sway_parse/

token.rs

use core::mem;
use extension_trait::extension_trait;
use num_bigint::BigUint;
use sway_ast::literal::{LitChar, LitInt, LitIntType, LitString, Literal};
use sway_ast::token::{
    Comment, CommentKind, CommentedGroup, CommentedTokenStream, CommentedTokenTree, DocComment,
    DocStyle, Punct, Spacing, TokenStream,
};
use sway_error::error::CompileError;
use sway_error::handler::{ErrorEmitted, Handler};
use sway_error::lex_error::{LexError, LexErrorKind};
use sway_types::span::Source;
use sway_types::{
    ast::{Delimiter, PunctKind},
    Ident, SourceId, Span, Spanned,
};
use unicode_bidi::format_chars::{ALM, FSI, LRE, LRI, LRM, LRO, PDF, PDI, RLE, RLI, RLM, RLO};
use unicode_xid::UnicodeXID;

#[extension_trait]
impl CharExt for char {
    /// Converts the character into an opening delimiter, if any.
    fn as_open_delimiter(self) -> Option<Delimiter> {
        match self {
            '(' => Some(Delimiter::Parenthesis),
            '{' => Some(Delimiter::Brace),
            '[' => Some(Delimiter::Bracket),
            _ => None,
        }
    }

    /// Converts the character into a closing delimiter, if any.
    fn as_close_delimiter(self) -> Option<Delimiter> {
        match self {
            ')' => Some(Delimiter::Parenthesis),
            '}' => Some(Delimiter::Brace),
            ']' => Some(Delimiter::Bracket),
            _ => None,
        }
    }

    /// Determines what sort of punctuation this character is, if any.
    fn as_punct_kind(self) -> Option<PunctKind> {
        match self {
            ';' => Some(PunctKind::Semicolon),
            ':' => Some(PunctKind::Colon),
            '/' => Some(PunctKind::ForwardSlash),
            ',' => Some(PunctKind::Comma),
            '*' => Some(PunctKind::Star),
            '+' => Some(PunctKind::Add),
            '-' => Some(PunctKind::Sub),
            '<' => Some(PunctKind::LessThan),
            '>' => Some(PunctKind::GreaterThan),
            '=' => Some(PunctKind::Equals),
            '.' => Some(PunctKind::Dot),
            '!' => Some(PunctKind::Bang),
            '%' => Some(PunctKind::Percent),
            '&' => Some(PunctKind::Ampersand),
            '^' => Some(PunctKind::Caret),
            '|' => Some(PunctKind::Pipe),
            '_' => Some(PunctKind::Underscore),
            '#' => Some(PunctKind::Sharp),
            _ => None,
        }
    }
}

struct CharIndicesInner<'a> {
    src: &'a str,
    position: usize,
}

impl Iterator for CharIndicesInner<'_> {
    type Item = (usize, char);

    fn next(&mut self) -> Option<(usize, char)> {
        let mut char_indices = self.src[self.position..].char_indices();
        let (_, c) = char_indices.next()?;
        let ret = (self.position, c);
        match char_indices.next() {
            Some((char_width, _)) => self.position += char_width,
            None => self.position = self.src.len(),
        };
        Some(ret)
    }
}

type CharIndices<'a> = std::iter::Peekable<CharIndicesInner<'a>>;
type Result<T> = core::result::Result<T, ErrorEmitted>;

struct Lexer<'l> {
    handler: &'l Handler,
    src: &'l Source,
    source_id: &'l Option<SourceId>,
    stream: &'l mut CharIndices<'l>,
}

pub fn lex(
    handler: &Handler,
    src: Source,
    start: usize,
    end: usize,
    source_id: Option<SourceId>,
) -> Result<TokenStream> {
    lex_commented(handler, src, start, end, &source_id).map(|stream| stream.strip_comments())
}

/// Identifier & path validation.
///
/// A *path* may optionally start with `::` and is otherwise a `::`‑separated
/// list of identifiers.  Identifiers follow these rules:
///
/// * Must not be empty.
/// * Must not be just `_`.
/// * Must not start with two underscores (`__`).
/// * First char: Unicode XID_Start or `_`.
/// * Remaining chars: Unicode XID_Continue.
///
/// Any colon that is *not* part of a `::` token is rejected, and empty path
/// segments such as `foo::` or `foo:::bar` are invalid.
pub fn is_valid_identifier_or_path(s: &str) -> bool {
    // Reject empty string early.
    if s.is_empty() {
        return false;
    }

    // Handle an optional leading `::`.
    let mut input = s;
    if let Some(rest) = input.strip_prefix("::") {
        input = rest;
        // Bare `::` is invalid.
        if input.is_empty() {
            return false;
        }
    }

    // Split on *exactly* two consecutive colons.  Any single `:` or triple
    // `:::` will leave stray `:` characters inside a segment and fail below.
    for segment in input.split("::") {
        if !is_valid_identifier(segment) {
            return false;
        }
    }

    true
}

/// Check a single identifier segment.
fn is_valid_identifier(ident: &str) -> bool {
    // Reject empty, bare underscore, or double‑underscore prefix.
    if ident.is_empty() || ident == "_" || ident.starts_with("__") {
        return false;
    }

    let mut chars = ident.chars();
    let first = chars.next().unwrap();

    // First char: Unicode XID_Start or underscore.
    if !(first.is_xid_start() || first == '_') {
        return false;
    }

    // Remaining chars: Unicode XID_Continue.
    chars.all(|c| c.is_xid_continue())
}

pub fn lex_commented(
    handler: &Handler,
    src: Source,
    start: usize,
    end: usize,
    source_id: &Option<SourceId>,
) -> Result<CommentedTokenStream> {
    let stream = &mut CharIndicesInner {
        src: &src.text[..end],
        position: start,
    }
    .peekable();
    let mut l = Lexer {
        handler,
        src: &src,
        source_id,
        stream,
    };
    let mut file_start_offset: usize = 0;

    let mut parent_token_trees = Vec::new();
    let mut token_trees = Vec::new();
    while let Some((mut index, mut character)) = l.stream.next() {
        if character.is_whitespace() {
            // if the beginning of a file starts with whitespace
            // we must keep track to ensure that the module level docs
            // will get inserted into the tree correctly
            if index - file_start_offset == 0 {
                file_start_offset += character.len_utf8();
            }
            continue;
        }
        if character == '/' {
            match l.stream.peek() {
                Some((_, '/')) => {
                    // search_end is the index at which we stop looking backwards for
                    // a newline
                    let search_end = token_trees
                        .last()
                        .map(|tt| {
                            if let CommentedTokenTree::Tree(t) = tt {
                                t.span().end()
                            } else {
                                0
                            }
                        })
                        .unwrap_or_default();

                    let has_newline = src.text[search_end..index]
                        .chars()
                        .rev()
                        .take_while(|c| c.is_whitespace())
                        .filter(|&c| c == '\n')
                        .count()
                        > 0;
                    // We found a comment at the start of file, which should be accounted for as a Newlined comment.
                    let start_of_file_found = search_end == 0 && index == 0;

                    let comment_kind = if has_newline || start_of_file_found {
                        CommentKind::Newlined
                    } else {
                        CommentKind::Trailing
                    };

                    let ctt = lex_line_comment(&mut l, end, index, comment_kind);
                    token_trees.push(ctt);
                    continue;
                }
                Some((_, '*')) => {
                    if let Some(token) = lex_block_comment(&mut l, index) {
                        token_trees.push(token);
                    }
                    continue;
                }
                Some(_) | None => {}
            }
        }

        if character.is_xid_start() || character == '_' {
            // Raw identifier, e.g., `r#foo`? Then mark as such, stripping the prefix `r#`.
            let is_raw_ident = character == 'r' && matches!(l.stream.peek(), Some((_, '#')));
            if is_raw_ident {
                l.stream.next();
                if let Some((next_index, next_character)) = l.stream.next() {
                    character = next_character;
                    index = next_index;
                }
                if !(character.is_xid_start() || character == '_') {
                    let kind = LexErrorKind::InvalidCharacter {
                        position: index,
                        character,
                    };
                    let span = span_one(&l, index, character);
                    error(l.handler, LexError { kind, span });
                    continue;
                }
            }

            // Don't accept just `_` as an identifier.
            let not_is_single_underscore = character != '_'
                || l.stream
                    .peek()
                    .is_some_and(|(_, next)| next.is_xid_continue());
            if not_is_single_underscore {
                // Consume until we hit other than `XID_CONTINUE`.
                while l.stream.next_if(|(_, c)| c.is_xid_continue()).is_some() {}
                let ident = Ident::new_with_raw(span_until(&mut l, index), is_raw_ident);
                token_trees.push(CommentedTokenTree::Tree(ident.into()));
                continue;
            }
        }
        if let Some(delimiter) = character.as_open_delimiter() {
            let token_trees = mem::take(&mut token_trees);
            parent_token_trees.push((token_trees, index, delimiter));
            continue;
        }
        if let Some(close_delimiter) = character.as_close_delimiter() {
            match parent_token_trees.pop() {
                None => {
                    // Recover by ignoring the unexpected closing delim,
                    // giving the parser opportunities to realize the need for an opening delim
                    // in e.g., this example:
                    //
                    // fn foo() // <-- Parser expects grouped tokens in `{ ... }` here.
                    //     let x = 0;
                    // } // <- This recovery.
                    let kind = LexErrorKind::UnexpectedCloseDelimiter {
                        position: index,
                        close_delimiter,
                    };
                    let span = span_one(&l, index, character);
                    error(l.handler, LexError { kind, span });
                }
                Some((parent, open_index, open_delimiter)) => {
                    if open_delimiter != close_delimiter {
                        // Recover on e.g., a `{ )` mismatch by having `)` interpreted as `}`.
                        let kind = LexErrorKind::MismatchedDelimiters {
                            open_position: open_index,
                            close_position: index,
                            open_delimiter,
                            close_delimiter,
                        };
                        let span = span_one(&l, index, character);
                        error(l.handler, LexError { kind, span });
                    }
                    token_trees = lex_close_delimiter(
                        &mut l,
                        index,
                        parent,
                        token_trees,
                        open_index,
                        open_delimiter,
                    );
                }
            }
            continue;
        }
        if let Some(token) = lex_string(&mut l, index, character)? {
            token_trees.push(token);
            continue;
        }
        if let Some(token) = lex_char(&mut l, index, character)? {
            token_trees.push(token);
            continue;
        }
        if let Some(token) = lex_int_lit(&mut l, index, character)? {
            token_trees.push(token);
            continue;
        }
        if let Some(token) = lex_punctuation(&mut l, index, character) {
            token_trees.push(token);
            continue;
        }

        // Recover by simply ignoring the character.
        // NOTE(Centril): I'm not sure how good of an idea this is... time will tell.
        let kind = LexErrorKind::InvalidCharacter {
            position: index,
            character,
        };
        let span = span_one(&l, index, character);
        error(l.handler, LexError { kind, span });
        continue;
    }

    // Recover all unclosed delimiters.
    while let Some((parent, open_index, open_delimiter)) = parent_token_trees.pop() {
        let kind = LexErrorKind::UnclosedDelimiter {
            open_position: open_index,
            open_delimiter,
        };
        let span = span_one(&l, open_index, open_delimiter.as_open_char());
        error(l.handler, LexError { kind, span });

        token_trees = lex_close_delimiter(
            &mut l,
            src.text.len(),
            parent,
            token_trees,
            open_index,
            open_delimiter,
        );
    }
    Ok(CommentedTokenStream {
        token_trees,
        full_span: span(&l, start, end),
    })
}

fn lex_close_delimiter(
    l: &mut Lexer<'_>,
    index: usize,
    mut parent: Vec<CommentedTokenTree>,
    token_trees: Vec<CommentedTokenTree>,
    open_index: usize,
    delimiter: Delimiter,
) -> Vec<CommentedTokenTree> {
    let start_index = open_index + delimiter.as_open_char().len_utf8();
    let full_span = span(l, start_index, index);
    let group = CommentedGroup {
        token_stream: CommentedTokenStream {
            token_trees,
            full_span,
        },
        delimiter,
        span: span_until(l, open_index),
    };
    parent.push(CommentedTokenTree::Tree(group.into()));
    parent
}

fn lex_line_comment(
    l: &mut Lexer<'_>,
    end: usize,
    index: usize,
    comment_kind: CommentKind,
) -> CommentedTokenTree {
    let _ = l.stream.next();

    // Find end; either at EOF or at `\n`.
    let end = l
        .stream
        .find(|(_, character)| *character == '\n')
        .map_or(end, |(end, _)| end);
    let sp = span(l, index, end);

    let doc_style = match (sp.as_str().chars().nth(2), sp.as_str().chars().nth(3)) {
        // `//!` is an inner line doc comment.
        (Some('!'), _) => Some(DocStyle::Inner),
        // `////` (more than 3 slashes) is not considered a doc comment.
        (Some('/'), Some('/')) => None,
        // `///` is an outer line doc comment.
        (Some('/'), _) => Some(DocStyle::Outer),
        _ => None,
    };

    if let Some(doc_style) = doc_style {
        let doc_comment = DocComment {
            span: sp,
            doc_style,
            content_span: span(l, index + 3, end),
        };
        CommentedTokenTree::Tree(doc_comment.into())
    } else {
        Comment {
            span: sp,
            comment_kind,
        }
        .into()
    }
}

fn lex_block_comment(l: &mut Lexer<'_>, index: usize) -> Option<CommentedTokenTree> {
    // Lexing a multi-line comment.
    let _ = l.stream.next();
    let mut unclosed_indices = vec![index];

    let unclosed_multiline_comment = |l: &Lexer<'_>, unclosed_indices: Vec<_>| {
        let span = span(l, *unclosed_indices.last().unwrap(), l.src.text.len() - 1);
        let kind = LexErrorKind::UnclosedMultilineComment { unclosed_indices };
        error(l.handler, LexError { kind, span });
        None
    };

    // We first start by assuming that block comments are inlined.
    let mut comment_kind = CommentKind::Inlined;

    loop {
        match l.stream.next() {
            None => return unclosed_multiline_comment(l, unclosed_indices),
            Some((_, '*')) => match l.stream.next() {
                None => return unclosed_multiline_comment(l, unclosed_indices),
                // Matched `*/`, so we're closing some multi-line comment. It could be nested.
                Some((slash_ix, '/')) => {
                    let start = unclosed_indices.pop().unwrap();
                    if unclosed_indices.is_empty() {
                        // For the purposes of lexing,
                        // nested multi-line comments constitute a single multi-line comment.
                        // We could represent them as several ones, but that's unnecessary.
                        let end = slash_ix + '/'.len_utf8();
                        let span = span(l, start, end);
                        return Some(Comment { span, comment_kind }.into());
                    }
                }
                Some(_) => {}
            },
            // Found nested multi-line comment.
            Some((next_index, '/')) => match l.stream.next() {
                None => return unclosed_multiline_comment(l, unclosed_indices),
                Some((_, '*')) => unclosed_indices.push(next_index),
                Some(_) => {}
            },
            Some((_, '\n')) => {
                // If we find a newline character while lexing, this means that the block comment is multiline.
                // Example:
                // /* this is a
                //    multilined block comment */
                comment_kind = CommentKind::Multilined;
            }
            Some(_) => {}
        }
    }
}

fn lex_string(
    l: &mut Lexer<'_>,
    index: usize,
    character: char,
) -> Result<Option<CommentedTokenTree>> {
    if character != '"' {
        return Ok(None);
    }
    let mut parsed = String::new();
    loop {
        let unclosed_string_lit = |l: &Lexer<'_>, end| {
            error(
                l.handler,
                LexError {
                    kind: LexErrorKind::UnclosedStringLiteral { position: index },
                    span: span(l, index, end),
                },
            )
        };
        let (next_index, next_character) = l.stream.next().ok_or_else(|| {
            // last character may not be a unicode boundary
            let mut end = l.src.text.len() - 1;
            while !l.src.text.is_char_boundary(end) {
                end -= 1;
            }
            unclosed_string_lit(l, end)
        })?;
        parsed.push(match next_character {
            '\\' => parse_escape_code(l)
                .map_err(|e| e.unwrap_or_else(|| unclosed_string_lit(l, l.src.text.len())))?,
            '"' => break,
            // do not allow text direction codepoints
            ALM | FSI | LRE | LRI | LRM | LRO | PDF | PDI | RLE | RLI | RLM | RLO => {
                let kind = LexErrorKind::UnicodeTextDirInLiteral {
                    position: next_index,
                    character: next_character,
                };
                let span = span_one(l, next_index, next_character);
                error(l.handler, LexError { span, kind });
                continue;
            }
            _ => next_character,
        });
    }
    let span = span_until(l, index);
    let literal = Literal::String(LitString { span, parsed });
    Ok(Some(CommentedTokenTree::Tree(literal.into())))
}

fn lex_char(
    l: &mut Lexer<'_>,
    index: usize,
    character: char,
) -> Result<Option<CommentedTokenTree>> {
    let is_quote = |c| c == '\'';
    if !is_quote(character) {
        return Ok(None);
    }

    let unclosed_char_lit = |l: &Lexer<'_>| {
        let err = LexError {
            kind: LexErrorKind::UnclosedCharLiteral { position: index },
            span: span(l, index, l.src.text.len()),
        };
        error(l.handler, err)
    };
    let next = |l: &mut Lexer<'_>| l.stream.next().ok_or_else(|| unclosed_char_lit(l));
    let escape = |l: &mut Lexer<'_>, next_char| {
        if next_char == '\\' {
            parse_escape_code(l).map_err(|e| e.unwrap_or_else(|| unclosed_char_lit(l)))
        } else {
            Ok(next_char)
        }
    };

    let (next_index, next_char) = next(l)?;
    // do not allow text direction codepoints
    if let ALM | FSI | LRE | LRI | LRM | LRO | PDF | PDI | RLE | RLI | RLM | RLO = next_char {
        let kind = LexErrorKind::UnicodeTextDirInLiteral {
            position: next_index,
            character: next_char,
        };
        let span = span_one(l, next_index, next_char);
        error(l.handler, LexError { span, kind });
    }

    let parsed = escape(l, next_char)?;

    // Consume the closing `'`.
    let (next_index, next_char) = next(l)?;
    let sp = span_until(l, index);

    // Not a closing quote? Then this is e.g., 'ab'.
    // Most likely the user meant a string literal, so recover as that instead.
    let literal = if !is_quote(next_char) {
        let mut string = String::new();
        string.push(parsed);
        string.push(escape(l, next_char)?);
        loop {
            let (_, next_char) = next(l)?;
            if is_quote(next_char) {
                break;
            }
            string.push(next_char);
        }

        // Emit the expected closing quote error.
        error(
            l.handler,
            LexError {
                kind: LexErrorKind::ExpectedCloseQuote {
                    position: next_index,
                },
                span: span(l, next_index, next_index + string.len()),
            },
        );

        Literal::String(LitString {
            span: sp,
            parsed: string,
        })
    } else {
        Literal::Char(LitChar { span: sp, parsed })
    };

    Ok(Some(CommentedTokenTree::Tree(literal.into())))
}

fn parse_escape_code(l: &mut Lexer<'_>) -> core::result::Result<char, Option<ErrorEmitted>> {
    let error = |kind, span| Err(Some(error(l.handler, LexError { kind, span })));

    match l.stream.next() {
        None => Err(None),
        Some((_, '"')) => Ok('"'),
        Some((_, '\'')) => Ok('\''),
        Some((_, 'n')) => Ok('\n'),
        Some((_, 'r')) => Ok('\r'),
        Some((_, 't')) => Ok('\t'),
        Some((_, '\\')) => Ok('\\'),
        Some((_, '0')) => Ok('\0'),
        Some((index, 'x')) => {
            let (high, low) = match (l.stream.next(), l.stream.next()) {
                (Some((_, high)), Some((_, low))) => (high, low),
                _ => return Err(None),
            };
            let (high, low) = match (high.to_digit(16), low.to_digit(16)) {
                (Some(high), Some(low)) => (high, low),
                _ => return error(LexErrorKind::InvalidHexEscape, span_until(l, index)),
            };
            let parsed_character = char::from_u32((high << 4) | low).unwrap();
            Ok(parsed_character)
        }
        Some((index, 'u')) => {
            match l.stream.next() {
                None => return Err(None),
                Some((_, '{')) => (),
                Some((_, unexpected_char)) => {
                    let span = span_one(l, index, unexpected_char);
                    let kind = LexErrorKind::UnicodeEscapeMissingBrace { position: index };
                    return error(kind, span);
                }
            }
            let mut digits_start_position_opt = None;
            let mut char_value = BigUint::from(0u32);
            let digits_end_position = loop {
                let (position, digit) = match l.stream.next() {
                    None => return Err(None),
                    Some((position, '}')) => break position,
                    Some((position, digit)) => (position, digit),
                };
                if digits_start_position_opt.is_none() {
                    digits_start_position_opt = Some(position);
                };
                let digit = match digit.to_digit(16) {
                    None => {
                        let span = span_one(l, position, digit);
                        let kind = LexErrorKind::InvalidUnicodeEscapeDigit { position };
                        return error(kind, span);
                    }
                    Some(digit) => digit,
                };
                char_value *= 16u32;
                char_value += digit;
            };
            let digits_start_position = digits_start_position_opt.unwrap_or(digits_end_position);
            let char_value = match u32::try_from(char_value) {
                Err(..) => {
                    let span = span(l, digits_start_position, digits_end_position);
                    let kind = LexErrorKind::UnicodeEscapeOutOfRange { position: index };
                    return error(kind, span);
                }
                Ok(char_value) => char_value,
            };
            let parsed_character = match char::from_u32(char_value) {
                None => {
                    let span_all = span_until(l, index);
                    let kind = LexErrorKind::UnicodeEscapeInvalidCharValue { span: span_all };
                    let span = span(l, digits_start_position, digits_end_position);
                    return error(kind, span);
                }
                Some(parsed_character) => parsed_character,
            };
            Ok(parsed_character)
        }
        Some((index, unexpected_char)) => error(
            LexErrorKind::InvalidEscapeCode { position: index },
            span_one(l, index, unexpected_char),
        ),
    }
}

fn lex_int_lit(
    l: &mut Lexer<'_>,
    index: usize,
    character: char,
) -> Result<Option<CommentedTokenTree>> {
    let digit = match character.to_digit(10) {
        None => return Ok(None),
        Some(d) => d,
    };

    let decimal_int_lit = |l, digit: u32| {
        let mut big_uint = BigUint::from(digit);
        let end_opt = parse_digits(&mut big_uint, l, 10);
        (big_uint, end_opt)
    };
    let (big_uint, end_opt) = if digit == 0 {
        let prefixed_int_lit = |l: &mut Lexer<'_>, radix| {
            let _ = l.stream.next();
            let d = l.stream.next();
            let incomplete_int_lit = |end| {
                let kind = match radix {
                    16 => LexErrorKind::IncompleteHexIntLiteral { position: index },
                    8 => LexErrorKind::IncompleteOctalIntLiteral { position: index },
                    2 => LexErrorKind::IncompleteBinaryIntLiteral { position: index },
                    _ => unreachable!(),
                };
                let span = span(l, index, end);
                error(l.handler, LexError { kind, span })
            };
            let (digit_pos, digit) = d.ok_or_else(|| incomplete_int_lit(l.src.text.len()))?;
            let radix_digit = digit
                .to_digit(radix)
                .ok_or_else(|| incomplete_int_lit(digit_pos))?;
            let mut big_uint = BigUint::from(radix_digit);
            let end_opt = parse_digits(&mut big_uint, l, radix);
            Ok((big_uint, end_opt))
        };

        match l.stream.peek() {
            Some((_, 'x')) => prefixed_int_lit(l, 16)?,
            Some((_, 'o')) => prefixed_int_lit(l, 8)?,
            Some((_, 'b')) => prefixed_int_lit(l, 2)?,
            Some((_, '_' | '0'..='9')) => decimal_int_lit(l, 0),
            Some(&(next_index, _)) => (BigUint::from(0u32), Some(next_index)),
            None => (BigUint::from(0u32), None),
        }
    } else {
        decimal_int_lit(l, digit)
    };

    let ty_opt = lex_int_ty_opt(l)?;

    let literal = Literal::Int(LitInt {
        span: span(l, index, end_opt.unwrap_or(l.src.text.len())),
        parsed: big_uint,
        ty_opt,
        is_generated_b256: false,
    });

    Ok(Some(CommentedTokenTree::Tree(literal.into())))
}

fn lex_int_ty_opt(l: &mut Lexer<'_>) -> Result<Option<(LitIntType, Span)>> {
    let (suffix_start_position, c) = match l.stream.next_if(|(_, c)| c.is_xid_continue()) {
        None => return Ok(None),
        Some(x) => x,
    };
    let mut suffix = String::from(c);
    let suffix_end_position = loop {
        match l.stream.peek() {
            Some((_, c)) if c.is_xid_continue() => {
                suffix.push(*c);
                let _ = l.stream.next();
            }
            Some((pos, _)) => break *pos,
            None => break l.src.text.len(),
        }
    };
    // Parse the suffix to a known one, or if unknown, recover by throwing it away.
    let ty = match parse_int_suffix(&suffix) {
        Some(s) => s,
        None => {
            let span = span(l, suffix_start_position, suffix_end_position);
            let kind = LexErrorKind::InvalidIntSuffix {
                suffix: Ident::new(span.clone()),
            };
            error(l.handler, LexError { kind, span });
            return Ok(None);
        }
    };
    let span = span_until(l, suffix_start_position);
    Ok(Some((ty, span)))
}

/// Interpret the given `suffix` string as a `LitIntType`.
pub fn parse_int_suffix(suffix: &str) -> Option<LitIntType> {
    Some(match suffix {
        "u8" => LitIntType::U8,
        "u16" => LitIntType::U16,
        "u32" => LitIntType::U32,
        "u64" => LitIntType::U64,
        "u256" => LitIntType::U256,
        "i8" => LitIntType::I8,
        "i16" => LitIntType::I16,
        "i32" => LitIntType::I32,
        "i64" => LitIntType::I64,
        _ => return None,
    })
}

fn parse_digits(big_uint: &mut BigUint, l: &mut Lexer<'_>, radix: u32) -> Option<usize> {
    loop {
        match l.stream.peek() {
            None => break None,
            Some((_, '_')) => {
                let _ = l.stream.next();
            }
            Some(&(index, character)) => match character.to_digit(radix) {
                None => break Some(index),
                Some(digit) => {
                    let _ = l.stream.next();
                    *big_uint *= radix;
                    *big_uint += digit;
                }
            },
        };
    }
}

fn lex_punctuation(l: &mut Lexer<'_>, index: usize, character: char) -> Option<CommentedTokenTree> {
    let punct = Punct {
        kind: character.as_punct_kind()?,
        spacing: match l.stream.peek() {
            Some((_, next_character)) if next_character.as_punct_kind().is_some() => Spacing::Joint,
            _ => Spacing::Alone,
        },
        span: span_until(l, index),
    };
    Some(CommentedTokenTree::Tree(punct.into()))
}

fn span_until(l: &mut Lexer<'_>, start: usize) -> Span {
    let end = l.stream.peek().map_or(l.src.text.len(), |(end, _)| *end);
    span(l, start, end)
}

fn span_one(l: &Lexer<'_>, start: usize, c: char) -> Span {
    span(l, start, start + c.len_utf8())
}

fn span(l: &Lexer<'_>, start: usize, end: usize) -> Span {
    Span::new(l.src.clone(), start, end, *l.source_id).unwrap()
}

/// Emit a lexer error.
fn error(handler: &Handler, error: LexError) -> ErrorEmitted {
    handler.emit_err(CompileError::Lex { error })
}

#[cfg(test)]
mod tests {
    use super::*;
    use assert_matches::assert_matches;
    use sway_ast::{
        literal::{LitChar, Literal},
        token::{
            Comment, CommentKind, CommentedTokenTree, CommentedTree, DocComment, DocStyle,
            TokenTree,
        },
    };
    use sway_error::{
        error::CompileError,
        handler::Handler,
        lex_error::{LexError, LexErrorKind},
    };

    #[test]
    fn lex_bidi() {
        let input = "
            script;
            use std::string::String;
            fn main() {
                let a = String::from_ascii_str(\"fuel\");
                let b = String::from_ascii_str(\"fuel\u{202E}\u{2066}// Same string again\u{2069}\u{2066}\");
                if a.as_bytes() == b.as_bytes() {
                    log(\"same\");
                } else {
                    log(\"different\");
                }
                let lrm = '\u{202E}';
                log(lrm);
            }
        ";
        let start = 0;
        let end = input.len();
        let path = None;
        let handler = Handler::default();
        let _stream = lex_commented(&handler, input.into(), start, end, &path).unwrap();
        let (errors, warnings, infos) = handler.consume();
        assert_eq!(infos.len(), 0);
        assert_eq!(warnings.len(), 0);
        assert_eq!(errors.len(), 5);
        for err in errors {
            assert_matches!(
                err,
                CompileError::Lex {
                    error: LexError {
                        span: _,
                        kind: LexErrorKind::UnicodeTextDirInLiteral {
                            position: _,
                            character: _
                        }
                    }
                }
            );
        }
    }

    #[test]
    fn lex_commented_token_stream() {
        let input = r#"
        //
        // Single-line comment.
        struct Foo {
            /* multi-
             * line-
             * comment */
            bar: i32, // trailing comment
        }
        "#;
        let start = 0;
        let end = input.len();
        let path = None;
        let handler = Handler::default();
        let stream = lex_commented(&handler, input.into(), start, end, &path).unwrap();
        assert!(handler.consume().0.is_empty());
        let mut tts = stream.token_trees().iter();
        assert_eq!(tts.next().unwrap().span().as_str(), "//");
        assert_eq!(
            tts.next().unwrap().span().as_str(),
            "// Single-line comment."
        );
        assert_eq!(tts.next().unwrap().span().as_str(), "struct");
        assert_eq!(tts.next().unwrap().span().as_str(), "Foo");
        {
            let group = match tts.next() {
                Some(CommentedTokenTree::Tree(CommentedTree::Group(group))) => group,
                _ => panic!("expected group"),
            };
            let mut tts = group.token_stream.token_trees().iter();
            assert_eq!(
                tts.next().unwrap().span().as_str(),
                "/* multi-\n             * line-\n             * comment */",
            );
            assert_eq!(tts.next().unwrap().span().as_str(), "bar");
            assert_eq!(tts.next().unwrap().span().as_str(), ":");
            assert_eq!(tts.next().unwrap().span().as_str(), "i32");
            assert_eq!(tts.next().unwrap().span().as_str(), ",");
            assert_matches!(
                tts.next(),
                Some(CommentedTokenTree::Comment(Comment {
                    span,
                    comment_kind: CommentKind::Trailing,
                })) if span.as_str() ==  "// trailing comment"
            );
            assert!(tts.next().is_none());
        }
        assert!(tts.next().is_none());
    }

    #[test]
    fn lex_comments_check_comment_kind() {
        let input = r#"
        // CommentKind::Newlined
        abi Foo {
            // CommentKind::Newlined
            fn bar(); // CommentKind::Trailing
            // CommentKind::Newlined
        }
        "#;
        let start = 0;
        let end = input.len();
        let path = None;
        let handler = Handler::default();
        let stream = lex_commented(&handler, input.into(), start, end, &path).unwrap();
        assert!(handler.consume().0.is_empty());
        let mut tts = stream.token_trees().iter();

        assert_matches!(
            tts.next(),
            Some(CommentedTokenTree::Comment(Comment {
                span,
                comment_kind: CommentKind::Newlined,
            })) if span.as_str() ==  "// CommentKind::Newlined"
        );
        assert_eq!(tts.next().unwrap().span().as_str(), "abi");
        assert_eq!(tts.next().unwrap().span().as_str(), "Foo");

        {
            let group = match tts.next() {
                Some(CommentedTokenTree::Tree(CommentedTree::Group(group))) => group,
                _ => panic!("expected group"),
            };
            let mut tts = group.token_stream.token_trees().iter();

            assert_matches!(
                tts.next(),
                Some(CommentedTokenTree::Comment(Comment {
                    span,
                    comment_kind: CommentKind::Newlined,
                })) if span.as_str() ==  "// CommentKind::Newlined"
            );
            assert_eq!(tts.next().unwrap().span().as_str(), "fn");
            assert_eq!(tts.next().unwrap().span().as_str(), "bar");
            assert_eq!(tts.next().unwrap().span().as_str(), "()");
            assert_eq!(tts.next().unwrap().span().as_str(), ";");
            assert_matches!(
                tts.next(),
                Some(CommentedTokenTree::Comment(Comment {
                    span,
                    comment_kind: CommentKind::Trailing,
                })) if span.as_str() ==  "// CommentKind::Trailing"
            );
            assert_matches!(
                tts.next(),
                Some(CommentedTokenTree::Comment(Comment {
                    span,
                    comment_kind: CommentKind::Newlined,
                })) if span.as_str() ==  "// CommentKind::Newlined"
            );
            assert!(tts.next().is_none());
        }
    }

    #[test]
    fn lex_doc_comments() {
        let input = r#"
        //none
        ////none
        //!inner
        //! inner
        ///outer
        /// outer
        "#;
        let start = 0;
        let end = input.len();
        let path = None;
        let handler = Handler::default();
        let stream = lex_commented(&handler, input.into(), start, end, &path).unwrap();
        assert!(handler.consume().0.is_empty());
        let mut tts = stream.token_trees().iter();
        assert_matches!(
            tts.next(),
            Some(CommentedTokenTree::Comment(Comment {
                span,
                comment_kind: CommentKind::Newlined,
            })) if span.as_str() ==  "//none"
        );
        assert_matches!(
            tts.next(),
            Some(CommentedTokenTree::Comment(Comment {
                span,
                comment_kind: CommentKind::Newlined,
            })) if span.as_str() ==  "////none"
        );
        assert_matches!(
            tts.next(),
            Some(CommentedTokenTree::Tree(CommentedTree::DocComment(DocComment {
                doc_style: DocStyle::Inner,
                span,
                content_span
            }))) if span.as_str() ==  "//!inner" && content_span.as_str() == "inner"
        );
        assert_matches!(
            tts.next(),
            Some(CommentedTokenTree::Tree(CommentedTree::DocComment(DocComment {
                doc_style: DocStyle::Inner,
                span,
                content_span
            }))) if span.as_str() ==  "//! inner" && content_span.as_str() == " inner"
        );
        assert_matches!(
            tts.next(),
            Some(CommentedTokenTree::Tree(CommentedTree::DocComment(DocComment {
                doc_style: DocStyle::Outer,
                span,
                content_span
            }))) if span.as_str() ==  "///outer" && content_span.as_str() == "outer"
        );
        assert_matches!(
            tts.next(),
            Some(CommentedTokenTree::Tree(CommentedTree::DocComment(DocComment {
                doc_style: DocStyle::Outer,
                span,
                content_span
            }))) if span.as_str() ==  "/// outer" && content_span.as_str() == " outer"
        );
        assert_eq!(tts.next(), None);
    }

    #[test]
    fn lex_char_escaped_quote() {
        let input = r"
        '\''
        ";
        let handler = Handler::default();
        let stream = lex(&handler, input.into(), 0, input.len(), None).unwrap();
        assert!(handler.consume().0.is_empty());
        let mut tts = stream.token_trees().iter();
        assert_matches!(
            tts.next(),
            Some(TokenTree::Literal(Literal::Char(LitChar {
                parsed: '\'',
                ..
            })))
        );
        assert_eq!(tts.next(), None);
    }

    use super::is_valid_identifier_or_path as valid;

    #[test]
    fn accepts_simple_identifiers() {
        assert!(valid("foo"));
        assert!(valid("Foo"));
        assert!(valid("_foo"));
        assert!(valid("foo123"));
        assert!(valid("føø"));
    }

    #[test]
    fn rejects_invalid_identifiers() {
        assert!(!valid(""));
        assert!(!valid("_"));
        assert!(!valid("__"));
        assert!(!valid("__invalid"));
        assert!(!valid(":foo"));
        assert!(!valid("foo:bar"));
    }

    #[test]
    fn accepts_paths() {
        assert!(valid("foo::bar"));
        assert!(valid("_foo::_bar"));
        assert!(valid("foo_bar::baz123"));
        assert!(valid("::some_module::in_the_same::package"));
    }

    #[test]
    fn rejects_malformed_paths() {
        assert!(!valid("foo:bar:baz"));
        assert!(!valid("foo::"));
        assert!(!valid("::"));
        assert!(!valid("foo:::bar"));
        assert!(!valid("foo::__bad"));
    }
}