schala/schala-lang/src/parsing/combinator.rs

use std::{cell::RefCell, rc::Rc};

use nom::{
    branch::alt,
    bytes::complete::{tag, take_till},
    character::complete::{
        alpha1, alphanumeric0, char, line_ending, multispace0, none_of, not_line_ending, one_of, space0,
        space1,
    },
    combinator::{map, not, opt, peek, recognize, value},
    error::{context, ParseError, VerboseError},
    multi::{fold_many1, many0, many1, separated_list0, separated_list1},
    sequence::{pair, preceded, tuple},
    Err, IResult, Parser,
};
use nom_locate::{position, LocatedSpan};

use crate::identifier::{Id, IdStore};

type StoreRef = Rc<RefCell<IdStore<ASTItem>>>;

type Span<'a> = LocatedSpan<&'a str, StoreRef>;
type ParseResult<'a, O> = IResult<Span<'a>, O, VerboseError<Span<'a>>>;

use crate::ast::*;

fn rc_string(s: &str) -> Rc<String> {
    Rc::new(s.to_string())
}
fn fresh_id(span: &Span) -> Id<ASTItem> {
    let mut table_handle = span.extra.borrow_mut();
    table_handle.fresh()
}

fn tok<'a, O>(
    input_parser: impl Parser<Span<'a>, O, VerboseError<Span<'a>>>,
) -> impl FnMut(Span<'a>) -> IResult<Span<'a>, O, VerboseError<Span<'a>>> {
    context("tok", map(tuple((ws0, input_parser)), |(_, output)| output))
}

fn kw<'a>(keyword_str: &'static str) -> impl FnMut(Span<'a>) -> ParseResult<()> {
    context("keyword", tok(value((), tag(keyword_str))))
}

// whitespace does consume at least one piece of whitespace - use ws0 for maybe none
fn whitespace(input: Span) -> ParseResult<()> {
    context("whitespace", alt((block_comment, line_comment, value((), space1))))(input)
}

fn ws0(input: Span) -> ParseResult<()> {
    context("WS0", value((), many0(whitespace)))(input)
}

fn line_comment(input: Span) -> ParseResult<()> {
    value((), tuple((tag("//"), not_line_ending)))(input)
}

fn block_comment(input: Span) -> ParseResult<()> {
    context(
        "Block-comment",
        value(
            (),
            tuple((
                tag("/*"),
                many0(alt((value((), none_of("*/")), value((), none_of("/*")), block_comment))),
                tag("*/"),
            )),
        ),
    )(input)
}

fn statement_delimiter(input: Span) -> ParseResult<()> {
    tok(alt((value((), line_ending), value((), char(';')))))(input)
}

fn block(input: Span) -> ParseResult<Block> {
    context(
        "block",
        map(
            tuple((
                tok(char('{')),
                many0(statement_delimiter),
                separated_list0(statement_delimiter, statement),
                many0(statement_delimiter),
                tok(char('}')),
            )),
            |(_, _, items, _, _)| items.into(),
        ),
    )(input)
}

fn statement(input: Span) -> ParseResult<Statement> {
    let (input, pos) = position(input)?;
    let pos: usize = pos.location_offset();
    let id = fresh_id(&input);
    context(
        "Parsing-statement",
        map(expression, move |expr| Statement {
            id,
            location: pos.into(),
            kind: StatementKind::Expression(expr),
        }),
    )(input)
}

fn expression(input: Span) -> ParseResult<Expression> {
    let id = fresh_id(&input);
    map(pair(expression_kind, opt(type_anno)), move |(kind, maybe_anno)| Expression::new(id, kind))(input)
}

fn type_anno(input: Span) -> ParseResult<TypeIdentifier> {
    preceded(kw(":"), type_identifier)(input)
}

fn type_identifier(input: Span) -> ParseResult<TypeIdentifier> {
    /*
    alt((
        tuple((kw("("), separated_list0(kw(","), type_identifier), kw(")"))),
        type_singleton_name
    ))(input)
    */
    unimplemented!()
}

fn type_singleton_name(input: Span) -> ParseResult<TypeSingletonName> {
    unimplemented!()
}

pub fn expression_kind(input: Span) -> ParseResult<ExpressionKind> {
    context("expression-kind", primary_expr)(input)
}

fn primary_expr(input: Span) -> ParseResult<ExpressionKind> {
    context("primary-expr", alt((number_literal, bool_literal, identifier_expr)))(input)
}

fn identifier_expr(input: Span) -> ParseResult<ExpressionKind> {
    context("identifier-expr", map(qualified_identifier, ExpressionKind::Value))(input)
}

fn qualified_identifier(input: Span) -> ParseResult<QualifiedName> {
    let id = fresh_id(&input);
    tok(map(separated_list1(tag("::"), map(identifier, |x| rc_string(x.fragment()))), move |items| {
        QualifiedName { id, components: items }
    }))(input)
}

fn identifier(input: Span) -> ParseResult<Span> {
    recognize(tuple((alt((tag("_"), alpha1)), alphanumeric0)))(input)
}

fn bool_literal(input: Span) -> ParseResult<ExpressionKind> {
    context(
        "bool-literal",
        alt((
            map(kw("true"), |_| ExpressionKind::BoolLiteral(true)),
            map(kw("false"), |_| ExpressionKind::BoolLiteral(false)),
        )),
    )(input)
}

fn number_literal(input: Span) -> ParseResult<ExpressionKind> {
    map(alt((tok(hex_literal), tok(bin_literal), tok(dec_literal))), ExpressionKind::NatLiteral)(input)
}

fn dec_literal(input: Span) -> ParseResult<u64> {
    map(digits(digit_group_dec), |chars: Vec<char>| {
        let s: String = chars.into_iter().collect();
        s.parse().unwrap()
    })(input)
}

fn hex_literal(input: Span) -> ParseResult<u64> {
    map(preceded(alt((tag("0x"), tag("0X"))), digits(digit_group_hex)), |chars: Vec<char>| {
        let s: String = chars.into_iter().collect();
        parse_hex(&s).unwrap()
    })(input)
}

fn bin_literal(input: Span) -> ParseResult<u64> {
    map(preceded(alt((tag("0b"), tag("0B"))), digits(digit_group_bin)), |chars: Vec<char>| {
        let s: String = chars.into_iter().collect();
        parse_binary(&s).unwrap()
    })(input)
}

fn digits<'a, E: ParseError<Span<'a>>>(
    digit_type: impl Parser<Span<'a>, Vec<char>, E>,
) -> impl FnMut(Span<'a>) -> IResult<Span<'a>, Vec<char>, E> {
    map(separated_list1(many1(char('_')), digit_type), |items: Vec<Vec<char>>| {
        items.into_iter().flatten().collect()
    })
}

fn digit_group_dec(input: Span) -> ParseResult<Vec<char>> {
    many1(one_of("0123456789"))(input)
}

fn digit_group_hex(input: Span) -> ParseResult<Vec<char>> {
    many1(one_of("0123456789abcdefABCDEF"))(input)
}

fn digit_group_bin(input: Span) -> ParseResult<Vec<char>> {
    many1(one_of("01"))(input)
}

fn parse_binary(digits: &str) -> Result<u64, &'static str> {
    let mut result: u64 = 0;
    let mut multiplier = 1;
    for d in digits.chars().rev() {
        match d {
            '1' => result += multiplier,
            '0' => (),
            '_' => continue,
            _ => unreachable!(),
        }
        multiplier = match multiplier.checked_mul(2) {
            Some(m) => m,
            None => return Err("Binary expression will overflow"),
        }
    }
    Ok(result)
}

fn parse_hex(digits: &str) -> Result<u64, &'static str> {
    let mut result: u64 = 0;
    let mut multiplier: u64 = 1;
    for d in digits.chars().rev() {
        if d == '_' {
            continue;
        }
        match d.to_digit(16) {
            Some(n) => result += n as u64 * multiplier,
            None => return Err("Internal parser error: invalid hex digit"),
        }
        multiplier = match multiplier.checked_mul(16) {
            Some(m) => m,
            None => return Err("Hexadecimal expression will overflow"),
        }
    }
    Ok(result)
}

#[cfg(test)]
mod test {
    use pretty_assertions::assert_eq;

    use super::*;

    fn rc(s: &str) -> Rc<String> {
        Rc::new(s.to_owned())
    }
    macro_rules! qn {
        ( $( $component:ident),* ) => {
            {
                let mut components = vec![];
                $(
                    components.push(rc(stringify!($component)));
                )*
                    QualifiedName { components, id: Default::default() }
            }
        };
    }

    macro_rules! span {
        ($func:expr, $input:expr) => {{
            let id_store: IdStore<ASTItem> = IdStore::new();
            let span = Span::new_extra($input, Rc::new(RefCell::new(id_store)));
            $func(span).map(|(span, x)| (*span.fragment(), x))
        }};
    }

    #[test]
    fn combinator_test1() {
        assert_eq!(span!(digits(digit_group_dec), "342"), Ok(("", vec!['3', '4', '2'])));
        assert_eq!(span!(bin_literal, "0b1111qsdf"), Ok(("qsdf", 15)));
    }

    #[test]
    fn combinator_test_ws0() {
        assert_eq!(span!(block_comment, "/*yolo*/"), Ok(("", ())));
        assert_eq!(span!(block_comment, "/*yolo*/ jumpy /*nah*/"), Ok((" jumpy /*nah*/", ())));
        assert_eq!(span!(ws0, "/*    yolo  */ "), Ok(("", ())));
        assert_eq!(span!(ws0, "/*  /* no */   yolo  */ "), Ok(("", ())));
    }

    #[test]
    fn combinator_test2() {
        for s in ["  15", "   0b1111", "   1_5_", "0XF__", "0Xf"].iter() {
            assert_eq!(span!(expression_kind, s).unwrap().1, ExpressionKind::NatLiteral(15));
        }

        assert_eq!(span!(expression_kind, " /*gay*/  true").unwrap().1, ExpressionKind::BoolLiteral(true));
        assert_eq!(
            span!(expression_kind, " /*yolo*/  barnaby").unwrap().1,
            ExpressionKind::Value(qn!(barnaby))
        );
    }

    #[test]
    fn combinator_test3() {
        let source = "{}";
        assert_eq!(span!(block, source).unwrap().1, vec![].into());
        let source = r#"{

            //hella
            4_5 //bog
            11; /*chutney*/0xf
            }"#;
        let parsed = span!(block, source).map_err(|err| match err {
            Err::Error(err) | Err::Failure(err) => {
                let err = VerboseError {
                    errors: err.errors.into_iter().map(|(sp, kind)| (*sp.fragment(), kind)).collect(),
                };
                nom::error::convert_error(source, err)
            }
            _ => panic!(),
        });

        if let Err(err) = parsed {
            println!("{}", err);
            panic!("parse error desu!");
        }

        assert_eq!(
            parsed.unwrap().1,
            vec![
                Statement {
                    id: Default::default(),
                    location: Default::default(),
                    kind: StatementKind::Expression(Expression::new(
                        Default::default(),
                        ExpressionKind::NatLiteral(45)
                    ))
                },
                Statement {
                    id: Default::default(),
                    location: Default::default(),
                    kind: StatementKind::Expression(Expression::new(
                        Default::default(),
                        ExpressionKind::NatLiteral(11)
                    ))
                },
                Statement {
                    id: Default::default(),
                    location: Default::default(),
                    kind: StatementKind::Expression(Expression::new(
                        Default::default(),
                        ExpressionKind::NatLiteral(15)
                    ))
                },
            ]
            .into()
        );
    }
}