Repair not following grammar

[duskybomb]

Hi, I was playing with your amazing tool and used a grammar very similar to the "calculator evaluator" given in Quickstart Guide.

I tried an arbitrary buggy calculator program: 1 + * ( 1 + ) ( + * 1 + 1 * ) ( * 1 )

These were the suggested repair sequences:

Parsing error at line 1 column 5. Repair sequences found:
   1: Delete *
   2: Insert INT
start (line: 1, col: 5), end (line: 1, col:6)

Parsing error at line 1 column 13. Repair sequences found:
   1: Delete ), Shift (, Delete +, Delete *
   2: Delete ), Shift (, Insert INT, Delete +
   3: Insert INT, Shift ), Delete (, Delete +
   4: Delete ), Shift (, Insert INT, Shift +, Insert INT
   5: Insert INT, Shift ), Delete (, Shift +, Insert INT
   6: Insert INT, Shift ), Delete (, Shift +, Delete *
   7: Delete ), Shift (, Insert INT, Shift +, Delete *

Parsing error at line 1 column 29. Repair sequences found:
   1: Insert INT, Shift ), Delete (
   2: Delete ), Shift (, Insert INT
start (line: 1, col: 29), end (line: 1, col:30)

However, the repaired program we get using the first repair sequence is: 1 + ( 1 + ( 1 + 1 * INT ) * 1 ), which doesn't conform to the grammar and the tool rejects the repair upon sending it through again. (gives error). Is there something that I missed?

calc.y

%start Expr
%%
Expr -> Result<u64, ()>:
      Expr 'PLUS' Term { Ok($1? + $3?) }
    | Term { $1 }
    ;

Term -> Result<u64, ()>:
      Term 'MUL' Factor { Ok($1? * $3?) }
    | Factor { $1 }
    ;

Factor -> Result<u64, ()>:
      'PARENOPEN' Expr 'PARENCLOSE' { $2 }
    | 'INT'
      {
          let v = $1.map_err(|_| ())?;
          parse_int($lexer.span_str(v.span()))
      }
    ;
%%
// Any functions here are in scope for all the grammar actions above.

fn parse_int(s: &str) -> Result<u64, ()> {
    match s.parse::<u64>() {
        Ok(val) => Ok(val),
        Err(_) => {
            eprintln!("{} cannot be represented as a u64", s);
            Err(())
        }
    }
}

calc.l

%%
[0-9]+ "INT"
\+ "PLUS"
\* "MUL"
\( "PARENOPEN"
\) "PARENCLOSE"
[\t ]+ ;

[ltratt]

If I try this in lrpar/examples/calc_actions I get:

>>> 1 + * ( 1 + ) ( + * 1 + 1 * ) ( * 1 )
Parsing error at line 1 column 5. Repair sequences found:
   1: Delete *
   2: Insert INT
Parsing error at line 1 column 13. Repair sequences found:
   1: Delete ), Shift (, Delete +, Delete *
   2: Delete ), Shift (, Insert INT, Delete +
   3: Insert INT, Shift ), Delete (, Delete +
   4: Delete ), Shift (, Insert INT, Shift +, Insert INT
   5: Insert INT, Shift ), Delete (, Shift +, Delete *
   6: Insert INT, Shift ), Delete (, Shift +, Insert INT
   7: Delete ), Shift (, Insert INT, Shift +, Delete *
Parsing error at line 1 column 29. Repair sequences found:
   1: Insert INT, Shift ), Delete (
   2: Delete ), Shift (, Insert INT
<evaluation aborted>

as you do. If I then follow its suggestions (noting that "Insert INT" means "insert a value of type INT" so I'll use 2 as an example) I get:

>>> 1 + ( 1 + ( 1 + 1 * 2 ) * 1 )  
Result: 5

which looks OK to me.

I guess it's the "insert INT" bit that causes you to make an input that can't be parsed? There is a UI issue here: sometimes we refer to lexeme values (e.g. Delete )) and sometimes lexeme types (e.g. Insert INT). I flapped around a bit on this originally because there's not an easy way in plain text to make clear that e.g. Insert ''' means "Insert a single quote".

The user can use the %epp feature to make the difference between these clearer though, perhaps, it would be better if the default was something like Insert <INT>. That would solve the "single quote" (and related problem), and I think the angle bracket problem is less of an issue.

[duskybomb]

Thank you for such an elaborate response. I don't know how did I miss that this was due to inserting INT and not an syntactic error. Extremely sorry for that!

However, I have come across the repaired program not conforming to grammar issue with one of the repairs being:

1 + ( 1 + ( INT + INT * 1 + 1 * ( * 1 )
Input program: 1 + * ( 1 + ) ( + * 1 + 1 * ) ( * 1 )
If I follow the repair sequence: 1, 4, 2 from the one mentioned in my first post.

1: Delete *
4: Delete ), Shift (, Insert INT, Shift +, Insert INT
2: Delete ), Shift (, Insert INT

I hope this isn't another silly mistake from my side.

[duskybomb]

If I use the %epp feature and mention %epp INT "1", I still get the <evaluation aborted> message. Is it because of timeout or something else?

[ltratt]

Ah, you can't intermix repair sequences in that way: depending on which repair sequence you take, the rest of the parse (and the parse errors) can be different. There's no practical way around this, unfortunately, without hitting a combinatorial explosion.

[duskybomb]

Thank you for the clarification!