|
| 1 | +// Copyright 2012-2014 The Rust Project Developers. See the COPYRIGHT |
| 2 | +// file at the top-level directory of this distribution and at |
| 3 | +// http://rust-lang.org/COPYRIGHT. |
| 4 | +// |
| 5 | +// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or |
| 6 | +// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license |
| 7 | +// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your |
| 8 | +// option. This file may not be copied, modified, or distributed |
| 9 | +// except according to those terms. |
| 10 | + |
| 11 | +use core::prelude::*; |
| 12 | + |
| 13 | +use core::cmp; |
| 14 | + |
| 15 | +use tables::grapheme::GraphemeCat; |
| 16 | + |
| 17 | +/// External iterator for grapheme clusters and byte offsets. |
| 18 | +#[derive(Clone)] |
| 19 | +pub struct GraphemeIndices<'a> { |
| 20 | + start_offset: usize, |
| 21 | + iter: Graphemes<'a>, |
| 22 | +} |
| 23 | + |
| 24 | +impl<'a> Iterator for GraphemeIndices<'a> { |
| 25 | + type Item = (usize, &'a str); |
| 26 | + |
| 27 | + #[inline] |
| 28 | + fn next(&mut self) -> Option<(usize, &'a str)> { |
| 29 | + self.iter.next().map(|s| (s.as_ptr() as usize - self.start_offset, s)) |
| 30 | + } |
| 31 | + |
| 32 | + #[inline] |
| 33 | + fn size_hint(&self) -> (usize, Option<usize>) { |
| 34 | + self.iter.size_hint() |
| 35 | + } |
| 36 | +} |
| 37 | + |
| 38 | +impl<'a> DoubleEndedIterator for GraphemeIndices<'a> { |
| 39 | + #[inline] |
| 40 | + fn next_back(&mut self) -> Option<(usize, &'a str)> { |
| 41 | + self.iter.next_back().map(|s| (s.as_ptr() as usize - self.start_offset, s)) |
| 42 | + } |
| 43 | +} |
| 44 | + |
| 45 | +/// External iterator for a string's |
| 46 | +/// [grapheme clusters](http://www.unicode.org/reports/tr29/#Grapheme_Cluster_Boundaries). |
| 47 | +#[derive(Clone)] |
| 48 | +pub struct Graphemes<'a> { |
| 49 | + string: &'a str, |
| 50 | + extended: bool, |
| 51 | + cat: Option<GraphemeCat>, |
| 52 | + catb: Option<GraphemeCat>, |
| 53 | +} |
| 54 | + |
| 55 | +// state machine for cluster boundary rules |
| 56 | +#[derive(PartialEq,Eq)] |
| 57 | +enum GraphemeState { |
| 58 | + Start, |
| 59 | + FindExtend, |
| 60 | + HangulL, |
| 61 | + HangulLV, |
| 62 | + HangulLVT, |
| 63 | + Regional, |
| 64 | +} |
| 65 | + |
| 66 | +impl<'a> Iterator for Graphemes<'a> { |
| 67 | + type Item = &'a str; |
| 68 | + |
| 69 | + #[inline] |
| 70 | + fn size_hint(&self) -> (usize, Option<usize>) { |
| 71 | + let slen = self.string.len(); |
| 72 | + (cmp::min(slen, 1), Some(slen)) |
| 73 | + } |
| 74 | + |
| 75 | + #[inline] |
| 76 | + fn next(&mut self) -> Option<&'a str> { |
| 77 | + use self::GraphemeState::*; |
| 78 | + use tables::grapheme as gr; |
| 79 | + if self.string.len() == 0 { |
| 80 | + return None; |
| 81 | + } |
| 82 | + |
| 83 | + let mut take_curr = true; |
| 84 | + let mut idx = 0; |
| 85 | + let mut state = Start; |
| 86 | + let mut cat = gr::GC_Any; |
| 87 | + for (curr, ch) in self.string.char_indices() { |
| 88 | + idx = curr; |
| 89 | + |
| 90 | + // retrieve cached category, if any |
| 91 | + // We do this because most of the time we would end up |
| 92 | + // looking up each character twice. |
| 93 | + cat = match self.cat { |
| 94 | + None => gr::grapheme_category(ch), |
| 95 | + _ => self.cat.take().unwrap() |
| 96 | + }; |
| 97 | + |
| 98 | + if match cat { |
| 99 | + gr::GC_Extend => true, |
| 100 | + gr::GC_SpacingMark if self.extended => true, |
| 101 | + _ => false |
| 102 | + } { |
| 103 | + state = FindExtend; // rule GB9/GB9a |
| 104 | + continue; |
| 105 | + } |
| 106 | + |
| 107 | + state = match state { |
| 108 | + Start if '\r' == ch => { |
| 109 | + let slen = self.string.len(); |
| 110 | + let nidx = idx + 1; |
| 111 | + if nidx != slen && self.string.char_at(nidx) == '\n' { |
| 112 | + idx = nidx; // rule GB3 |
| 113 | + } |
| 114 | + break; // rule GB4 |
| 115 | + } |
| 116 | + Start => match cat { |
| 117 | + gr::GC_Control => break, |
| 118 | + gr::GC_L => HangulL, |
| 119 | + gr::GC_LV | gr::GC_V => HangulLV, |
| 120 | + gr::GC_LVT | gr::GC_T => HangulLVT, |
| 121 | + gr::GC_Regional_Indicator => Regional, |
| 122 | + _ => FindExtend |
| 123 | + }, |
| 124 | + FindExtend => { // found non-extending when looking for extending |
| 125 | + take_curr = false; |
| 126 | + break; |
| 127 | + }, |
| 128 | + HangulL => match cat { // rule GB6: L x (L|V|LV|LVT) |
| 129 | + gr::GC_L => continue, |
| 130 | + gr::GC_LV | gr::GC_V => HangulLV, |
| 131 | + gr::GC_LVT => HangulLVT, |
| 132 | + _ => { |
| 133 | + take_curr = false; |
| 134 | + break; |
| 135 | + } |
| 136 | + }, |
| 137 | + HangulLV => match cat { // rule GB7: (LV|V) x (V|T) |
| 138 | + gr::GC_V => continue, |
| 139 | + gr::GC_T => HangulLVT, |
| 140 | + _ => { |
| 141 | + take_curr = false; |
| 142 | + break; |
| 143 | + } |
| 144 | + }, |
| 145 | + HangulLVT => match cat { // rule GB8: (LVT|T) x T |
| 146 | + gr::GC_T => continue, |
| 147 | + _ => { |
| 148 | + take_curr = false; |
| 149 | + break; |
| 150 | + } |
| 151 | + }, |
| 152 | + Regional => match cat { // rule GB8a |
| 153 | + gr::GC_Regional_Indicator => continue, |
| 154 | + _ => { |
| 155 | + take_curr = false; |
| 156 | + break; |
| 157 | + } |
| 158 | + } |
| 159 | + } |
| 160 | + } |
| 161 | + |
| 162 | + self.cat = if take_curr { |
| 163 | + idx = idx + self.string.char_at(idx).len_utf8(); |
| 164 | + None |
| 165 | + } else { |
| 166 | + Some(cat) |
| 167 | + }; |
| 168 | + |
| 169 | + let retstr = &self.string[..idx]; |
| 170 | + self.string = &self.string[idx..]; |
| 171 | + Some(retstr) |
| 172 | + } |
| 173 | +} |
| 174 | + |
| 175 | +impl<'a> DoubleEndedIterator for Graphemes<'a> { |
| 176 | + #[inline] |
| 177 | + fn next_back(&mut self) -> Option<&'a str> { |
| 178 | + use self::GraphemeState::*; |
| 179 | + use tables::grapheme as gr; |
| 180 | + if self.string.len() == 0 { |
| 181 | + return None; |
| 182 | + } |
| 183 | + |
| 184 | + let mut take_curr = true; |
| 185 | + let mut idx = self.string.len(); |
| 186 | + let mut previdx = idx; |
| 187 | + let mut state = Start; |
| 188 | + let mut cat = gr::GC_Any; |
| 189 | + for (curr, ch) in self.string.char_indices().rev() { |
| 190 | + previdx = idx; |
| 191 | + idx = curr; |
| 192 | + |
| 193 | + // cached category, if any |
| 194 | + cat = match self.catb { |
| 195 | + None => gr::grapheme_category(ch), |
| 196 | + _ => self.catb.take().unwrap() |
| 197 | + }; |
| 198 | + |
| 199 | + // a matching state machine that runs *backwards* across an input string |
| 200 | + // note that this has some implications for the Hangul matching, since |
| 201 | + // we now need to know what the rightward letter is: |
| 202 | + // |
| 203 | + // Right to left, we have: |
| 204 | + // L x L |
| 205 | + // V x (L|V|LV) |
| 206 | + // T x (V|T|LV|LVT) |
| 207 | + // HangulL means the letter to the right is L |
| 208 | + // HangulLV means the letter to the right is V |
| 209 | + // HangulLVT means the letter to the right is T |
| 210 | + state = match state { |
| 211 | + Start if '\n' == ch => { |
| 212 | + if idx > 0 && '\r' == self.string.char_at_reverse(idx) { |
| 213 | + idx -= 1; // rule GB3 |
| 214 | + } |
| 215 | + break; // rule GB4 |
| 216 | + }, |
| 217 | + Start | FindExtend => match cat { |
| 218 | + gr::GC_Extend => FindExtend, |
| 219 | + gr::GC_SpacingMark if self.extended => FindExtend, |
| 220 | + gr::GC_L | gr::GC_LV | gr::GC_LVT => HangulL, |
| 221 | + gr::GC_V => HangulLV, |
| 222 | + gr::GC_T => HangulLVT, |
| 223 | + gr::GC_Regional_Indicator => Regional, |
| 224 | + gr::GC_Control => { |
| 225 | + take_curr = Start == state; |
| 226 | + break; |
| 227 | + }, |
| 228 | + _ => break |
| 229 | + }, |
| 230 | + HangulL => match cat { // char to right is an L |
| 231 | + gr::GC_L => continue, // L x L is the only legal match |
| 232 | + _ => { |
| 233 | + take_curr = false; |
| 234 | + break; |
| 235 | + } |
| 236 | + }, |
| 237 | + HangulLV => match cat { // char to right is a V |
| 238 | + gr::GC_V => continue, // V x V, right char is still V |
| 239 | + gr::GC_L | gr::GC_LV => HangulL, // (L|V) x V, right char is now L |
| 240 | + _ => { |
| 241 | + take_curr = false; |
| 242 | + break; |
| 243 | + } |
| 244 | + }, |
| 245 | + HangulLVT => match cat { // char to right is a T |
| 246 | + gr::GC_T => continue, // T x T, right char is still T |
| 247 | + gr::GC_V => HangulLV, // V x T, right char is now V |
| 248 | + gr::GC_LV | gr::GC_LVT => HangulL, // (LV|LVT) x T, right char is now L |
| 249 | + _ => { |
| 250 | + take_curr = false; |
| 251 | + break; |
| 252 | + } |
| 253 | + }, |
| 254 | + Regional => match cat { // rule GB8a |
| 255 | + gr::GC_Regional_Indicator => continue, |
| 256 | + _ => { |
| 257 | + take_curr = false; |
| 258 | + break; |
| 259 | + } |
| 260 | + } |
| 261 | + } |
| 262 | + } |
| 263 | + |
| 264 | + self.catb = if take_curr { |
| 265 | + None |
| 266 | + } else { |
| 267 | + idx = previdx; |
| 268 | + Some(cat) |
| 269 | + }; |
| 270 | + |
| 271 | + let retstr = &self.string[idx..]; |
| 272 | + self.string = &self.string[..idx]; |
| 273 | + Some(retstr) |
| 274 | + } |
| 275 | +} |
| 276 | + |
| 277 | +#[inline] |
| 278 | +pub fn new_graphemes<'b>(s: &'b str, is_extended: bool) -> Graphemes<'b> { |
| 279 | + Graphemes { string: s, extended: is_extended, cat: None, catb: None } |
| 280 | +} |
| 281 | + |
| 282 | +#[inline] |
| 283 | +pub fn new_grapheme_indices<'b>(s: &'b str, is_extended: bool) -> GraphemeIndices<'b> { |
| 284 | + GraphemeIndices { start_offset: s.as_ptr() as usize, iter: new_graphemes(s, is_extended) } |
| 285 | +} |
0 commit comments