src/lexer.ml @@ -11,6 +11,8 @@ modern programming languages (think Python, Ruby, or Haskell). It's not intended to appeal to anyone other than myself; it's more like a sandbox for my experiments. +The current implementation is written in OCaml. + --- Somersault is syntactically similar to the ML family of languages. README @@ -0,0 +1,40 @@ +{ + open Parser + open Printf +} + + +let digit = ['0'-'9'] +let identifier = ['a'-'z'] ['a'-'z' 'A'-'Z' '_']* + +rule tokenize = parse + [' ' '\t' '\r' '\n'] { tokenize lexbuf } + | digit+ as number + { printf "(NUMBER, %i)\n" (int_of_string number); + tokenize lexbuf + } + | identifier as id + { printf "(IDENTIFIER, %s)\n" id; + tokenize lexbuf + } + | _ as c + { printf "Unrecognized token: %c\n" c; + tokenize lexbuf + } + + | eof { } + + +{ + let main () = + let cin = + if Array.length Sys.argv > 1 then + open_in Sys.argv.(1) + else + stdin + in + let lexbuf = Lexing.from_channel cin in + tokenize lexbuf + +let _ = Printexc.print main () +} src/lexer.mll @@ -0,0 +1,40 @@ +%token EOF + +%token T_BOOL +%token T_DUMMY +%token T_INT +%token T_NIL +%token T_STRING + +%token IDENTIFIER + +%token LPAREN +%token RPAREN + +/* reserved */ +%token ARROW +%token DUMMY +%token FALSE +%token FUN +%token LET +%token NIL +%token REC +%token TRUE +%token WHERE +%token WITHIN + + +/* operators */ +%token NEG +%token PLUS MINUS TIMES DIV +%token EQUAL LESS GREAT LESS_E GREAT_E +%token AT + +%start expr +%type unit +%% + +expr: + EOF { () } + +%% src/parser.mly ba99f5d1554264776763a740cf443c559bd613fb Kenny Sanders sandersk@gmail.com http://github.com/kobs/somersault/commit/837c63e11ba9d6739cee505ffd2f68bdb0c7b957 837c63e11ba9d6739cee505ffd2f68bdb0c7b957 2009-05-25T21:25:40-07:00 2009-05-25T21:25:40-07:00 basic lexer 4e04d9b2e45b6d8f6121987eedc51f73023e60ec Kenny Sanders sandersk@gmail.com