From 63676de59f9d21193bb4ee1bfbb3768d10452b7f Mon Sep 17 00:00:00 2001
From: Priya Srikumar <priyasrikumar1@gmail.com>
Date: Wed, 2 Aug 2023 22:14:06 -0700
Subject: [PATCH] parser tweaks

---
 vcalyx/ocaml/lexer.mll  | 10 +++++++---
 vcalyx/ocaml/parser.mly |  4 ++--
 vcalyx/ocaml/vcx.ml     | 32 +++++++++++++++++++++++++++-----
 3 files changed, 36 insertions(+), 10 deletions(-)

diff --git a/vcalyx/ocaml/lexer.mll b/vcalyx/ocaml/lexer.mll
index 4fe67b4fdc..8c7ac23a0d 100644
--- a/vcalyx/ocaml/lexer.mll
+++ b/vcalyx/ocaml/lexer.mll
@@ -1,13 +1,17 @@
 { 
   open Parser 
-  exception ParseError of string
+  exception SyntaxError of string
 }
 
 let id = ['a'-'z' 'A'-'Z' '_'] ['a'-'z' 'A'-'Z' '0'-'9' '_' '.']*
+let whitespace = [' ' '\t']+
+let newline = '\r' | '\n' | "\r\n"
 
 rule tokens = parse 
 (* i.e., 1'd1 *)
-| ['0'-'9']+"'d"['0'-'9']+ as i { INT (int_of_string i) } 
+| ['0'-'9']+ as i { INT (int_of_string i) } 
+| whitespace      { tokens lexbuf }
+| newline         { Lexing.new_line lexbuf; tokens lexbuf }
 | "("             { LPAREN }
 | ")"             { RPAREN }
 | "components"    { COMPONENTS }
@@ -41,4 +45,4 @@ rule tokens = parse
 | "assignments"   { ASSIGNMENTS }
 | eof             { EOF }
 | id as x         { ID x }
-| _ { raise (ParseError (Printf.sprintf "At offset %d: unexpected character.\n" (Lexing.lexeme_start lexbuf))) }
\ No newline at end of file
+| _ { raise (SyntaxError (Printf.sprintf "At offset %d: unexpected character %s" (Lexing.lexeme_start lexbuf) (Lexing.lexeme lexbuf))) }
\ No newline at end of file
diff --git a/vcalyx/ocaml/parser.mly b/vcalyx/ocaml/parser.mly
index 1d2de03eed..7a4304f254 100644
--- a/vcalyx/ocaml/parser.mly
+++ b/vcalyx/ocaml/parser.mly
@@ -17,13 +17,13 @@
 %token DIRECTION
 %token ASSIGNMENTS
 
-%start <Extr.context> main
+%start <Extr.context option> main
 %%
 
 main: 
 | LPAREN; LPAREN; COMPONENTS; LPAREN; comps = list(component); RPAREN; RPAREN; 
 LPAREN; ENTRYPOINT; entry = ID; RPAREN; RPAREN; EOF
-  { {ctx_comps = comps; ctx_entrypoint = entry} }
+  { Some {ctx_comps = comps; ctx_entrypoint = entry} }
 
 attrs_clause:
   | LPAREN; ATTRIBUTES; LPAREN; attrs = list(attribute); RPAREN; RPAREN
diff --git a/vcalyx/ocaml/vcx.ml b/vcalyx/ocaml/vcx.ml
index b84128b4d5..fc83099e49 100644
--- a/vcalyx/ocaml/vcx.ml
+++ b/vcalyx/ocaml/vcx.ml
@@ -1,5 +1,28 @@
 open Core
 open Vcalyx
+open Lexing
+
+(* from https://dev.realworldocaml.org/parsing-with-ocamllex-and-menhir.html *)
+let print_position outx lexbuf =
+  let pos = lexbuf.lex_curr_p in
+  fprintf outx "%s:%d:%d" pos.pos_fname pos.pos_lnum
+    (pos.pos_cnum - pos.pos_bol + 1)
+
+let parse_with_error lexbuf =
+  try Parser.main Lexer.tokens lexbuf with
+  (* | SyntaxError msg ->
+    fprintf stderr "%a: %s\n" print_position lexbuf msg;
+    None *)
+  | Parser.Error ->
+    fprintf stderr "%a: syntax error\n" print_position lexbuf;
+    exit (-1)
+
+let rec parse_and_print source_str source_location =
+  match parse_with_error source_str with
+  | Some _ ->
+    Printf.printf "Successfully parsed %s.\n" source_location;
+    parse_and_print source_str source_location
+  | None -> ()
 
 let vcx_parse : Command.t =
   let open Command.Let_syntax in
@@ -9,14 +32,13 @@ let vcx_parse : Command.t =
       fun () ->
         let source_chan = In_channel.create source_location in
         let source_str = Lexing.from_channel source_chan in
-        source_str.lex_curr_p <- { source_str.lex_curr_p with pos_fname = source_location };
-        In_channel.close source_chan;
-        match Parser.main Lexer.tokens source_str with
-        | _ -> Printf.printf "Successfully parsed %s.\n" source_location]
+        source_str.lex_curr_p <-
+          { source_str.lex_curr_p with pos_fname = source_location };
+        parse_and_print source_str source_location;
+        In_channel.close source_chan]
 
 let vcx_cmd : Command.t =
   Command.group ~summary:"vcx: the vcalyx command-line interface"
     [ ("parse", vcx_parse) ]
-(* ("json-test", json_test)  *)
 
 let () = Command_unix.run ~version:"dev" vcx_cmd