
-time(lexing_simp(WHILE_REGS, prog2 * i))
-}
-val fib = """
-write "Fib";
-read n;
-minus1 := 0;
-minus2 := 1;
-while n > 0 do {
-temp := minus2;
-minus2 := minus1 + minus2;
-minus1 := temp;
-n := n - 1
-};
 write "Result";
 write minus2
 """
-println(env(lexing_simp(WHILE_REGS, prog2)).filterNot{_._1 == "w"})
+println("lexing Fib")
+println(escape(lexing_simp(WHILE_REGS, prog2)).mkString("\n"))
+val prog3 = """
+start := 1000;
+x := start;
+y := start;
+z := start;
+while 0 < x do {
+while 0 < y do {
+while 0 < z do {
+z := z - 1
+};
+z := start;
+y := y - 1
+};
+y := start;
+x := x - 1
+}
+"""
+println("lexing Loops")
+println(escape(lexing_simp(WHILE_REGS, prog3)).mkString("\n"))

changeset 642	2eecbcbc7c51
parent 625	f8d3e2c7d8b2
child 670	67c60bf4f4f5