| 1 |
// Copyright Mozilla Foundation. See the COPYRIGHT |
| 2 |
// file at the top-level directory of this distribution. |
| 3 |
// |
| 4 |
// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or |
| 5 |
// https://www.apache.org/licenses/LICENSE-2.0> or the MIT license |
| 6 |
// <LICENSE-MIT or https://opensource.org/licenses/MIT>, at your |
| 7 |
// option. This file may not be copied, modified, or distributed |
| 8 |
// except according to those terms. |
| 9 |
|
| 10 |
use super::*; |
| 11 |
use crate::data::*; |
| 12 |
use crate::handles::*; |
| 13 |
use crate::variant::*; |
| 14 |
// Rust 1.14.0 requires the following despite the asterisk above. |
| 15 |
use super::in_inclusive_range16; |
| 16 |
use super::in_range16; |
| 17 |
|
| 18 |
enum Gb18030Pending { |
| 19 |
None, |
| 20 |
One(u8), |
| 21 |
Two(u8, u8), |
| 22 |
Three(u8, u8, u8), |
| 23 |
} |
| 24 |
|
| 25 |
impl Gb18030Pending { |
| 26 |
fn is_none(&self) -> bool { |
| 27 |
match *self { |
| 28 |
Gb18030Pending::None => true, |
| 29 |
_ => false, |
| 30 |
} |
| 31 |
} |
| 32 |
|
| 33 |
fn count(&self) -> usize { |
| 34 |
match *self { |
| 35 |
Gb18030Pending::None => 0, |
| 36 |
Gb18030Pending::One(_) => 1, |
| 37 |
Gb18030Pending::Two(_, _) => 2, |
| 38 |
Gb18030Pending::Three(_, _, _) => 3, |
| 39 |
} |
| 40 |
} |
| 41 |
} |
| 42 |
|
| 43 |
pub struct Gb18030Decoder { |
| 44 |
first: Option<u8>, |
| 45 |
second: Option<u8>, |
| 46 |
third: Option<u8>, |
| 47 |
pending: Gb18030Pending, |
| 48 |
pending_ascii: Option<u8>, |
| 49 |
} |
| 50 |
|
| 51 |
impl Gb18030Decoder { |
| 52 |
pub fn new() -> VariantDecoder { |
| 53 |
VariantDecoder::Gb18030(Gb18030Decoder { |
| 54 |
first: None, |
| 55 |
second: None, |
| 56 |
third: None, |
| 57 |
pending: Gb18030Pending::None, |
| 58 |
pending_ascii: None, |
| 59 |
}) |
| 60 |
} |
| 61 |
|
| 62 |
pub fn in_neutral_state(&self) -> bool { |
| 63 |
self.first.is_none() |
| 64 |
&& self.second.is_none() |
| 65 |
&& self.third.is_none() |
| 66 |
&& self.pending.is_none() |
| 67 |
&& self.pending_ascii.is_none() |
| 68 |
} |
| 69 |
|
| 70 |
fn extra_from_state(&self, byte_length: usize) -> Option<usize> { |
| 71 |
byte_length.checked_add( |
| 72 |
self.pending.count() |
| 73 |
+ match self.first { |
| 74 |
None => 0, |
| 75 |
Some(_) => 1, |
| 76 |
} |
| 77 |
+ match self.second { |
| 78 |
None => 0, |
| 79 |
Some(_) => 1, |
| 80 |
} |
| 81 |
+ match self.third { |
| 82 |
None => 0, |
| 83 |
Some(_) => 1, |
| 84 |
} |
| 85 |
+ match self.pending_ascii { |
| 86 |
None => 0, |
| 87 |
Some(_) => 1, |
| 88 |
}, |
| 89 |
) |
| 90 |
} |
| 91 |
|
| 92 |
pub fn max_utf16_buffer_length(&self, byte_length: usize) -> Option<usize> { |
| 93 |
// ASCII: 1 to 1 (worst case) |
| 94 |
// gbk: 2 to 1 |
| 95 |
// ranges: 4 to 1 or 4 to 2 |
| 96 |
checked_add(1, self.extra_from_state(byte_length)) |
| 97 |
} |
| 98 |
|
| 99 |
pub fn max_utf8_buffer_length_without_replacement(&self, byte_length: usize) -> Option<usize> { |
| 100 |
// ASCII: 1 to 1 |
| 101 |
// gbk: 2 to 2 or 2 to 3 |
| 102 |
// ranges: 4 to 2, 4 to 3 or 4 to 4 |
| 103 |
// 0x80: 1 to 3 (worst case) |
| 104 |
self.max_utf8_buffer_length(byte_length) |
| 105 |
} |
| 106 |
|
| 107 |
pub fn max_utf8_buffer_length(&self, byte_length: usize) -> Option<usize> { |
| 108 |
checked_add(1, checked_mul(3, self.extra_from_state(byte_length))) |
| 109 |
} |
| 110 |
|
| 111 |
gb18030_decoder_functions!( |
| 112 |
{ |
| 113 |
// If first is between 0x81 and 0xFE, inclusive, |
| 114 |
// subtract offset 0x81. |
| 115 |
let non_ascii_minus_offset = non_ascii.wrapping_sub(0x81); |
| 116 |
if non_ascii_minus_offset > (0xFE - 0x81) { |
| 117 |
if non_ascii == 0x80 { |
| 118 |
handle.write_upper_bmp(0x20ACu16); |
| 119 |
continue 'outermost; |
| 120 |
} |
| 121 |
return (DecoderResult::Malformed(1, 0), |
| 122 |
source.consumed(), |
| 123 |
handle.written()); |
| 124 |
} |
| 125 |
non_ascii_minus_offset |
| 126 |
}, |
| 127 |
{ |
| 128 |
// Two-byte (or error) |
| 129 |
if first_minus_offset >= 0x20 { |
| 130 |
// Not the gbk ideograph range above GB2312 |
| 131 |
let trail_minus_offset = second.wrapping_sub(0xA1); |
| 132 |
if trail_minus_offset <= (0xFE - 0xA1) { |
| 133 |
// GB2312 |
| 134 |
let hanzi_lead = first_minus_offset.wrapping_sub(0x2F); |
| 135 |
if hanzi_lead < (0x77 - 0x2F) { |
| 136 |
// Level 1 Hanzi, Level 2 Hanzi |
| 137 |
// or one of the 5 PUA code |
| 138 |
// points in between. |
| 139 |
let hanzi_pointer = mul_94(hanzi_lead) + trail_minus_offset as usize; |
| 140 |
let upper_bmp = GB2312_HANZI[hanzi_pointer]; |
| 141 |
handle.write_upper_bmp(upper_bmp) |
| 142 |
} else if first_minus_offset == 0x20 { |
| 143 |
// Symbols (starting with ideographic space) |
| 144 |
let bmp = GB2312_SYMBOLS[trail_minus_offset as usize]; |
| 145 |
handle.write_bmp_excl_ascii(bmp) |
| 146 |
} else if first_minus_offset == 0x25 && ((trail_minus_offset.wrapping_sub(63) as usize) < GB2312_SYMBOLS_AFTER_GREEK.len()) { |
| 147 |
handle.write_bmp_excl_ascii(GB2312_SYMBOLS_AFTER_GREEK[trail_minus_offset.wrapping_sub(63) as usize]) |
| 148 |
} else if first_minus_offset == 0x27 && (trail_minus_offset as usize) < GB2312_PINYIN.len() { |
| 149 |
handle.write_bmp_excl_ascii(GB2312_PINYIN[trail_minus_offset as usize]) |
| 150 |
} else if first_minus_offset > 0x76 { |
| 151 |
// Bottom PUA |
| 152 |
let pua = (0xE234 + mul_94(first_minus_offset - 0x77) + trail_minus_offset as usize) as u16; |
| 153 |
handle.write_upper_bmp(pua) |
| 154 |
} else { |
| 155 |
let bmp = gb2312_other_decode((mul_94(first_minus_offset - 0x21) + (trail_minus_offset as usize)) as u16); |
| 156 |
handle.write_bmp_excl_ascii(bmp) |
| 157 |
} |
| 158 |
} else { |
| 159 |
// gbk range on the left |
| 160 |
let mut trail_minus_offset = second.wrapping_sub(0x40); |
| 161 |
if trail_minus_offset > (0x7E - 0x40) { |
| 162 |
let trail_minus_range_start = second.wrapping_sub(0x80); |
| 163 |
if trail_minus_range_start > (0xA0 - 0x80) { |
| 164 |
if second < 0x80 { |
| 165 |
return (DecoderResult::Malformed(1, 0), |
| 166 |
unread_handle_second.unread(), |
| 167 |
handle.written()); |
| 168 |
} |
| 169 |
return (DecoderResult::Malformed(2, 0), |
| 170 |
unread_handle_second.consumed(), |
| 171 |
handle.written()); |
| 172 |
} |
| 173 |
trail_minus_offset = second - 0x41; |
| 174 |
} |
| 175 |
// Zero-base lead |
| 176 |
let left_lead = first_minus_offset - 0x20; |
| 177 |
let left_pointer = left_lead as usize * (190 - 94) + |
| 178 |
trail_minus_offset as usize; |
| 179 |
let gbk_left_ideograph_pointer = left_pointer.wrapping_sub((0x29 - 0x20) * (190 - 94)); |
| 180 |
if gbk_left_ideograph_pointer < (((0x7D - 0x29) * (190 - 94)) - 5) { |
| 181 |
let upper_bmp = gbk_left_ideograph_decode(gbk_left_ideograph_pointer as u16); |
| 182 |
handle.write_upper_bmp(upper_bmp) |
| 183 |
} else if left_pointer < ((0x29 - 0x20) * (190 - 94)) { |
| 184 |
let bmp = gbk_other_decode(left_pointer as u16); |
| 185 |
handle.write_bmp_excl_ascii(bmp) |
| 186 |
} else { |
| 187 |
let bottom_pointer = left_pointer - (((0x7D - 0x20) * (190 - 94)) - 5); |
| 188 |
let upper_bmp = GBK_BOTTOM[bottom_pointer]; |
| 189 |
handle.write_upper_bmp(upper_bmp) |
| 190 |
} |
| 191 |
} |
| 192 |
} else { |
| 193 |
// gbk ideograph range above GB2312 |
| 194 |
let mut trail_minus_offset = second.wrapping_sub(0x40); |
| 195 |
if trail_minus_offset > (0x7E - 0x40) { |
| 196 |
let trail_minus_range_start = second.wrapping_sub(0x80); |
| 197 |
if trail_minus_range_start > (0xFE - 0x80) { |
| 198 |
if second < 0x80 { |
| 199 |
return (DecoderResult::Malformed(1, 0), |
| 200 |
unread_handle_second.unread(), |
| 201 |
handle.written()); |
| 202 |
} |
| 203 |
return (DecoderResult::Malformed(2, 0), |
| 204 |
unread_handle_second.consumed(), |
| 205 |
handle.written()); |
| 206 |
} |
| 207 |
trail_minus_offset = second - 0x41; |
| 208 |
} |
| 209 |
let pointer = first_minus_offset as usize * 190usize + |
| 210 |
trail_minus_offset as usize; |
| 211 |
let upper_bmp = gbk_top_ideograph_decode(pointer as u16); |
| 212 |
handle.write_upper_bmp(upper_bmp) |
| 213 |
} |
| 214 |
}, |
| 215 |
{ |
| 216 |
// If third is between 0x81 and 0xFE, inclusive, |
| 217 |
// subtract offset 0x81. |
| 218 |
let third_minus_offset = third.wrapping_sub(0x81); |
| 219 |
if third_minus_offset > (0xFE - 0x81) { |
| 220 |
// We have an error. Let's inline what's going |
| 221 |
// to happen when `second` is |
| 222 |
// reprocessed. (`third` gets unread.) |
| 223 |
// `second` is guaranteed ASCII, so let's |
| 224 |
// put it in `pending_ascii`. Recompute |
| 225 |
// `second` from `second_minus_offset`. |
| 226 |
self.pending_ascii = Some(second_minus_offset + 0x30); |
| 227 |
// Now unread `third` and designate the previous |
| 228 |
// `first` as being in error. |
| 229 |
return (DecoderResult::Malformed(1, 1), |
| 230 |
unread_handle_third.unread(), |
| 231 |
handle.written()); |
| 232 |
} |
| 233 |
third_minus_offset |
| 234 |
}, |
| 235 |
{ |
| 236 |
// If fourth is between 0x30 and 0x39, inclusive, |
| 237 |
// subtract offset 0x30. |
| 238 |
// |
| 239 |
// If we have an error, we'll inline what's going |
| 240 |
// to happen when `second` and `third` are |
| 241 |
// reprocessed. (`fourth` gets unread.) |
| 242 |
// `second` is guaranteed ASCII, so let's |
| 243 |
// put it in `pending_ascii`. Recompute |
| 244 |
// `second` from `second_minus_offset` to |
| 245 |
// make this block reusable when `second` |
| 246 |
// is not in scope. |
| 247 |
// |
| 248 |
// `third` is guaranteed to be in the range |
| 249 |
// that makes it become the new `self.first`. |
| 250 |
// |
| 251 |
// `fourth` gets unread and the previous |
| 252 |
// `first` gets designates as being in error. |
| 253 |
let fourth_minus_offset = fourth.wrapping_sub(0x30); |
| 254 |
if fourth_minus_offset > (0x39 - 0x30) { |
| 255 |
self.pending_ascii = Some(second_minus_offset + 0x30); |
| 256 |
self.pending = Gb18030Pending::One(third_minus_offset); |
| 257 |
return (DecoderResult::Malformed(1, 2), |
| 258 |
unread_handle_fourth.unread(), |
| 259 |
handle.written()); |
| 260 |
} |
| 261 |
let pointer = (first_minus_offset as usize * (10 * 126 * 10)) + |
| 262 |
(second_minus_offset as usize * (10 * 126)) + |
| 263 |
(third_minus_offset as usize * 10) + |
| 264 |
fourth_minus_offset as usize; |
| 265 |
if pointer <= 39419 { |
| 266 |
// BMP |
| 267 |
if pointer == 7457 { |
| 268 |
handle.write_upper_bmp(0xE7C7) |
| 269 |
} else { |
| 270 |
handle.write_bmp_excl_ascii(gb18030_range_decode(pointer as u16)) |
| 271 |
} |
| 272 |
} else if pointer >= 189_000 && pointer <= 1_237_575 { |
| 273 |
// Astral |
| 274 |
handle.write_astral((pointer - (189_000usize - 0x1_0000usize)) as u32) |
| 275 |
} else { |
| 276 |
return (DecoderResult::Malformed(4, 0), |
| 277 |
unread_handle_fourth.consumed(), |
| 278 |
handle.written()); |
| 279 |
} |
| 280 |
}, |
| 281 |
self, |
| 282 |
non_ascii, |
| 283 |
first_minus_offset, |
| 284 |
second, |
| 285 |
second_minus_offset, |
| 286 |
unread_handle_second, |
| 287 |
third, |
| 288 |
third_minus_offset, |
| 289 |
unread_handle_third, |
| 290 |
fourth, |
| 291 |
fourth_minus_offset, |
| 292 |
unread_handle_fourth, |
| 293 |
source, |
| 294 |
handle, |
| 295 |
'outermost); |
| 296 |
} |
| 297 |
|
| 298 |
// XXX Experiment with inline directives |
| 299 |
fn gbk_encode_non_unified(bmp: u16) -> Option<(usize, usize)> { |
| 300 |
// Try ideographic punctuation first as it's the most likely case. |
| 301 |
// Throwing in the check for full-width currencies and tilde is probably |
| 302 |
// more size-efficient here than elsewhere. |
| 303 |
if in_inclusive_range16(bmp, 0x2014, 0x3017) || in_inclusive_range16(bmp, 0xFF04, 0xFFE1) { |
| 304 |
if let Some(pos) = position(&GB2312_SYMBOLS[..], bmp) { |
| 305 |
return Some((0xA1, pos + 0xA1)); |
| 306 |
} |
| 307 |
} |
| 308 |
// Ext A |
| 309 |
if in_range16(bmp, 0x3400, 0x4E00) { |
| 310 |
return position(&GBK_BOTTOM[21..100], bmp).map(|pos| { |
| 311 |
( |
| 312 |
0xFE, |
| 313 |
pos + if pos < (0x3F - 16) { |
| 314 |
0x40 + 16 |
| 315 |
} else { |
| 316 |
0x41 + 16 |
| 317 |
}, |
| 318 |
) |
| 319 |
}); |
| 320 |
} |
| 321 |
// Compatibility ideographs |
| 322 |
if in_range16(bmp, 0xF900, 0xFB00) { |
| 323 |
return position(&GBK_BOTTOM[0..21], bmp).map(|pos| { |
| 324 |
if pos < 5 { |
| 325 |
// end of second to last row |
| 326 |
(0xFD, pos + (190 - 94 - 5 + 0x41)) |
| 327 |
} else { |
| 328 |
// last row |
| 329 |
(0xFE, pos + (0x40 - 5)) |
| 330 |
} |
| 331 |
}); |
| 332 |
} |
| 333 |
// Handle everything below U+02CA, which is in GBK_OTHER. |
| 334 |
if bmp < 0x02CA { |
| 335 |
if in_range16(bmp, 0x00E0, 0x0262) && bmp != 0x00F7 { |
| 336 |
// Pinyin except U+1E3F |
| 337 |
if let Some(pos) = position(&GB2312_PINYIN[..], bmp) { |
| 338 |
return Some((0xA8, pos + 0xA1)); |
| 339 |
} |
| 340 |
} else if in_inclusive_range16(bmp, 0x00A4, 0x00F7) |
| 341 |
|| in_inclusive_range16(bmp, 0x02C7, 0x02C9) |
| 342 |
{ |
| 343 |
// Diacritics and Latin 1 symbols |
| 344 |
if let Some(pos) = position(&GB2312_SYMBOLS[3..(0xAC - 0x60)], bmp) { |
| 345 |
return Some((0xA1, pos + 0xA1 + 3)); |
| 346 |
} |
| 347 |
} |
| 348 |
return None; |
| 349 |
} |
| 350 |
if bmp >= 0xE794 { |
| 351 |
// Various brackets, all in PUA or full-width regions |
| 352 |
if let Some(pos) = position(&GB2312_SYMBOLS_AFTER_GREEK[..], bmp) { |
| 353 |
return Some((0xA6, pos + (0x9F - 0x60 + 0xA1))); |
| 354 |
} |
| 355 |
} else if bmp == 0x1E3F { |
| 356 |
// The one Pinyin placed elsewhere on the BMP |
| 357 |
return Some((0xA8, 0x7B - 0x60 + 0xA1)); |
| 358 |
} else if in_range16(bmp, 0xA000, 0xD800) { |
| 359 |
// Since Korean has usage in China, let's spend a branch to fast-track |
| 360 |
// Hangul. |
| 361 |
return None; |
| 362 |
} |
| 363 |
// GB2312 other (except bottom PUA and PUA between Hanzi levels). |
| 364 |
if let Some(other_pointer) = gb2312_other_encode(bmp) { |
| 365 |
let other_lead = other_pointer as usize / 94; |
| 366 |
let other_trail = other_pointer as usize % 94; |
| 367 |
return Some((0xA2 + other_lead, 0xA1 + other_trail)); |
| 368 |
} |
| 369 |
// At this point, we've handled all mappable characters above U+02D9 but |
| 370 |
// below U+2010. Let's check for that range in order to let lower BMP |
| 371 |
// characters used for minority languages in China avoid the subsequent |
| 372 |
// search that deals mainly with various symbols. |
| 373 |
if in_range16(bmp, 0x02DA, 0x2010) { |
| 374 |
return None; |
| 375 |
} |
| 376 |
// GBK other (except radicals and PUA in GBK_BOTTOM). |
| 377 |
if let Some(other_pointer) = gbk_other_encode(bmp) { |
| 378 |
let other_lead = other_pointer as usize / (190 - 94); |
| 379 |
let other_trail = other_pointer as usize % (190 - 94); |
| 380 |
let offset = if other_trail < 0x3F { 0x40 } else { 0x41 }; |
| 381 |
return Some((other_lead + (0x81 + 0x20), other_trail + offset)); |
| 382 |
} |
| 383 |
// CJK Radicals Supplement or PUA in GBK_BOTTOM |
| 384 |
if in_inclusive_range16(bmp, 0x2E81, 0x2ECA) || in_inclusive_range16(bmp, 0xE816, 0xE864) { |
| 385 |
if let Some(pos) = position(&GBK_BOTTOM[21..], bmp) { |
| 386 |
let trail = pos + 16; |
| 387 |
let offset = if trail < 0x3F { 0x40 } else { 0x41 }; |
| 388 |
return Some((0xFE, trail + offset)); |
| 389 |
} |
| 390 |
} |
| 391 |
// GB2312 bottom PUA |
| 392 |
let bmp_minus_gb2312_bottom_pua = bmp.wrapping_sub(0xE234); |
| 393 |
if bmp_minus_gb2312_bottom_pua <= (0xE4C5 - 0xE234) { |
| 394 |
let pua_lead = bmp_minus_gb2312_bottom_pua as usize / 94; |
| 395 |
let pua_trail = bmp_minus_gb2312_bottom_pua as usize % 94; |
| 396 |
return Some((0x81 + 0x77 + pua_lead, 0xA1 + pua_trail)); |
| 397 |
} |
| 398 |
// PUA between Hanzi Levels |
| 399 |
let bmp_minus_pua_between_hanzi = bmp.wrapping_sub(0xE810); |
| 400 |
if bmp_minus_pua_between_hanzi < 5 { |
| 401 |
return Some((0x81 + 0x56, 0xFF - 5 + bmp_minus_pua_between_hanzi as usize)); |
| 402 |
} |
| 403 |
None |
| 404 |
} |
| 405 |
|
| 406 |
#[cfg(not(feature = "fast-gb-hanzi-encode"))] |
| 407 |
#[inline(always)] |
| 408 |
fn encode_hanzi(bmp: u16, _: u16) -> (u8, u8) { |
| 409 |
if let Some((lead, trail)) = gb2312_level1_hanzi_encode(bmp) { |
| 410 |
(lead, trail) |
| 411 |
} else if let Some(hanzi_pointer) = gb2312_level2_hanzi_encode(bmp) { |
| 412 |
let hanzi_lead = (hanzi_pointer / 94) + (0xD8); |
| 413 |
let hanzi_trail = (hanzi_pointer % 94) + 0xA1; |
| 414 |
(hanzi_lead as u8, hanzi_trail as u8) |
| 415 |
} else { |
| 416 |
let (lead, gbk_trail) = if bmp < 0x72DC { |
| 417 |
// Above GB2312 |
| 418 |
let pointer = gbk_top_ideograph_encode(bmp) as usize; |
| 419 |
let lead = (pointer / 190) + 0x81; |
| 420 |
let gbk_trail = pointer % 190; |
| 421 |
(lead, gbk_trail) |
| 422 |
} else { |
| 423 |
// To the left of GB2312 |
| 424 |
let gbk_left_ideograph_pointer = gbk_left_ideograph_encode(bmp) as usize; |
| 425 |
let lead = (gbk_left_ideograph_pointer / (190 - 94)) + (0x81 + 0x29); |
| 426 |
let gbk_trail = gbk_left_ideograph_pointer % (190 - 94); |
| 427 |
(lead, gbk_trail) |
| 428 |
}; |
| 429 |
let offset = if gbk_trail < 0x3F { 0x40 } else { 0x41 }; |
| 430 |
(lead as u8, (gbk_trail + offset) as u8) |
| 431 |
} |
| 432 |
} |
| 433 |
|
| 434 |
#[cfg(feature = "fast-gb-hanzi-encode")] |
| 435 |
#[inline(always)] |
| 436 |
fn encode_hanzi(_: u16, bmp_minus_unified_start: u16) -> (u8, u8) { |
| 437 |
gbk_hanzi_encode(bmp_minus_unified_start) |
| 438 |
} |
| 439 |
|
| 440 |
pub struct Gb18030Encoder { |
| 441 |
extended: bool, |
| 442 |
} |
| 443 |
|
| 444 |
impl Gb18030Encoder { |
| 445 |
pub fn new(encoding: &'static Encoding, extended_range: bool) -> Encoder { |
| 446 |
Encoder::new( |
| 447 |
encoding, |
| 448 |
VariantEncoder::Gb18030(Gb18030Encoder { |
| 449 |
extended: extended_range, |
| 450 |
}), |
| 451 |
) |
| 452 |
} |
| 453 |
|
| 454 |
pub fn max_buffer_length_from_utf16_without_replacement( |
| 455 |
&self, |
| 456 |
u16_length: usize, |
| 457 |
) -> Option<usize> { |
| 458 |
if self.extended { |
| 459 |
u16_length.checked_mul(4) |
| 460 |
} else { |
| 461 |
// Need to add, because space check is done with the four-byte |
| 462 |
// assumption. |
| 463 |
checked_add(2, u16_length.checked_mul(2)) |
| 464 |
} |
| 465 |
} |
| 466 |
|
| 467 |
pub fn max_buffer_length_from_utf8_without_replacement( |
| 468 |
&self, |
| 469 |
byte_length: usize, |
| 470 |
) -> Option<usize> { |
| 471 |
if self.extended { |
| 472 |
// 1 to 1 |
| 473 |
// 2 to 2 |
| 474 |
// 3 to 2 |
| 475 |
// 2 to 4 (worst) |
| 476 |
// 3 to 4 |
| 477 |
// 4 to 4 |
| 478 |
checked_add(2, byte_length.checked_mul(2)) |
| 479 |
} else { |
| 480 |
// 1 to 1 |
| 481 |
// 2 to 2 |
| 482 |
// 3 to 2 |
| 483 |
// Need to add, because space check is done with the four-byte |
| 484 |
// assumption. |
| 485 |
byte_length.checked_add(3) |
| 486 |
} |
| 487 |
} |
| 488 |
|
| 489 |
ascii_compatible_encoder_functions!( |
| 490 |
{ |
| 491 |
let bmp_minus_unified_start = bmp.wrapping_sub(0x4E00); |
| 492 |
if bmp_minus_unified_start < (0x9FA6 - 0x4E00) { |
| 493 |
// CJK Unified Ideographs |
| 494 |
// Can't fail now, since all are |
| 495 |
// mapped. |
| 496 |
let (lead, trail) = encode_hanzi(bmp, bmp_minus_unified_start); |
| 497 |
handle.write_two(lead, trail) |
| 498 |
} else if bmp == 0xE5E5 { |
| 499 |
// It's not optimal to check for the unmappable |
| 500 |
// and for euro at this stage, but getting |
| 501 |
// the out of the way makes the rest of the |
| 502 |
// code less messy. |
| 503 |
return ( |
| 504 |
EncoderResult::unmappable_from_bmp(bmp), |
| 505 |
source.consumed(), |
| 506 |
handle.written(), |
| 507 |
); |
| 508 |
} else if bmp == 0x20AC && !self.extended { |
| 509 |
handle.write_one(0x80u8) |
| 510 |
} else { |
| 511 |
match gbk_encode_non_unified(bmp) { |
| 512 |
Some((lead, trail)) => handle.write_two(lead as u8, trail as u8), |
| 513 |
None => { |
| 514 |
if !self.extended { |
| 515 |
return ( |
| 516 |
EncoderResult::unmappable_from_bmp(bmp), |
| 517 |
source.consumed(), |
| 518 |
handle.written(), |
| 519 |
); |
| 520 |
} |
| 521 |
let range_pointer = gb18030_range_encode(bmp); |
| 522 |
let first = range_pointer / (10 * 126 * 10); |
| 523 |
let rem_first = range_pointer % (10 * 126 * 10); |
| 524 |
let second = rem_first / (10 * 126); |
| 525 |
let rem_second = rem_first % (10 * 126); |
| 526 |
let third = rem_second / 10; |
| 527 |
let fourth = rem_second % 10; |
| 528 |
handle.write_four( |
| 529 |
(first + 0x81) as u8, |
| 530 |
(second + 0x30) as u8, |
| 531 |
(third + 0x81) as u8, |
| 532 |
(fourth + 0x30) as u8, |
| 533 |
) |
| 534 |
} |
| 535 |
} |
| 536 |
} |
| 537 |
}, |
| 538 |
{ |
| 539 |
if !self.extended { |
| 540 |
return ( |
| 541 |
EncoderResult::Unmappable(astral), |
| 542 |
source.consumed(), |
| 543 |
handle.written(), |
| 544 |
); |
| 545 |
} |
| 546 |
let range_pointer = astral as usize + (189_000usize - 0x1_0000usize); |
| 547 |
let first = range_pointer / (10 * 126 * 10); |
| 548 |
let rem_first = range_pointer % (10 * 126 * 10); |
| 549 |
let second = rem_first / (10 * 126); |
| 550 |
let rem_second = rem_first % (10 * 126); |
| 551 |
let third = rem_second / 10; |
| 552 |
let fourth = rem_second % 10; |
| 553 |
handle.write_four( |
| 554 |
(first + 0x81) as u8, |
| 555 |
(second + 0x30) as u8, |
| 556 |
(third + 0x81) as u8, |
| 557 |
(fourth + 0x30) as u8, |
| 558 |
) |
| 559 |
}, |
| 560 |
bmp, |
| 561 |
astral, |
| 562 |
self, |
| 563 |
source, |
| 564 |
handle, |
| 565 |
copy_ascii_to_check_space_four, |
| 566 |
check_space_four, |
| 567 |
false |
| 568 |
); |
| 569 |
} |
| 570 |
|
| 571 |
// Any copyright to the test code below this comment is dedicated to the |
| 572 |
// Public Domain. http://creativecommons.org/publicdomain/zero/1.0/ |
| 573 |
|
| 574 |
#[cfg(all(test, feature = "alloc"))] |
| 575 |
mod tests { |
| 576 |
use super::super::testing::*; |
| 577 |
use super::super::*; |
| 578 |
|
| 579 |
fn decode_gb18030(bytes: &[u8], expect: &str) { |
| 580 |
decode(GB18030, bytes, expect); |
| 581 |
} |
| 582 |
|
| 583 |
fn encode_gb18030(string: &str, expect: &[u8]) { |
| 584 |
encode(GB18030, string, expect); |
| 585 |
} |
| 586 |
|
| 587 |
fn encode_gbk(string: &str, expect: &[u8]) { |
| 588 |
encode(GBK, string, expect); |
| 589 |
} |
| 590 |
|
| 591 |
#[test] |
| 592 |
fn test_gb18030_decode() { |
| 593 |
// Empty |
| 594 |
decode_gb18030(b"", &""); |
| 595 |
|
| 596 |
// ASCII |
| 597 |
decode_gb18030(b"\x61\x62", "\u{0061}\u{0062}"); |
| 598 |
|
| 599 |
// euro |
| 600 |
decode_gb18030(b"\x80", "\u{20AC}"); |
| 601 |
decode_gb18030(b"\xA2\xE3", "\u{20AC}"); |
| 602 |
|
| 603 |
// two bytes |
| 604 |
decode_gb18030(b"\x81\x40", "\u{4E02}"); |
| 605 |
decode_gb18030(b"\x81\x7E", "\u{4E8A}"); |
| 606 |
decode_gb18030(b"\x81\x7F", "\u{FFFD}\u{007F}"); |
| 607 |
decode_gb18030(b"\x81\x80", "\u{4E90}"); |
| 608 |
decode_gb18030(b"\x81\xFE", "\u{4FA2}"); |
| 609 |
decode_gb18030(b"\xFE\x40", "\u{FA0C}"); |
| 610 |
decode_gb18030(b"\xFE\x7E", "\u{E843}"); |
| 611 |
decode_gb18030(b"\xFE\x7F", "\u{FFFD}\u{007F}"); |
| 612 |
decode_gb18030(b"\xFE\x80", "\u{4723}"); |
| 613 |
decode_gb18030(b"\xFE\xFE", "\u{E4C5}"); |
| 614 |
|
| 615 |
// The difference from the original GB18030 |
| 616 |
decode_gb18030(b"\xA3\xA0", "\u{3000}"); |
| 617 |
decode_gb18030(b"\xA1\xA1", "\u{3000}"); |
| 618 |
|
| 619 |
// 0xFF |
| 620 |
decode_gb18030(b"\xFF\x40", "\u{FFFD}\u{0040}"); |
| 621 |
decode_gb18030(b"\xE3\xFF\x9A\x33", "\u{FFFD}\u{FFFD}"); // not \u{FFFD}\u{FFFD}\u{0033} ! |
| 622 |
decode_gb18030(b"\xFF\x32\x9A\x33", "\u{FFFD}\u{0032}\u{FFFD}"); // not \u{FFFD}\u{0032}\u{FFFD}\u{0033} ! |
| 623 |
decode_gb18030(b"\xFF\x40\x00", "\u{FFFD}\u{0040}\u{0000}"); |
| 624 |
decode_gb18030(b"\xE3\xFF\x9A\x33\x00", "\u{FFFD}\u{FFFD}\u{0033}\u{0000}"); |
| 625 |
decode_gb18030( |
| 626 |
b"\xFF\x32\x9A\x33\x00", |
| 627 |
"\u{FFFD}\u{0032}\u{FFFD}\u{0033}\u{0000}", |
| 628 |
); |
| 629 |
|
| 630 |
// Four bytes |
| 631 |
decode_gb18030(b"\x81\x30\x81\x30", "\u{0080}"); |
| 632 |
decode_gb18030(b"\x81\x35\xF4\x37", "\u{E7C7}"); |
| 633 |
decode_gb18030(b"\x81\x37\xA3\x30", "\u{2603}"); |
| 634 |
decode_gb18030(b"\x94\x39\xDA\x33", "\u{1F4A9}"); |
| 635 |
decode_gb18030(b"\xE3\x32\x9A\x35", "\u{10FFFF}"); |
| 636 |
decode_gb18030(b"\xE3\x32\x9A\x36\x81\x30", "\u{FFFD}\u{FFFD}"); |
| 637 |
decode_gb18030(b"\xE3\x32\x9A\x36\x81\x40", "\u{FFFD}\u{4E02}"); |
| 638 |
decode_gb18030(b"\xE3\x32\x9A", "\u{FFFD}"); // not \u{FFFD}\u{0032}\u{FFFD} ! |
| 639 |
decode_gb18030(b"\xE3\x32\x9A\x00", "\u{FFFD}\u{0032}\u{FFFD}\u{0000}"); |
| 640 |
} |
| 641 |
|
| 642 |
#[test] |
| 643 |
fn test_gb18030_encode() { |
| 644 |
// Empty |
| 645 |
encode_gb18030("", b""); |
| 646 |
|
| 647 |
// ASCII |
| 648 |
encode_gb18030("\u{0061}\u{0062}", b"\x61\x62"); |
| 649 |
|
| 650 |
// euro |
| 651 |
encode_gb18030("\u{20AC}", b"\xA2\xE3"); |
| 652 |
|
| 653 |
// two bytes |
| 654 |
encode_gb18030("\u{4E02}", b"\x81\x40"); |
| 655 |
encode_gb18030("\u{4E8A}", b"\x81\x7E"); |
| 656 |
if !cfg!(miri) { |
| 657 |
// Miri is too slow |
| 658 |
encode_gb18030("\u{4E90}", b"\x81\x80"); |
| 659 |
encode_gb18030("\u{4FA2}", b"\x81\xFE"); |
| 660 |
encode_gb18030("\u{FA0C}", b"\xFE\x40"); |
| 661 |
encode_gb18030("\u{E843}", b"\xFE\x7E"); |
| 662 |
encode_gb18030("\u{4723}", b"\xFE\x80"); |
| 663 |
encode_gb18030("\u{E4C5}", b"\xFE\xFE"); |
| 664 |
} |
| 665 |
|
| 666 |
// The difference from the original GB18030 |
| 667 |
encode_gb18030("\u{E5E5}", b""); |
| 668 |
encode_gb18030("\u{3000}", b"\xA1\xA1"); |
| 669 |
|
| 670 |
// Four bytes |
| 671 |
encode_gb18030("\u{0080}", b"\x81\x30\x81\x30"); |
| 672 |
encode_gb18030("\u{E7C7}", b"\x81\x35\xF4\x37"); |
| 673 |
if !cfg!(miri) { |
| 674 |
// Miri is too slow |
| 675 |
encode_gb18030("\u{2603}", b"\x81\x37\xA3\x30"); |
| 676 |
encode_gb18030("\u{1F4A9}", b"\x94\x39\xDA\x33"); |
| 677 |
encode_gb18030("\u{10FFFF}", b"\xE3\x32\x9A\x35"); |
| 678 |
} |
| 679 |
|
| 680 |
// Edge cases |
| 681 |
encode_gb18030("\u{00F7}", b"\xA1\xC2"); |
| 682 |
} |
| 683 |
|
| 684 |
#[test] |
| 685 |
fn test_gbk_encode() { |
| 686 |
// Empty |
| 687 |
encode_gbk("", b""); |
| 688 |
|
| 689 |
// ASCII |
| 690 |
encode_gbk("\u{0061}\u{0062}", b"\x61\x62"); |
| 691 |
|
| 692 |
// euro |
| 693 |
encode_gbk("\u{20AC}", b"\x80"); |
| 694 |
|
| 695 |
// two bytes |
| 696 |
encode_gbk("\u{4E02}", b"\x81\x40"); |
| 697 |
encode_gbk("\u{4E8A}", b"\x81\x7E"); |
| 698 |
if !cfg!(miri) { |
| 699 |
// Miri is too slow |
| 700 |
encode_gbk("\u{4E90}", b"\x81\x80"); |
| 701 |
encode_gbk("\u{4FA2}", b"\x81\xFE"); |
| 702 |
encode_gbk("\u{FA0C}", b"\xFE\x40"); |
| 703 |
encode_gbk("\u{E843}", b"\xFE\x7E"); |
| 704 |
encode_gbk("\u{4723}", b"\xFE\x80"); |
| 705 |
encode_gbk("\u{E4C5}", b"\xFE\xFE"); |
| 706 |
} |
| 707 |
|
| 708 |
// The difference from the original gb18030 |
| 709 |
encode_gbk("\u{E5E5}", b""); |
| 710 |
encode_gbk("\u{3000}", b"\xA1\xA1"); |
| 711 |
|
| 712 |
// Four bytes |
| 713 |
encode_gbk("\u{0080}", b"€"); |
| 714 |
encode_gbk("\u{E7C7}", b""); |
| 715 |
if !cfg!(miri) { |
| 716 |
// Miri is too slow |
| 717 |
encode_gbk("\u{2603}", b"☃"); |
| 718 |
encode_gbk("\u{1F4A9}", b"💩"); |
| 719 |
encode_gbk("\u{10FFFF}", b""); |
| 720 |
} |
| 721 |
|
| 722 |
// Edge cases |
| 723 |
encode_gbk("\u{00F7}", b"\xA1\xC2"); |
| 724 |
} |
| 725 |
|
| 726 |
#[test] |
| 727 |
#[cfg_attr(miri, ignore)] // Miri is too slow |
| 728 |
fn test_gb18030_decode_all() { |
| 729 |
let input = include_bytes!("test_data/gb18030_in.txt"); |
| 730 |
let expectation = include_str!("test_data/gb18030_in_ref.txt"); |
| 731 |
let (cow, had_errors) = GB18030.decode_without_bom_handling(input); |
| 732 |
assert!(!had_errors, "Should not have had errors."); |
| 733 |
assert_eq!(&cow[..], expectation); |
| 734 |
} |
| 735 |
|
| 736 |
#[test] |
| 737 |
#[cfg_attr(miri, ignore)] // Miri is too slow |
| 738 |
fn test_gb18030_encode_all() { |
| 739 |
let input = include_str!("test_data/gb18030_out.txt"); |
| 740 |
let expectation = include_bytes!("test_data/gb18030_out_ref.txt"); |
| 741 |
let (cow, encoding, had_errors) = GB18030.encode(input); |
| 742 |
assert!(!had_errors, "Should not have had errors."); |
| 743 |
assert_eq!(encoding, GB18030); |
| 744 |
assert_eq!(&cow[..], &expectation[..]); |
| 745 |
} |
| 746 |
|
| 747 |
#[test] |
| 748 |
fn test_gb18030_encode_from_utf16_max_length() { |
| 749 |
let mut output = [0u8; 20]; |
| 750 |
let mut encoder = GB18030.new_encoder(); |
| 751 |
{ |
| 752 |
let needed = encoder |
| 753 |
.max_buffer_length_from_utf16_without_replacement(1) |
| 754 |
.unwrap(); |
| 755 |
let (result, read, written) = encoder.encode_from_utf16_without_replacement( |
| 756 |
&[0x3000], |
| 757 |
&mut output[..needed], |
| 758 |
true, |
| 759 |
); |
| 760 |
assert_eq!(result, EncoderResult::InputEmpty); |
| 761 |
assert_eq!(read, 1); |
| 762 |
assert_eq!(written, 2); |
| 763 |
assert_eq!(output[0], 0xA1); |
| 764 |
assert_eq!(output[1], 0xA1); |
| 765 |
} |
| 766 |
} |
| 767 |
} |