libreda-lefdef 0.0.5

// SPDX-FileCopyrightText: 2021-2023 Thomas Kramer <code@tkramer.ch>
//
// SPDX-License-Identifier: AGPL-3.0-or-later

//! Functions for parsing ASCII-based formats from iterators over bytes.

use itertools::{Itertools, PeekingNext};
use std::fmt;
use std::iter::Peekable;
use std::num::ParseIntError;
use std::str::FromStr;

use libreda_stream_parser::{Lexer, ParserError};

/// Error while parsing LEF or DEF.
/// TODO: Separate lexer errors from LEF/DEF specific errors.
#[derive(Clone, Debug)]
pub enum LefDefParseError {
    /// Error during parsing.
    ParserError(ParserError<char>),
    /// Encountered invalid character.
    InvalidCharacter,
    /// Reached end of file before end of library arrived.
    UnexpectedEndOfFile,
    /// Expected and actual token.
    UnexpectedToken(String, String),
    /// Unknown token. The token is given as a string.
    UnknownToken(String),
    /// Unknown literal. The literal is given as a string.
    InvalidLiteral(String),
    /// Illegal value for bus bit chars.
    IllegalBusBitChars(char, char),
    /// Something is not yet implemented.
    NotImplemented(&'static str),
    /// Using a property name that has not been defined in PROPERTYDEFINITIONS.
    UndefinedProperty(String),
    /// Failed to parse an integer.
    ParseIntError(ParseIntError),
    /// Some other error defined by a string.
    Other(&'static str),
}

impl From<ParserError<char>> for LefDefParseError {
    fn from(e: ParserError<char>) -> Self {
        Self::ParserError(e)
    }
}

impl fmt::Display for LefDefParseError {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        match self {
            LefDefParseError::InvalidCharacter => write!(f, "Invalid character."),
            LefDefParseError::UnexpectedEndOfFile => write!(f, "Unexpected end of file."),
            LefDefParseError::UnexpectedToken(actual, exp) => {
                write!(f, "Unexpected token. '{}' instead of '{}'", actual, exp)
            }
            LefDefParseError::UnknownToken(t) => write!(f, "Unknown token: '{}'.", t),
            LefDefParseError::InvalidLiteral(n) => write!(f, "Invalid literal: '{}'.", n),
            LefDefParseError::IllegalBusBitChars(a, b) => {
                write!(f, "Illegal bus bit chars: '{} {}'.", a, b)
            }
            LefDefParseError::NotImplemented(n) => write!(f, "Not implemented: '{}'.", n),
            LefDefParseError::UndefinedProperty(p) => write!(f, "Undefined property: '{}'.", p),
            LefDefParseError::Other(msg) => write!(f, "'{}'.", msg),
            LefDefParseError::ParseIntError(e) => write!(f, "Illegal integer: '{}'", e),
            LefDefParseError::ParserError(e) => write!(f, "{}", e),
        }
    }
}

impl From<ParseIntError> for LefDefParseError {
    fn from(e: ParseIntError) -> Self {
        Self::ParseIntError(e)
    }
}

pub struct LefDefLexer {}

impl Lexer for LefDefLexer {
    type Char = char;

    fn consume_next_token(
        &mut self,
        iter: &mut (impl Iterator<Item = char> + PeekingNext),
        mut output: impl FnMut(char),
    ) -> Result<(), ParserError<char>> {
        loop {
            // Skip whitespace.
            let _n = iter.peeking_take_while(|c| c.is_whitespace()).count();

            // Look ahead.
            if let Some(c) = iter.peeking_next(|_| true) {
                debug_assert!(!c.is_whitespace());

                match c {
                    '#' => {
                        // Skip comments.
                        iter.peeking_take_while(|&c| c != '\n' && c != '\r').count();
                    }
                    '"' | '\'' => {
                        // Quoted string.
                        let quote_char = c;

                        let mut prev = None;
                        while let Some(c) = iter.next() {
                            if prev != Some('\\') && c == quote_char {
                                // Abort on quote char.
                                break;
                            }
                            output(c);
                            prev = Some(c);
                        }
                        return Ok(());
                    }
                    _ => {
                        // Normal token.
                        let mut prev = Some(c);
                        output(c);

                        while let Some(c) = iter.next() {
                            if prev != Some('\\') && c.is_whitespace() {
                                // Abort on unmasked whitespace.
                                break;
                            }

                            output(c);
                            prev = Some(c);
                        }
                        return Ok(());
                    }
                }
            } else {
                return Ok(());
            }
        }
    }
}

/// Read simple tokens and skip comments.
#[test]
fn test_read_token() {
    let data = r#"
        # Comment 1

        # Comment 2

        token1

        # Comment 3

        token2 token3

        "quoted token"

        token4
    "#;

    let mut iter = data.chars().inspect(|c| print!("{}", c)).peekable();

    let mut buffer = String::new();

    let mut tk = libreda_stream_parser::tokenize(iter, LefDefLexer {});

    tk.advance().unwrap();

    tk.expect_str("token1").unwrap();
    tk.expect_str("token2").unwrap();
    tk.expect_str("token3").unwrap();
    tk.expect_str("quoted token").unwrap();
    tk.expect_str("token4").unwrap();
    assert!(tk.current_token_ref().is_none());
}