js-regex 0.1.0

// Copyright (C) 2020 Quentin M. Kniep <hello@quentinkniep.com>
// Distributed under terms of the MIT license.

use std::collections::VecDeque;

#[derive(Debug)]
pub struct Reader {
    unicode: bool,
    src: String,
    index: usize,
    end: usize,
    cps: VecDeque<char>,
    widths: VecDeque<usize>,
}

impl Reader {
    pub fn new() -> Self {
        Self {
            unicode: false,
            src: "".to_owned(),
            index: 0,
            end: 0,
            cps: VecDeque::with_capacity(4),
            widths: VecDeque::with_capacity(3),
        }
    }

    pub fn source(&self) -> &str {
        &self.src
    }

    pub fn index(&self) -> usize {
        self.index
    }

    pub fn code_point_with_offset(&self, offset: usize) -> Option<char> {
        self.cps.get(offset).cloned()
    }

    pub fn reset(&mut self, source: &str, start: usize, end: usize, u_flag: bool) {
        self.unicode = u_flag;
        self.src = source.to_owned();
        self.end = end;
        self.rewind(start);
    }

    pub fn rewind(&mut self, index: usize) {
        self.index = index;
        self.cps.clear();
        self.widths.clear();
        for i in 0..4 {
            let w_sum: usize = self.widths.iter().take(i).sum();
            if let Some(c) = self.at(index + w_sum) {
                self.cps.push_back(c);
                self.widths.push_back(self.width(c));
            } else {
                break;
            }
        }
    }

    pub fn advance(&mut self) {
        if self.cps.get(0).is_some() {
            self.index += self.widths[0];
            self.cps.pop_front();
            self.widths.pop_front();
            let w_sum: usize = self.widths.iter().sum();
            if let Some(c) = self.at(self.index + w_sum) {
                self.widths.push_back(self.width(*self.cps.back().unwrap()));
                self.cps.push_back(c);
            }
        }
    }

    pub fn eat(&mut self, cp: char) -> bool {
        let opt = self.cps.get(0);
        if opt.is_some() && *opt.unwrap() == cp {
            self.advance();
            return true;
        }
        return false;
    }

    pub fn eat2(&mut self, cp1: char, cp2: char) -> bool {
        let (opt1, opt2) = (self.cps.get(0), self.cps.get(1));
        if opt1.is_some() && opt2.is_some() && *opt1.unwrap() == cp1 && *opt2.unwrap() == cp2 {
            self.advance();
            self.advance();
            return true;
        }
        return false;
    }

    pub fn eat3(&mut self, cp1: char, cp2: char, cp3: char) -> bool {
        let (opt1, opt2, opt3) = (self.cps.get(0), self.cps.get(1), self.cps.get(2));
        if opt1.is_some()
            && opt2.is_some()
            && opt3.is_some()
            && *opt1.unwrap() == cp1
            && *opt2.unwrap() == cp2
            && *opt3.unwrap() == cp3
        {
            self.advance();
            self.advance();
            self.advance();
            return true;
        }
        return false;
    }

    fn at(&self, i: usize) -> Option<char> {
        //println!("{:?}", self.src.as_bytes());
        //println!("{:?}", self.src.encode_utf16().collect::<Vec<u16>>());
        if i >= self.end {
            None
        } else if self.unicode {
            // TODO: read non ASCII as UTF-8
            //let c: char = self.src.as_bytes()[i].into();
            let c = self.src.chars().skip(i).next().unwrap();
            Some(c)
        } else {
            // TODO: move the conversion out of this method and make it safe
            unsafe {
                let c: char = std::char::from_u32_unchecked(self.src.encode_utf16().skip(i).next().unwrap() as u32);
                Some(c)
            }
        }
    }

    fn width(&self, c: char) -> usize {
        if self.unicode && c > '\u{FFFF}' {
            1
        } else {
            1
        }
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn eat_test() {
        let mut reader = Reader::new();
        reader.reset("abcdefghijk", 0, 11, true);
        assert_eq!(reader.eat('a'), true);
        assert_eq!(reader.eat3('b', 'd', 'd'), false);
        assert_eq!(reader.eat3('b', 'c', 'd'), true);
        assert_eq!(reader.eat2('e', 'f'), true);
        assert_eq!(reader.eat('h'), false);
        assert_eq!(reader.eat('g'), true);
        assert_eq!(reader.eat2('h', 'i'), true);
        assert_eq!(reader.eat3('j', 'k', 'a'), false);
    }

    #[test]
    fn rewind_test() {
        let mut reader = Reader::new();
        reader.reset("abcd", 0, 4, true);
        assert_eq!(reader.eat('a'), true);
        assert_eq!(reader.eat3('b', 'd', 'd'), false);
        assert_eq!(reader.eat3('b', 'c', 'd'), true);
        reader.rewind(0);
        assert_eq!(reader.eat('a'), true);
        assert_eq!(reader.eat3('b', 'd', 'd'), false);
        assert_eq!(reader.eat3('b', 'c', 'd'), true);
    }

    #[test]
    fn at_test_es_compliance() {
        let mut reader = Reader::new();
        // without unicode flag
        reader.reset("Hello", 0, 5, false);
        assert_eq!(reader.at(1).unwrap() as u32, 101);
        reader.reset("􀃃a🩢☃★♲", 0, 6, false);
        assert_eq!(reader.at(0).unwrap() as u32, 56256);
        reader.reset("􀃃ello", 0, 6, false);
        assert_eq!(reader.at(0).unwrap() as u32, 56256);
        reader.reset("􀃃ello", 0, 6, false);
        assert_eq!(reader.at(1).unwrap() as u32, 56515);
        // with unicode flag
        reader.reset("Hello", 0, 5, true);
        assert_eq!(reader.at(1).unwrap() as u32, 101);
        reader.reset("􀃃a🩢☃★♲", 0, 6, true);
        assert_eq!(reader.at(0).unwrap() as u32, 1048771);
        reader.reset("􀃃ello", 0, 6, true);
        assert_eq!(reader.at(0).unwrap() as u32, 1048771);
    }
}