| 1 | // pest. The Elegant Parser |
| 2 | // Copyright (c) 2018 Dragoș Tiselice |
| 3 | // |
| 4 | // Licensed under the Apache License, Version 2.0 |
| 5 | // <LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0> or the MIT |
| 6 | // license <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your |
| 7 | // option. All files in the project carrying such notice may not be copied, |
| 8 | // modified, or distributed except according to those terms. |
| 9 | |
| 10 | use alloc::rc::Rc; |
| 11 | use alloc::vec::Vec; |
| 12 | use core::fmt; |
| 13 | |
| 14 | use super::line_index::LineIndex; |
| 15 | use super::pair::{self, Pair}; |
| 16 | use super::queueable_token::QueueableToken; |
| 17 | use super::tokens::{self, Tokens}; |
| 18 | use crate::RuleType; |
| 19 | |
| 20 | /// An iterator over [`Pair`]s. It is created by [`Pairs::flatten`]. |
| 21 | /// |
| 22 | /// [`Pair`]: struct.Pair.html |
| 23 | /// [`Pairs::flatten`]: struct.Pairs.html#method.flatten |
| 24 | pub struct FlatPairs<'i, R> { |
| 25 | queue: Rc<Vec<QueueableToken<'i, R>>>, |
| 26 | input: &'i str, |
| 27 | start: usize, |
| 28 | end: usize, |
| 29 | line_index: Rc<LineIndex>, |
| 30 | } |
| 31 | |
| 32 | pub fn new<'i, R: RuleType>( |
| 33 | queue: Rc<Vec<QueueableToken<'i, R>>>, |
| 34 | input: &'i str, |
| 35 | line_index: Rc<LineIndex>, |
| 36 | start: usize, |
| 37 | end: usize, |
| 38 | ) -> FlatPairs<'i, R> { |
| 39 | FlatPairs { |
| 40 | queue, |
| 41 | input, |
| 42 | line_index, |
| 43 | start, |
| 44 | end, |
| 45 | } |
| 46 | } |
| 47 | |
| 48 | impl<'i, R: RuleType> FlatPairs<'i, R> { |
| 49 | /// Returns the `Tokens` for these pairs. |
| 50 | /// |
| 51 | /// # Examples |
| 52 | /// |
| 53 | /// ``` |
| 54 | /// # use std::rc::Rc; |
| 55 | /// # use pest; |
| 56 | /// # #[allow (non_camel_case_types)] |
| 57 | /// # #[derive(Clone, Copy, Debug, Eq, Hash, Ord, PartialEq, PartialOrd)] |
| 58 | /// enum Rule { |
| 59 | /// a |
| 60 | /// } |
| 61 | /// |
| 62 | /// let input = "" ; |
| 63 | /// let pairs = pest::state(input, |state| { |
| 64 | /// // generating Token pair with Rule::a ... |
| 65 | /// # state.rule(Rule::a, |s| Ok(s)) |
| 66 | /// }).unwrap(); |
| 67 | /// let tokens: Vec<_> = pairs.flatten().tokens().collect(); |
| 68 | /// |
| 69 | /// assert_eq!(tokens.len(), 2); |
| 70 | /// ``` |
| 71 | #[inline ] |
| 72 | pub fn tokens(self) -> Tokens<'i, R> { |
| 73 | tokens::new(self.queue, self.input, self.start, self.end) |
| 74 | } |
| 75 | |
| 76 | fn next_start(&mut self) { |
| 77 | self.start += 1; |
| 78 | |
| 79 | while self.start < self.end && !self.is_start(self.start) { |
| 80 | self.start += 1; |
| 81 | } |
| 82 | } |
| 83 | |
| 84 | fn next_start_from_end(&mut self) { |
| 85 | self.end -= 1; |
| 86 | |
| 87 | while self.end >= self.start && !self.is_start(self.end) { |
| 88 | self.end -= 1; |
| 89 | } |
| 90 | } |
| 91 | |
| 92 | fn is_start(&self, index: usize) -> bool { |
| 93 | match self.queue[index] { |
| 94 | QueueableToken::Start { .. } => true, |
| 95 | QueueableToken::End { .. } => false, |
| 96 | } |
| 97 | } |
| 98 | } |
| 99 | |
| 100 | impl<'i, R: RuleType> ExactSizeIterator for FlatPairs<'i, R> { |
| 101 | fn len(&self) -> usize { |
| 102 | // Tokens len is exactly twice as flatten pairs len |
| 103 | (self.end - self.start) >> 1 |
| 104 | } |
| 105 | } |
| 106 | |
| 107 | impl<'i, R: RuleType> Iterator for FlatPairs<'i, R> { |
| 108 | type Item = Pair<'i, R>; |
| 109 | |
| 110 | fn next(&mut self) -> Option<Self::Item> { |
| 111 | if self.start >= self.end { |
| 112 | return None; |
| 113 | } |
| 114 | |
| 115 | let pair: Pair<'_, R> = pair::new( |
| 116 | queue:Rc::clone(&self.queue), |
| 117 | self.input, |
| 118 | line_index:Rc::clone(&self.line_index), |
| 119 | self.start, |
| 120 | ); |
| 121 | self.next_start(); |
| 122 | |
| 123 | Some(pair) |
| 124 | } |
| 125 | |
| 126 | fn size_hint(&self) -> (usize, Option<usize>) { |
| 127 | let len: usize = <Self as ExactSizeIterator>::len(self); |
| 128 | (len, Some(len)) |
| 129 | } |
| 130 | } |
| 131 | |
| 132 | impl<'i, R: RuleType> DoubleEndedIterator for FlatPairs<'i, R> { |
| 133 | fn next_back(&mut self) -> Option<Self::Item> { |
| 134 | if self.end <= self.start { |
| 135 | return None; |
| 136 | } |
| 137 | |
| 138 | self.next_start_from_end(); |
| 139 | |
| 140 | let pair: Pair<'_, R> = pair::new( |
| 141 | queue:Rc::clone(&self.queue), |
| 142 | self.input, |
| 143 | line_index:Rc::clone(&self.line_index), |
| 144 | self.end, |
| 145 | ); |
| 146 | |
| 147 | Some(pair) |
| 148 | } |
| 149 | } |
| 150 | |
| 151 | impl<'i, R: RuleType> fmt::Debug for FlatPairs<'i, R> { |
| 152 | fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { |
| 153 | f&mut DebugStruct<'_, '_>.debug_struct("FlatPairs" ) |
| 154 | .field(name:"pairs" , &self.clone().collect::<Vec<_>>()) |
| 155 | .finish() |
| 156 | } |
| 157 | } |
| 158 | |
| 159 | impl<'i, R: Clone> Clone for FlatPairs<'i, R> { |
| 160 | fn clone(&self) -> FlatPairs<'i, R> { |
| 161 | FlatPairs { |
| 162 | queue: Rc::clone(&self.queue), |
| 163 | input: self.input, |
| 164 | line_index: Rc::clone(&self.line_index), |
| 165 | start: self.start, |
| 166 | end: self.end, |
| 167 | } |
| 168 | } |
| 169 | } |
| 170 | |
| 171 | #[cfg (test)] |
| 172 | mod tests { |
| 173 | use super::super::super::macros::tests::*; |
| 174 | use super::super::super::Parser; |
| 175 | use alloc::vec; |
| 176 | use alloc::vec::Vec; |
| 177 | |
| 178 | #[test ] |
| 179 | fn iter_for_flat_pairs() { |
| 180 | let pairs = AbcParser::parse(Rule::a, "abcde" ).unwrap(); |
| 181 | |
| 182 | assert_eq!( |
| 183 | pairs.flatten().map(|p| p.as_rule()).collect::<Vec<Rule>>(), |
| 184 | vec![Rule::a, Rule::b, Rule::c] |
| 185 | ); |
| 186 | } |
| 187 | |
| 188 | #[test ] |
| 189 | fn double_ended_iter_for_flat_pairs() { |
| 190 | let pairs = AbcParser::parse(Rule::a, "abcde" ).unwrap(); |
| 191 | assert_eq!( |
| 192 | pairs |
| 193 | .flatten() |
| 194 | .rev() |
| 195 | .map(|p| p.as_rule()) |
| 196 | .collect::<Vec<Rule>>(), |
| 197 | vec![Rule::c, Rule::b, Rule::a] |
| 198 | ); |
| 199 | } |
| 200 | |
| 201 | #[test ] |
| 202 | fn test_line_col() { |
| 203 | let mut pairs = AbcParser::parse(Rule::a, "abcNe \nabcde" ).unwrap().flatten(); |
| 204 | |
| 205 | let pair = pairs.next().unwrap(); |
| 206 | assert_eq!(pair.as_str(), "abc" ); |
| 207 | assert_eq!(pair.line_col(), (1, 1)); |
| 208 | assert_eq!(pair.line_col(), pair.as_span().start_pos().line_col()); |
| 209 | |
| 210 | let pair = pairs.next().unwrap(); |
| 211 | assert_eq!(pair.as_str(), "b" ); |
| 212 | assert_eq!(pair.line_col(), (1, 2)); |
| 213 | assert_eq!(pair.line_col(), pair.as_span().start_pos().line_col()); |
| 214 | |
| 215 | let pair = pairs.next().unwrap(); |
| 216 | assert_eq!(pair.as_str(), "e" ); |
| 217 | assert_eq!(pair.line_col(), (1, 5)); |
| 218 | assert_eq!(pair.line_col(), pair.as_span().start_pos().line_col()); |
| 219 | } |
| 220 | |
| 221 | #[test ] |
| 222 | fn exact_size_iter_for_pairs() { |
| 223 | let pairs = AbcParser::parse(Rule::a, "abc \nefgh" ).unwrap().flatten(); |
| 224 | assert_eq!(pairs.len(), pairs.count()); |
| 225 | |
| 226 | let pairs = AbcParser::parse(Rule::a, "我很漂亮efgh" ).unwrap().flatten(); |
| 227 | assert_eq!(pairs.len(), pairs.count()); |
| 228 | |
| 229 | let pairs = AbcParser::parse(Rule::a, "abc \nefgh" ).unwrap().flatten(); |
| 230 | let pairs = pairs.rev(); |
| 231 | assert_eq!(pairs.len(), pairs.count()); |
| 232 | |
| 233 | let mut pairs = AbcParser::parse(Rule::a, "abc \nefgh" ).unwrap().flatten(); |
| 234 | let pairs_len = pairs.len(); |
| 235 | let _ = pairs.next().unwrap(); |
| 236 | assert_eq!(pairs.count() + 1, pairs_len); |
| 237 | } |
| 238 | } |
| 239 | |