Cleanup parser

Author: davesnx

source/Parser.mly

@@ -35,7 +35,6 @@
 %token CLOSE_BRACE
 %token EOF
 
-/* %left OPEN_BRACKET */
 /* according to https://github.com/stedolan/jq/issues/1326 */
 %right PIPE /* lowest precedence */
 %nonassoc COMMA
@@ -55,28 +54,29 @@ program:
   | EOF;
     { Identity }
 
-str_or_id:
+string_or_identifier:
   | key = IDENTIFIER { Literal (String key) }
   | key = STRING { Literal (String key) }
 
-key_val(E):
-  | key = str_or_id
+key_value (E):
+  | key = string_or_identifier
     { key, None }
   | OPEN_PARENT; e1 = E CLOSE_PARENT; COLON; e2 = E
     { e1, Some e2 }
-  | key = str_or_id; COLON; e = E
+  | key = string_or_identifier; COLON; e = E
     { key, Some e }
 
 elif_term:
   | ELIF cond = item_expr THEN e = term
     { cond, e }
 
 // sequence_expr handles the lowest precedence operators: comma and pipe
+// while item_expr handles the higher precedence operators
 sequence_expr:
   | left = sequence_expr; COMMA; right = sequence_expr;
     { Comma (left, right) }
 
-  | left = sequence_expr; PIPE; right = item_expr; // Pipe binds tighter than comma, but less than others
+  | left = sequence_expr; PIPE; right = item_expr;
     { Pipe (left, right) }
 
   | e = item_expr
@@ -96,10 +96,10 @@ sequence_expr:
   | AND {And}
   | OR {Or}
 
-// item_expr handles operators with higher precedence than COMMA and PIPE
 item_expr:
   | left = item_expr; op = operator; right = item_expr;
     { Operation (left, op, right) }
+
   | e = term
     { e }
 
@@ -124,8 +124,7 @@ term:
     { Literal(Null) }
   | RANGE; OPEN_PARENT; nl = separated_nonempty_list(SEMICOLON, number); CLOSE_PARENT;
     {
-      let nl = List.map int_of_float nl in
-      match nl with
+      match (List.map Int.of_float nl) with
       | [] -> assert false (* nonempty_list *)
       | x :: [] -> Range (x, None, None)
       | x :: y :: [] -> Range (x, Some y, None)
@@ -211,12 +210,13 @@ term:
   | OPEN_BRACE; CLOSE_BRACE;
     { Object [] }
 
-  | e = delimited(OPEN_BRACE, separated_nonempty_list(COMMA, key_val(term)), CLOSE_BRACE);
+  | e = delimited(OPEN_BRACE, separated_nonempty_list(COMMA, key_value (term)), CLOSE_BRACE);
     { Object e }
 
   // Parentheses allow a full sequence_expr inside, reducing to an item_expr
   | OPEN_PARENT; e = sequence_expr; CLOSE_PARENT;
     { e }
+
   | e = term; OPEN_BRACKET; i = number; CLOSE_BRACKET
     { Pipe (e, Index (int_of_float i)) }
 

source/Tokenizer.ml

@@ -1,6 +1,5 @@
 open Sedlexing.Utf8
 
-let dot = [%sedlex.regexp? '.']
 let digit = [%sedlex.regexp? '0' .. '9']
 let number = [%sedlex.regexp? Plus digit, Opt '.', Opt (Plus digit)]
 let space = [%sedlex.regexp? Plus ('\n' | '\t' | ' ')]
@@ -51,7 +50,7 @@ type token =
   | EOF
 [@@deriving show]
 
-let string buf =
+let tokenize_string buf =
   let buffer = Buffer.create 10 in
   let rec read_string buf =
     [%sedlex
@@ -108,13 +107,13 @@ let rec tokenize buf =
   | "else" -> Ok ELSE
   | "elif" -> Ok ELIF
   | "end" -> Ok END
-  | dot -> Ok DOT
+  | "." -> Ok DOT
   | ".." -> Ok RECURSE
-  | '"' -> string buf
+  | '"' -> tokenize_string buf
   | identifier -> tokenize_apply buf
   | number ->
-      let num = lexeme buf |> float_of_string in
-      Ok (NUMBER num)
+      let num = lexeme buf in
+      Ok (NUMBER (Float.of_string num))
   | space -> tokenize buf
   | any -> Error ("Unexpected character '" ^ lexeme buf ^ "'")
   | _ -> Error "Unexpected character"