/*
 * meli - text_processing crate.
 *
 * Copyright 2017-2020 Manos Pitsidianakis
 *
 * This file is part of meli.
 *
 * meli is free software: you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation, either version 3 of the License, or
 * (at your option) any later version.
 *
 * meli is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with meli. If not, see <http://www.gnu.org/licenses/>.
 */

extern crate unicode_segmentation;
use self::unicode_segmentation::UnicodeSegmentation;
use super::grapheme_clusters::TextProcessing;
use super::tables::LINE_BREAK_RULES;
use super::types::LineBreakClass;
use super::types::Reflow;
use core::cmp::Ordering;
use core::iter::Peekable;
use core::str::FromStr;
use LineBreakClass::*;

#[derive(Debug, PartialEq, Copy, Clone)]
pub enum LineBreakCandidate {
    MandatoryBreak,
    BreakAllowed,
    NoBreak, // Not used.
}

impl Default for LineBreakCandidate {
    fn default() -> Self {
        LineBreakCandidate::NoBreak
    }
}

use LineBreakCandidate::*;

pub struct LineBreakCandidateIter<'a> {
    text: &'a str,
    iter: Peekable<unicode_segmentation::GraphemeIndices<'a>>,
    pos: usize,
    /* Needed for rule LB30a */
    reg_ind_streak: u32,
}

impl<'a> LineBreakCandidateIter<'a> {
    pub fn new(text: &'a str) -> Self {
        LineBreakCandidateIter {
            text,
            pos: 0,
            iter: UnicodeSegmentation::grapheme_indices(text, true).peekable(),
            reg_ind_streak: 0,
        }
    }
}

macro_rules! get_base_character {
    ($grapheme:ident) => {{
        char::from_str($grapheme.get(0..1).unwrap_or_else(|| {
            $grapheme.get(0..2).unwrap_or_else(|| {
                $grapheme
                    .get(0..3)
                    .unwrap_or_else(|| $grapheme.get(0..4).unwrap())
            })
        }))
    }};
    ($grapheme:expr) => {{
        char::from_str($grapheme.get(0..1).unwrap_or_else(|| {
            $grapheme.get(0..2).unwrap_or_else(|| {
                $grapheme
                    .get(0..3)
                    .unwrap_or_else(|| $grapheme.get(0..4).unwrap())
            })
        }))
    }};
}

/// Side effects: none
macro_rules! get_class {
    ($grapheme:ident) => {{
        get_base_character!($grapheme)
            .map(|char| search_table(char as u32, LINE_BREAK_RULES))
            .unwrap_or(XX)
    }};
    ($grapheme:expr) => {{
        get_base_character!($grapheme)
            .map(|char| search_table(char as u32, LINE_BREAK_RULES))
            .unwrap_or(XX)
    }};
}

/// Side effects: Updates $graph_iter and potentially $idx and $grapheme
macro_rules! next_grapheme_class {
    ($graph_iter:ident, $grapheme:ident) => ({
        if let Some((_, g)) = $graph_iter.next() {
            $grapheme = g;
            Some(get_class!(g))
        } else { None }
    });
    (($next_char:ident is $class:expr)) => ({
        $next_char.is_some() && get_class!(($next_char.unwrap().1)) == $class
    });
    (($next_char:ident is $($class:ident),+)) => ({
        $next_char.is_some() && ($(get_class!(($next_char.unwrap().1)) == $class)||+)
    });
}

/// Returns positions where breaks can happen
/// Examples:
/// ```
/// use melib::text_processing::{self, LineBreakCandidate::{self, *}};
/// use melib::text_processing::line_break::LineBreakCandidateIter;
///
/// assert!(LineBreakCandidateIter::new("").collect::<Vec<(usize, LineBreakCandidate)>>().is_empty());
/// assert_eq!(&[(7, BreakAllowed), (12, MandatoryBreak)],
///            LineBreakCandidateIter::new("Sample Text.").collect::<Vec<(usize, LineBreakCandidate)>>().as_slice());
/// assert_eq!(&[(3, MandatoryBreak), (7, MandatoryBreak), (10, BreakAllowed), (17, MandatoryBreak)],
///            LineBreakCandidateIter::new("Sa\nmp\r\nle T(e)xt.").collect::<Vec<(usize, LineBreakCandidate)>>().as_slice());
/// ```
impl<'a> Iterator for LineBreakCandidateIter<'a> {
    type Item = (usize, LineBreakCandidate);
    fn next(&mut self) -> Option<Self::Item> {
        loop {
            // After end of text, there are no breaks.
            if self.pos >= self.text.len() {
                return None;
            }
            // LB3 Always break at the end of text
            if self.pos + 1 == self.text.len() {
                self.pos += 1;
                return Some((self.pos, MandatoryBreak));
            }

            let LineBreakCandidateIter {
                ref mut iter,
                ref text,
                ref mut reg_ind_streak,
                ref mut pos,
            } = self;
            let (idx, mut grapheme) = iter.next().unwrap();
            let iter = iter.by_ref();

            debug_assert_eq!(idx, *pos);

            // LB2 Never break at the start of text
            if idx == 0 {
                *pos += grapheme.len();
                continue;
            }

            let class = get_class!(grapheme);

            if class != RI {
                *reg_ind_streak = 0;
            }

            /* LB1 Assign a line breaking class to each code point of the input. Resolve AI, CB, CJ,
             * SA, SG, and XX into other line breaking classes depending on criteria outside the scope
             * of this algorithm.
             *
             * In the absence of such criteria all characters with a specific combination of original
             * class and General_Category property value are resolved as follows:
             * Resolved Original     General_Category
             * AL       AI, SG, XX   Any
             * CM       SA           Only Mn or Mc
             * AL       SA           Any except Mn and Mc
             * NS       SJ           Any
             */

            // TODO: LB1

            /* Check if next character class allows breaks before it */
            let next_char: Option<&(usize, &str)> = iter.peek();

            match class {
                BK => {
                    // LB4 Always Break after hard line breaks.
                    *pos += grapheme.len();
                    return Some((*pos, MandatoryBreak));
                }
                // LB5 Treat CR followed by LF, as well as CR, LF, and NL as hard line breaks
                CR if next_grapheme_class!((next_char is LF)) => {
                    *pos += grapheme.len();
                    assert!(Some(LF) == next_grapheme_class!(iter, grapheme));
                    *pos += grapheme.len();
                    return Some((*pos, MandatoryBreak));
                }
                CR | LF | NL => {
                    *pos += grapheme.len();
                    return Some((*pos, MandatoryBreak));
                }
                _ => {}
            }
            if let Some((_, next_grapheme)) = next_char {
                let next_class = get_class!(next_grapheme);
                match next_class {
                    /* LB6 Do not break before hard line breaks.  × ( BK | CR | LF | NL ) */
                    BK | CR | LF | NL => {
                        *pos += grapheme.len();
                        continue;
                    }
                    /* LB7 Do not break before spaces or zero width
                     * space. × SP × ZW */
                    SP | ZW => {
                        *pos += grapheme.len();
                        continue;
                    }
                    _ => {}
                }
            }
            match class {
                ZW => {
                    // LB8 Break before any character following a zero-width space, even if one or more
                    // spaces intervene
                    // ZW SP* ÷
                    *pos += grapheme.len();
                    while Some(SP) == next_grapheme_class!(iter, grapheme) {
                        *pos += grapheme.len();
                    }
                    return Some((*pos, MandatoryBreak));
                }
                ZWJ => {
                    // LB8a Do not break after a zero width joiner.
                    *pos += grapheme.len();
                    continue;
                }

                CM => {
                    // LB9 Do not break a combining character sequence; treat it as if it has the line
                    // breaking class of the base character in all of the following rules. Treat ZWJ as
                    // if it were CM.
                    // Treat X (CM | ZWJ)* as if it were X.
                    // where X is any line break class except BK, CR, LF, NL, SP, or ZW.

                    *pos += grapheme.len();
                    continue;
                }
                WJ => {
                    /*: LB11 Do not break before or after Word joiner and related characters.*/
                    *pos += grapheme.len();
                    /* Get next grapheme */
                    if next_grapheme_class!(iter, grapheme).is_some() {
                        *pos += grapheme.len();
                    }
                    continue;
                }
                GL => {
                    /*LB12 Non-breaking characters: LB12 Do not break after NBSP and related characters.*/
                    *pos += grapheme.len();
                    continue;
                }
                _ => {}
            }
            if let Some((next_idx, next_grapheme)) = next_char {
                let next_class = get_class!(next_grapheme);
                match next_class {
                    GL if ![SP, BA, HY].contains(&class) => {
                        /* LB12a Do not break before NBSP and related characters, except after spaces and
                         * hyphens.  [^SP BA HY] × GL
                         * Also LB12 Do not break after NBSP and related characters */
                        *pos += grapheme.len();
                        continue;
                    }
                    /* LB13 Do not break before ‘]’ or ‘!’ or ‘;’ or ‘/’, even after spaces. */
                    CL | CP | EX | IS | SY => {
                        *pos = *next_idx;
                        continue;
                    }
                    _ => {}
                }
            }

            match class {
                /* LB13 Do not break before ‘]’ or ‘!’ or ‘;’ or ‘/’, even after spaces. */
                SP if [CL, CP, EX, IS, SY].contains(&get_class!(text[idx..].trim_start())) => {
                    *pos += grapheme.len();
                    while ![CL, CP, EX, IS, SY]
                        .contains(&next_grapheme_class!(iter, grapheme).unwrap())
                    {
                        *pos += grapheme.len();
                    }
                    *pos += grapheme.len();
                    continue;
                }
                OP => {
                    /* LB14 Do not break after ‘[’, even after spaces.
                     * OP SP* ×
                     */
                    while let Some((idx, grapheme)) = self.iter.next() {
                        *pos = idx + grapheme.len();
                        if !(get_class!(grapheme) == SP) {
                            break;
                        }
                    }
                    continue;
                }
                QU if get_class!(text[idx..].trim_start()) == OP => {
                    /* LB15 Do not break within ‘”[’, even with intervening spaces.
                     * QU SP* × OP */
                    *pos += grapheme.len();
                    while Some(SP) == next_grapheme_class!(iter, grapheme) {
                        *pos += grapheme.len();
                    }
                    *pos = idx;
                    continue;
                }
                QU => {
                    /* LB19 Do not break before or after quotation marks, such as ‘ ” ’. */
                    *pos += grapheme.len();
                    if let Some((_, g)) = self.iter.next() {
                        *pos += g.len();
                    }
                    continue;
                }
                LineBreakClass::CL | LineBreakClass::CP
                    if get_class!(text[idx..].trim_start()) == NS =>
                {
                    /* LB16 Do not break between closing punctuation and a nonstarter (lb=NS), even with
                     * intervening spaces.
                     * (CL | CP) SP* × NS */
                    *pos += grapheme.len();
                    while Some(SP) == next_grapheme_class!(iter, grapheme) {
                        *pos += grapheme.len();
                    }
                    continue;
                }
                B2 if get_class!(text[idx..].trim_start()) == B2 => {
                    *pos += grapheme.len();
                    while Some(SP) == next_grapheme_class!(iter, grapheme) {
                        *pos += grapheme.len();
                    }
                    continue;
                }
                SP => {
                    /* LB18 Break after spaces.  SP ÷ */
                    // Space 0x20 is 1 byte long.
                    *pos += 1;
                    return Some((*pos, BreakAllowed));
                }
                _ => {}
            }
            if let Some((next_idx, next_grapheme)) = next_char {
                let next_class = get_class!(next_grapheme);
                match next_class {
                    QU if class != SP => {
                        /* LB19 Do not break before or after quotation marks, such as ‘ ” ’. */
                        *pos = *next_idx + next_grapheme.len();
                        self.iter.next();
                        continue;
                    }
                    _ => {}
                }
            }
            match class {
                CB => {
                    /* LB20 Break before and after unresolved CB. */
                    *pos += grapheme.len();
                    return Some((*pos - 1, BreakAllowed));
                }
                /* LB21 Do not break before hyphen-minus, other hyphens, fixed-width spaces, small
                 * kana, and other non-starters, or after acute accents.  × BA,  × HY, × NS,  BB × */
                BB => {
                    *pos += grapheme.len();
                    continue;
                }
                _ => {}
            }

            if let Some((_, next_grapheme)) = next_char {
                let next_class = get_class!(next_grapheme);
                match next_class {
                    BA | HY | NS => {
                        /* LB21 Do not break before hyphen-minus, other hyphens, fixed-width spaces, small
                         * kana, and other non-starters, or after acute accents.  × BA,  × HY, × NS,  BB × */
                        *pos += grapheme.len();
                        continue;
                    }
                    _ => {}
                }
            }
            match class {
                HL if next_grapheme_class!((next_char is HY, BA)) => {
                    /* LB21a Don’t break after Hebrew + Hyphen.  HL (HY | BA) × */
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* LB21b Don’t break between ,Solidus and Hebrew letters.  SY × HL */
                SY if next_grapheme_class!((next_char is HL)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    /* bypass next_char */
                    self.iter.next().unwrap();
                    if let Some((idx, next_grapheme)) = self.iter.next() {
                        *pos = idx + next_grapheme.len();
                    }
                    continue;
                }
                /*  LB22 Do not break between two ellipses, or between letters, numbers or excla-
                 *  mations and ellipsis.
                 *  Examples: ‘9...’, ‘a...’, ‘H...’
                 *  (AL | HL) × IN */
                AL | HL if next_grapheme_class!((next_char is IN)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /*  EX × IN */
                EX if next_grapheme_class!((next_char is IN)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                EX => {
                    // LB13
                    *pos += grapheme.len();
                    continue;
                }
                /*  (ID | EB | EM) × IN */
                ID | EB | EM if next_grapheme_class!((next_char is IN)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /*  IN × IN */
                IN if next_grapheme_class!((next_char is IN)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /*  NU × IN */
                NU if next_grapheme_class!((next_char is IN)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* LB23 Do not break between digits and letters.
                 * (AL | HL) × NU */
                AL | HL if next_grapheme_class!((next_char is NU)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* NU × (AL | HL) */
                NU if next_grapheme_class!((next_char is AL, HL)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* LB23a Do not break between numeric prefixes and ideographs, or between ideographs
                 * and numeric postfixes.
                 * PR × (ID | EB | EM) */
                PR if next_grapheme_class!((next_char is ID, EB, EM)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* (ID | EB | EM) × PO */
                ID | EB | EM if next_grapheme_class!((next_char is PO)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* B24 Do not break between numeric prefix/postfix and letters, or between
                letters and prefix/postfix.
                (PR | PO) × (AL | HL)*/
                PR | PO if next_grapheme_class!((next_char is AL, HL)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /*(AL | HL) × (PR | PO) */
                AL | HL if next_grapheme_class!((next_char is PR, PO)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* LB25 Do not break between the following pairs of classes relevant to numbers:
                 * CL × PO */
                CL if next_grapheme_class!((next_char is PO)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* CP × PO */
                CP if next_grapheme_class!((next_char is PO)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* CL × PR */
                CL if next_grapheme_class!((next_char is PR)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* CP × PR */
                CP if next_grapheme_class!((next_char is PR)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* NU × PO */
                NU if next_grapheme_class!((next_char is PO)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* NU × PR */
                NU if next_grapheme_class!((next_char is PR)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* PO × OP */
                PO if next_grapheme_class!((next_char is OP)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* PO × NU */
                PO if next_grapheme_class!((next_char is NU)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* PR × OP */
                PR if next_grapheme_class!((next_char is OP)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* PR × NU */
                PR if next_grapheme_class!((next_char is NU)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* HY × NU */
                HY if next_grapheme_class!((next_char is NU)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* IS × NU */
                IS if next_grapheme_class!((next_char is NU)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* NU × NU */
                NU if next_grapheme_class!((next_char is NU)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* SY × NU */
                SY if next_grapheme_class!((next_char is NU)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* LB26 Do not break a Korean syllable.
                 * JL × (JL | JV | H2 | H3) */
                JL if next_grapheme_class!((next_char is JL, JV, H2, H3)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* (JV | H2) × (JV | JT) */
                JV | H2 if next_grapheme_class!((next_char is JV, JT)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* (JT | H3) × JT */
                JT | H3 if next_grapheme_class!((next_char is JT)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* LB27 Treat a Korean Syllable Block the same as ID.
                 * (JL | JV | JT | H2 | H3) × IN */
                JL | JV | JT | H2 | H3 if next_grapheme_class!((next_char is IN)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* (JL | JV | JT | H2 | H3) × PO */
                JL | JV | JT | H2 | H3 if next_grapheme_class!((next_char is PO)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* PR × (JL | JV | JT | H2 | H3) */
                PR if next_grapheme_class!((next_char is JL, JV, JT, H2, H3)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* LB28 Do not break between alphabetics (“at”).
                (AL | HL) × (AL | HL) */
                AL | HL if next_grapheme_class!((next_char is AL, HL)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* LB29 Do not break between numeric punctuation and alphabetics (“e.g.”).
                IS × (AL | HL) */
                IS if next_grapheme_class!((next_char is AL, HL)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* LB30 Do not break between letters, numbers, or ordinary symbols and opening
                or closing parentheses.
                (AL | HL | NU) × OP */
                AL | HL | NU if next_grapheme_class!((next_char is OP)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /* CP × (AL | HL | NU) */
                CP if next_grapheme_class!((next_char is AL, HL , NU)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                /*LB30b Do not break between an emoji base and an emoji modifier.
                 * EB × EM */
                EB if next_grapheme_class!((next_char is EM)) => {
                    let (idx, next_grapheme) = next_char.unwrap();
                    *pos = idx + next_grapheme.len();
                    self.iter.next();
                    continue;
                }
                RI => {
                    /* LB30a Break between two regional indicator symbols if and only if there are an
                     * even number of regional indicators preceding the position of the break.
                     * sot (RI RI)* RI × RI
                     * [^RI] (RI RI)* RI × RI */
                    *reg_ind_streak += 1;
                    *pos += grapheme.len();
                    if *reg_ind_streak % 2 == 1 {
                        return Some((*pos - grapheme.len(), BreakAllowed));
                    }
                    self.iter.next();
                    continue;
                }
                _ if next_char.is_none() => {
                    return None;
                }
                _ => {
                    *pos += grapheme.len();
                    return Some((*pos - grapheme.len(), BreakAllowed));
                }
            }
        }
    }
}

fn search_table(c: u32, t: &'static [(u32, u32, LineBreakClass)]) -> LineBreakClass {
    match t.binary_search_by(|&(lo, hi, _)| {
        if lo <= c && c <= hi {
            Ordering::Equal
        } else if hi < c {
            Ordering::Less
        } else {
            Ordering::Greater
        }
    }) {
        Ok(idx) => t[idx].2,
        Err(_) => XX,
    }
}

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

    #[test]
    fn test_line_breaks() {
        let s = "Fell past it.\n\n‘Well!’ thought Alice to herself.";
        let breaks = LineBreakCandidateIter::new(s).collect::<Vec<(usize, LineBreakCandidate)>>();
        let mut prev = 0;
        for b in breaks {
            println!("{:?}", &s[prev..b.0]);
            prev = b.0;
        }
        println!("{:?}", &s[prev..]);

        let s = r#"Τ' άστρα τα κοντά -στη γλυκιά σελήνη
την ειδή των κρύβουν - τη διαμαντένια,
άμα φως λαμπρό -στη γή πάσα χύνει,
όλη ασημένια."#;
        let breaks = LineBreakCandidateIter::new(s).collect::<Vec<(usize, LineBreakCandidate)>>();
        let mut prev = 0;
        for b in breaks {
            println!("{:?}", &s[prev..b.0]);
            prev = b.0;
        }
        println!("{:?}", &s[prev..]);
    }
}

pub use alg::linear;

mod alg {
    use super::super::grapheme_clusters::TextProcessing;
    use super::super::*;
    fn cost(i: usize, j: usize, width: usize, minima: &[usize], offsets: &[usize]) -> usize {
        let w = offsets[j] + j - offsets[i] - i - 1;
        if w > width {
            return 65536 * (w - width);
        }
        minima[i] + (width - w) * (width - w)
    }

    fn smawk(
        rows: &mut Vec<usize>,
        columns: &mut Vec<usize>,
        minima: &mut Vec<usize>,
        breaks: &mut Vec<usize>,
        width: usize,
        offsets: &Vec<usize>,
    ) {
        let mut stack = Vec::new();
        let mut i = 0;
        while i < rows.len() {
            if !stack.is_empty() {
                let c = columns[stack.len() - 1];
                if cost(*stack.iter().last().unwrap(), c, width, minima, offsets)
                    < cost(rows[i], c, width, minima, offsets)
                {
                    if stack.len() < columns.len() {
                        stack.push(rows[i]);
                    }
                    i += 1;
                } else {
                    stack.pop();
                }
            } else {
                stack.push(rows[i]);
                i += 1;
            }
        }
        let rows = &mut stack;
        if columns.len() > 1 {
            let mut odd_columns = columns.iter().skip(1).step_by(2).cloned().collect();
            smawk(rows, &mut odd_columns, minima, breaks, width, offsets);
            for (i, o) in odd_columns.into_iter().enumerate() {
                columns[2 * i + 1] = o;
            }
        }
        let mut i = 0;
        let mut j = 0;
        while j < columns.len() {
            let end = if j + 1 < columns.len() {
                breaks[columns[j + 1]]
            } else {
                *rows.iter().last().unwrap()
            };
            let c = cost(rows[i], columns[j], width, minima, offsets);
            if c < minima[columns[j]] {
                minima[columns[j]] = c;
                breaks[columns[j]] = rows[i];
            }
            if rows[i] < end {
                i += 1;
            } else {
                j += 2;
            }
        }
    }

    pub fn linear(text: &str, width: usize) -> Vec<String> {
        let mut words = Vec::new();
        let breaks =
            LineBreakCandidateIter::new(text).collect::<Vec<(usize, LineBreakCandidate)>>();
        {
            let mut prev = 0;
            for b in breaks {
                if text[prev..b.0].ends_with('\n') && text[b.0..].starts_with('\n') {
                    words.push(text[prev..b.0].trim_end_matches('\n'));
                    words.push("\n\n");
                } else if &text[prev..b.0] != "\n" {
                    words.push(text[prev..b.0].trim_end_matches('\n'));
                    if text[prev..b.0].ends_with('\n') {
                        words.push(" ");
                    }
                }
                prev = b.0;
            }
            if &text[prev..] != "\n" {
                words.push(text[prev..].trim_end_matches("\n"));
            }
        }
        let count = words.len();
        let mut minima = vec![std::usize::MAX - 1; count + 1];
        minima[0] = 0;
        let mut offsets = Vec::with_capacity(words.len());
        offsets.push(0);
        for w in words.iter() {
            if *w == "\n\n" {
                offsets.push(offsets.iter().last().unwrap() + width - 1);
            } else {
                offsets.push(offsets.iter().last().unwrap() + w.grapheme_len().saturating_sub(1));
            }
        }

        let mut breaks = vec![0; count + 1];

        let mut n = count + 1;
        let mut i = 1;
        let mut offset = 0;
        loop {
            let r = std::cmp::min(n, 2 * i);
            let edge = i + offset;
            smawk(
                &mut (offset..edge).collect(),
                &mut (edge..(r + offset)).collect(),
                &mut minima,
                &mut breaks,
                width,
                &offsets,
            );
            let x = minima[r - 1 + offset];
            let mut for_was_broken = false;
            for j in i..(r - 1) {
                let y = cost(j + offset, r - 1 + offset, width, &minima, &offsets);
                if y <= x {
                    n -= j;
                    i = 1;
                    offset += j;
                    for_was_broken = true;
                    break;
                }
            }

            if !for_was_broken {
                if r == n {
                    break;
                }
                i *= 2;
            }
        }
        let paragraphs = text.split("\n\n").count();
        let mut lines = Vec::new();
        let mut j = count;
        let mut p_i = 0;
        while j > 0 {
            let mut line = String::new();
            for i in breaks[j]..j {
                line.push_str(words[i]);
            }
            lines.push(line);
            if p_i + 1 < paragraphs {
                lines.push(String::new());
                p_i += 1;
            }
            j = breaks[j];
        }
        lines.reverse();
        lines
    }
}

pub fn split_lines_reflow(text: &str, reflow: Reflow, width: Option<usize>) -> Vec<String> {
    match reflow {
        Reflow::FormatFlowed => {
            /* rfc3676 - The Text/Plain Format and DelSp Parameters
             * https://tools.ietf.org/html/rfc3676 */

            let mut ret = Vec::new();
            /*
             * - Split lines with indices using str::match_indices()
             * - Iterate and reflow flow regions, and pass fixed regions through
             */
            let lines_indices: Vec<usize> = text.match_indices('\n').map(|(i, _)| i).collect();
            let mut prev_index = 0;
            let mut in_paragraph = false;
            let mut paragraph_start = 0;

            let mut prev_quote_depth = 0;
            for i in &lines_indices {
                let line = &text[prev_index..*i];
                let mut trimmed = line.trim_start().lines().next().unwrap_or("");
                let mut quote_depth = 0;
                let p_str: usize = trimmed
                    .as_bytes()
                    .iter()
                    .position(|&b| {
                        if b != b'>' {
                            /* position() is short-circuiting */
                            true
                        } else {
                            quote_depth += 1;
                            false
                        }
                    })
                    .unwrap_or(0);
                trimmed = &trimmed[p_str..];
                if trimmed.starts_with(' ') {
                    /* Remove space stuffing before checking for ending space character.
                     * [rfc3676#section-4.4] */
                    trimmed = &trimmed[1..];
                }

                if trimmed.ends_with(' ') {
                    if !in_paragraph {
                        in_paragraph = true;
                        paragraph_start = prev_index;
                    } else if prev_quote_depth == quote_depth {
                        /* This becomes part of the paragraph we're in */
                    } else {
                        /*Malformed line, different quote depths can't be in the same paragraph. */
                        let paragraph = &text[paragraph_start..prev_index];
                        reflow_helper(&mut ret, paragraph, prev_quote_depth, in_paragraph, width);

                        paragraph_start = prev_index;
                    }
                } else {
                    if prev_quote_depth == quote_depth || !in_paragraph {
                        let paragraph = &text[paragraph_start..*i];
                        reflow_helper(&mut ret, paragraph, quote_depth, in_paragraph, width);
                    } else {
                        /*Malformed line, different quote depths can't be in the same paragraph. */
                        let paragraph = &text[paragraph_start..prev_index];
                        reflow_helper(&mut ret, paragraph, prev_quote_depth, in_paragraph, width);
                        let paragraph = &text[prev_index..*i];
                        reflow_helper(&mut ret, paragraph, quote_depth, false, width);
                    }
                    paragraph_start = *i;
                    in_paragraph = false;
                }
                prev_quote_depth = quote_depth;
                prev_index = *i;
            }
            let paragraph = &text[paragraph_start..text.len()];
            reflow_helper(&mut ret, paragraph, prev_quote_depth, in_paragraph, width);
            ret
        }
        Reflow::All => {
            if let Some(width) = width {
                let mut ret = Vec::new();
                let width = width.saturating_sub(2);

                for line in text.lines() {
                    if line.grapheme_len() <= width {
                        ret.push(line.to_string());
                        continue;
                    }

                    let breaks = LineBreakCandidateIter::new(line)
                        .collect::<Vec<(usize, LineBreakCandidate)>>();
                    if breaks.len() < 2 {
                        split(&mut ret, line, width);
                        continue;
                    }

                    let mut prev = 0;
                    let mut prev_line_offset = 0;
                    while prev < breaks.len() {
                        let new_off = match breaks[prev..].binary_search_by(|(offset, _)| {
                            line[prev_line_offset..*offset].grapheme_len().cmp(&width)
                        }) {
                            Ok(v) => v,
                            Err(v) => v,
                        } + prev;
                        let end_offset = if new_off >= breaks.len() {
                            line.len()
                        } else {
                            breaks[new_off].0
                        };
                        if !line[prev_line_offset..end_offset].is_empty() {
                            if prev_line_offset == 0 {
                                ret.push(line[prev_line_offset..end_offset].to_string());
                            } else {
                                ret.push(format!("⤷{}", &line[prev_line_offset..end_offset]));
                            }
                        }
                        if prev_line_offset == end_offset && prev == new_off {
                            break;
                        }
                        prev_line_offset = end_offset;
                        prev = new_off;
                    }
                }
                ret
            } else {
                text.trim().split('\n').map(str::to_string).collect()
            }
        }
        Reflow::No => text.trim().split('\n').map(str::to_string).collect(),
    }
}

fn split(ret: &mut Vec<String>, mut line: &str, width: usize) {
    while !line.is_empty() {
        let mut chop_index = std::cmp::min(line.len().saturating_sub(1), width);
        while chop_index > 0 && !line.is_char_boundary(chop_index) {
            chop_index -= 1;
        }
        if chop_index == 0 {
            ret.push(format!("⤷{}", line));
            return;
        } else {
            ret.push(format!("⤷{}", &line[..chop_index]));
        }
        line = &line[chop_index..];
    }
}
fn reflow_helper(
    ret: &mut Vec<String>,
    paragraph: &str,
    quote_depth: usize,
    in_paragraph: bool,
    width: Option<usize>,
) {
    if quote_depth > 0 {
        let quotes: String = ">".repeat(quote_depth);
        let paragraph = paragraph
            .trim_start_matches(&quotes)
            .replace(&format!("\n{}", &quotes), "")
            .replace("\n", "")
            .replace("\r", "");
        if in_paragraph {
            if let Some(width) = width {
                ret.extend(
                    linear(&paragraph, width.saturating_sub(quote_depth))
                        .into_iter()
                        .map(|l| format!("{}{}", &quotes, l)),
                );
            } else {
                ret.push(format!("{}{}", &quotes, &paragraph));
            }
        } else {
            ret.push(format!("{}{}", &quotes, &paragraph));
        }
    } else {
        let paragraph = paragraph.replace("\n", "").replace("\r", "");

        if in_paragraph {
            if let Some(width) = width {
                let ex = linear(&paragraph, width);
                ret.extend(ex.into_iter());
            } else {
                ret.push(paragraph);
            }
        } else {
            ret.push(paragraph);
        }
    }
}

#[test]
fn test_reflow() {
    let text = r#"`Take some more tea,' the March Hare said to Alice, very 
earnestly.

`I've had nothing yet,' Alice replied in an offended tone, `so 
I can't take more.'

`You mean you can't take LESS,' said the Hatter: `it's very 
easy to take MORE than nothing.'"#;
    for l in split_lines_reflow(text, Reflow::FormatFlowed, Some(30)) {
        println!("{}", l);
    }
    println!();
    for l in split_lines_reflow(text, Reflow::No, Some(30)) {
        println!("{}", l);
    }
    println!();
    let text = r#">>>Take some more tea.
>>I've had nothing yet, so I can't take more.
>You mean you can't take LESS, it's very easy to take 
>MORE than nothing."#;
    for l in split_lines_reflow(text, Reflow::FormatFlowed, Some(20)) {
        println!("{}", l);
    }
    println!();
    for l in split_lines_reflow(text, Reflow::No, Some(20)) {
        println!("{}", l);
    }
    println!();
    use super::_ALICE_CHAPTER_1;
    for l in split_lines_reflow(_ALICE_CHAPTER_1, Reflow::FormatFlowed, Some(72)) {
        println!("{}", l);
    }
}