Better approach: now we create a token SQUOTE "'" in the lexer, then we decide

if it is a string or a type variable in the parser. Still some errors in the
parser (see make test).

parser/ulexer.ml

@@ -46,11 +46,10 @@ end
 type token =
   | KEYWORD of string
   | IDENT of string
-  | PVAR of string
   | ANY_IN_NS of string
   | INT of string
-  | STRING1 of string
-  | STRING2 of string
+  | STRING of string
+  | SQUOTE
   | EOI
 
 module Token = struct
@@ -65,10 +64,9 @@ module Token = struct
     function
     | KEYWORD s -> sf "KEYWORD %S" s
     | IDENT s -> sf "IDENT %S" s
-    | PVAR s -> sf "PVAR \"%s\"" s
     | INT s -> sf "INT %s" s
-    | STRING1 s -> sf "STRING \"%s\"" s
-    | STRING2 s -> sf "STRING \"%s\"" s
+    | SQUOTE -> sf "'"
+    | STRING s -> sf "STRING \"%s\"" s
                       (* here it's not %S since the string is already escaped *)
     | ANY_IN_NS s -> sf "ANY_IN_NS %S" s
     | EOI -> sf "EOI"
@@ -82,7 +80,7 @@ module Token = struct
 
   let extract_string =
     function
-      | KEYWORD s | IDENT s | PVAR s | INT s | STRING1 s | STRING2 s |
+      | KEYWORD s | IDENT s | INT s | STRING s |
 	  ANY_IN_NS s -> s
       | tok ->
         invalid_arg ("Cannot extract a string from this token: "^
@@ -212,14 +210,10 @@ let rec token = lexer
      return lexbuf (KEYWORD (L.utf8_lexeme lexbuf))
  | '"' ->
      let start = L.lexeme_start lexbuf in
-     string_double (L.lexeme_start lexbuf) lexbuf;
+     string (L.lexeme_start lexbuf) lexbuf;
      let s = get_stored_string () in
-     return_loc start (L.lexeme_end lexbuf) (STRING2 s)
- | "'" ->
-     let start = L.lexeme_start lexbuf in
-     let b = string_simple (L.lexeme_start lexbuf) 0 lexbuf in
-     let s = get_stored_string () in
-     return_loc start (L.lexeme_end lexbuf) (if b then STRING1(s) else PVAR(s))
+     return_loc start (L.lexeme_end lexbuf) (STRING s)
+ | "'" -> return lexbuf SQUOTE
  | "(*" ->
      in_comment := true;
      comment (L.lexeme_start lexbuf) lexbuf;
@@ -241,12 +235,8 @@ and comment start = lexer
       comment start lexbuf
   | "*)" ->
       ()
-  | '"' ->
-      string_double (L.lexeme_start lexbuf) lexbuf;
-      clear_buff ();
-      comment start lexbuf
-  | "'" ->
-      ignore(string_simple (L.lexeme_start lexbuf) 0 lexbuf);
+  | '"' | "'" ->
+      string (L.lexeme_start lexbuf) lexbuf;
       clear_buff ();
       comment start lexbuf
   | eof ->
@@ -262,38 +252,20 @@ and tcomment start = lexer
   | _ ->
       tcomment start lexbuf
 
-and string_double start = lexer
+and string start = lexer
   | '"' -> ()
   | '\\' ['\\' '"' '\''] ->
-    store_ascii (L.latin1_lexeme_char lexbuf 1); string_double start lexbuf
-  | "\\n" -> store_ascii '\n'; string_double start lexbuf
-  | "\\t" -> store_ascii '\t'; string_double start lexbuf
-  | "\\r" -> store_ascii '\r'; string_double start lexbuf
+    store_ascii (L.latin1_lexeme_char lexbuf 1); string start lexbuf
+  | "\\n" -> store_ascii '\n'; string start lexbuf
+  | "\\t" -> store_ascii '\t'; string start lexbuf
+  | "\\r" -> store_ascii '\r'; string start lexbuf
   | '\\' ['0'-'9']+ ';' ->
-    store_code (parse_char lexbuf 10 1); string_double start lexbuf
+    store_code (parse_char lexbuf 10 1); string start lexbuf
   | '\\' 'x' ['0'-'9' 'a'-'f' 'A'-'F']+ ';' ->
-      store_code (parse_char lexbuf 16 2); string_double start lexbuf
+      store_code (parse_char lexbuf 16 2); string start lexbuf
   | '\\' -> illegal lexbuf;
-  | eof -> error start (start+1) "Unterminated string_double"
-  | _ -> store_lexeme lexbuf; string_double start lexbuf
-
-and string_simple start nb = lexer
-  | "'" -> true
-  | ' ' -> if nb != 0 then false
-    else (store_lexeme lexbuf; string_simple (nb+1) start lexbuf)
-  | '\\' ['\\' '"' '\''] -> store_ascii (L.latin1_lexeme_char lexbuf 1);
-    string_simple (nb+1) start lexbuf
-  | "\\n" -> store_ascii '\n'; string_simple (nb+1) start lexbuf
-  | "\\t" -> store_ascii '\t'; string_simple (nb+1) start lexbuf
-  | "\\r" -> store_ascii '\r'; string_simple (nb+1) start lexbuf
-  | '\\' ['0'-'9']+ ';' ->
-    store_code (parse_char lexbuf 10 1); string_simple (nb+1) start lexbuf
-  | '\\' 'x' ['0'-'9' 'a'-'f' 'A'-'F']+ ';' ->
-      store_code (parse_char lexbuf 16 2); string_simple (nb+1) start lexbuf
-  | '\\' -> illegal lexbuf;
-  | eof -> error start (start+1) "Unterminated string_simple (nb+1)"
-  | _ -> store_lexeme lexbuf; string_simple (nb+1) start lexbuf
-
+  | eof -> error start (start+1) "Unterminated string"
+  | _ -> store_lexeme lexbuf; string start lexbuf
 
 let lexbuf = ref None
 let last_tok = ref (KEYWORD "DUMMY")

parser/ulexer.mli

@@ -3,11 +3,10 @@ open Camlp4.Sig
 type token =
   | KEYWORD of string
   | IDENT of string
-  | PVAR of string
   | ANY_IN_NS of string
   | INT of string
-  | STRING1 of string
-  | STRING2 of string
+  | STRING of string
+  | SQUOTE
   | EOI
 
 exception Error of int * int * string