Ben Murdoch | c561043 | 2016-08-08 18:44:38 +0100 | [diff] [blame] | 1 | // Copyright 2016 the V8 project authors. All rights reserved. |
| 2 | // Use of this source code is governed by a BSD-style license that can be |
| 3 | // found in the LICENSE file. |
| 4 | |
| 5 | #include "src/uri.h" |
| 6 | |
| 7 | #include "src/char-predicates-inl.h" |
| 8 | #include "src/handles.h" |
| 9 | #include "src/isolate-inl.h" |
| 10 | #include "src/list.h" |
Ben Murdoch | 61f157c | 2016-09-16 13:49:30 +0100 | [diff] [blame] | 11 | #include "src/string-search.h" |
Ben Murdoch | c561043 | 2016-08-08 18:44:38 +0100 | [diff] [blame] | 12 | |
| 13 | namespace v8 { |
| 14 | namespace internal { |
| 15 | |
Ben Murdoch | 61f157c | 2016-09-16 13:49:30 +0100 | [diff] [blame] | 16 | namespace { // anonymous namespace for DecodeURI helper functions |
| 17 | bool IsReservedPredicate(uc16 c) { |
| 18 | switch (c) { |
| 19 | case '#': |
| 20 | case '$': |
| 21 | case '&': |
| 22 | case '+': |
| 23 | case ',': |
| 24 | case '/': |
| 25 | case ':': |
| 26 | case ';': |
| 27 | case '=': |
| 28 | case '?': |
| 29 | case '@': |
| 30 | return true; |
| 31 | default: |
| 32 | return false; |
| 33 | } |
| 34 | } |
| 35 | |
| 36 | bool IsReplacementCharacter(const uint8_t* octets, int length) { |
| 37 | // The replacement character is at codepoint U+FFFD in the Unicode Specials |
| 38 | // table. Its UTF-8 encoding is 0xEF 0xBF 0xBD. |
| 39 | if (length != 3 || octets[0] != 0xef || octets[1] != 0xbf || |
| 40 | octets[2] != 0xbd) { |
| 41 | return false; |
| 42 | } |
| 43 | return true; |
| 44 | } |
| 45 | |
| 46 | bool DecodeOctets(const uint8_t* octets, int length, List<uc16>* buffer) { |
| 47 | size_t cursor = 0; |
| 48 | uc32 value = unibrow::Utf8::ValueOf(octets, length, &cursor); |
| 49 | if (value == unibrow::Utf8::kBadChar && |
| 50 | !IsReplacementCharacter(octets, length)) { |
| 51 | return false; |
| 52 | } |
| 53 | |
| 54 | if (value <= unibrow::Utf16::kMaxNonSurrogateCharCode) { |
| 55 | buffer->Add(value); |
| 56 | } else { |
| 57 | buffer->Add(unibrow::Utf16::LeadSurrogate(value)); |
| 58 | buffer->Add(unibrow::Utf16::TrailSurrogate(value)); |
| 59 | } |
| 60 | return true; |
| 61 | } |
| 62 | |
| 63 | int TwoDigitHex(uc16 character1, uc16 character2) { |
| 64 | if (character1 > 'f') return -1; |
| 65 | int high = HexValue(character1); |
| 66 | if (high == -1) return -1; |
| 67 | if (character2 > 'f') return -1; |
| 68 | int low = HexValue(character2); |
| 69 | if (low == -1) return -1; |
| 70 | return (high << 4) + low; |
| 71 | } |
| 72 | |
| 73 | template <typename T> |
| 74 | void AddToBuffer(uc16 decoded, String::FlatContent* uri_content, int index, |
| 75 | bool is_uri, List<T>* buffer) { |
| 76 | if (is_uri && IsReservedPredicate(decoded)) { |
| 77 | buffer->Add('%'); |
| 78 | uc16 first = uri_content->Get(index + 1); |
| 79 | uc16 second = uri_content->Get(index + 2); |
| 80 | DCHECK_GT(std::numeric_limits<T>::max(), first); |
| 81 | DCHECK_GT(std::numeric_limits<T>::max(), second); |
| 82 | |
| 83 | buffer->Add(first); |
| 84 | buffer->Add(second); |
| 85 | } else { |
| 86 | buffer->Add(decoded); |
| 87 | } |
| 88 | } |
| 89 | |
| 90 | bool IntoTwoByte(int index, bool is_uri, int uri_length, |
| 91 | String::FlatContent* uri_content, List<uc16>* buffer) { |
| 92 | for (int k = index; k < uri_length; k++) { |
| 93 | uc16 code = uri_content->Get(k); |
| 94 | if (code == '%') { |
| 95 | uc16 decoded; |
| 96 | if (k + 2 >= uri_length || |
| 97 | (decoded = TwoDigitHex(uri_content->Get(k + 1), |
| 98 | uri_content->Get(k + 2))) < 0) { |
| 99 | return false; |
| 100 | } |
| 101 | k += 2; |
| 102 | if (decoded > unibrow::Utf8::kMaxOneByteChar) { |
| 103 | uint8_t octets[unibrow::Utf8::kMaxEncodedSize]; |
| 104 | octets[0] = decoded; |
| 105 | |
| 106 | int number_of_continuation_bytes = 0; |
| 107 | while ((decoded << ++number_of_continuation_bytes) & 0x80) { |
| 108 | if (number_of_continuation_bytes > 3 || k + 3 >= uri_length) { |
| 109 | return false; |
| 110 | } |
| 111 | |
| 112 | uc16 continuation_byte; |
| 113 | |
| 114 | if (uri_content->Get(++k) != '%' || |
| 115 | (continuation_byte = TwoDigitHex(uri_content->Get(k + 1), |
| 116 | uri_content->Get(k + 2))) < 0) { |
| 117 | return false; |
| 118 | } |
| 119 | k += 2; |
| 120 | octets[number_of_continuation_bytes] = continuation_byte; |
| 121 | } |
| 122 | |
| 123 | if (!DecodeOctets(octets, number_of_continuation_bytes, buffer)) { |
| 124 | return false; |
| 125 | } |
| 126 | } else { |
| 127 | AddToBuffer(decoded, uri_content, k - 2, is_uri, buffer); |
| 128 | } |
| 129 | } else { |
| 130 | buffer->Add(code); |
| 131 | } |
| 132 | } |
| 133 | return true; |
| 134 | } |
| 135 | |
| 136 | bool IntoOneAndTwoByte(Handle<String> uri, bool is_uri, |
| 137 | List<uint8_t>* one_byte_buffer, |
| 138 | List<uc16>* two_byte_buffer) { |
| 139 | DisallowHeapAllocation no_gc; |
| 140 | String::FlatContent uri_content = uri->GetFlatContent(); |
| 141 | |
| 142 | int uri_length = uri->length(); |
| 143 | for (int k = 0; k < uri_length; k++) { |
| 144 | uc16 code = uri_content.Get(k); |
| 145 | if (code == '%') { |
| 146 | uc16 decoded; |
| 147 | if (k + 2 >= uri_length || |
| 148 | (decoded = TwoDigitHex(uri_content.Get(k + 1), |
| 149 | uri_content.Get(k + 2))) < 0) { |
| 150 | return false; |
| 151 | } |
| 152 | |
| 153 | if (decoded > unibrow::Utf8::kMaxOneByteChar) { |
| 154 | return IntoTwoByte(k, is_uri, uri_length, &uri_content, |
| 155 | two_byte_buffer); |
| 156 | } |
| 157 | |
| 158 | AddToBuffer(decoded, &uri_content, k, is_uri, one_byte_buffer); |
| 159 | k += 2; |
| 160 | } else { |
| 161 | if (code > unibrow::Utf8::kMaxOneByteChar) { |
| 162 | return IntoTwoByte(k, is_uri, uri_length, &uri_content, |
| 163 | two_byte_buffer); |
| 164 | } |
| 165 | one_byte_buffer->Add(code); |
| 166 | } |
| 167 | } |
| 168 | return true; |
| 169 | } |
| 170 | |
| 171 | } // anonymous namespace |
| 172 | |
| 173 | MaybeHandle<String> Uri::Decode(Isolate* isolate, Handle<String> uri, |
| 174 | bool is_uri) { |
| 175 | uri = String::Flatten(uri); |
| 176 | List<uint8_t> one_byte_buffer; |
| 177 | List<uc16> two_byte_buffer; |
| 178 | |
| 179 | if (!IntoOneAndTwoByte(uri, is_uri, &one_byte_buffer, &two_byte_buffer)) { |
| 180 | THROW_NEW_ERROR(isolate, NewURIError(), String); |
| 181 | } |
| 182 | |
| 183 | if (two_byte_buffer.is_empty()) { |
| 184 | return isolate->factory()->NewStringFromOneByte( |
| 185 | one_byte_buffer.ToConstVector()); |
| 186 | } |
| 187 | |
| 188 | Handle<SeqTwoByteString> result; |
| 189 | ASSIGN_RETURN_ON_EXCEPTION( |
| 190 | isolate, result, isolate->factory()->NewRawTwoByteString( |
| 191 | one_byte_buffer.length() + two_byte_buffer.length()), |
| 192 | String); |
| 193 | |
| 194 | CopyChars(result->GetChars(), one_byte_buffer.ToConstVector().start(), |
| 195 | one_byte_buffer.length()); |
| 196 | CopyChars(result->GetChars() + one_byte_buffer.length(), |
| 197 | two_byte_buffer.ToConstVector().start(), two_byte_buffer.length()); |
| 198 | |
| 199 | return result; |
| 200 | } |
| 201 | |
Ben Murdoch | c561043 | 2016-08-08 18:44:38 +0100 | [diff] [blame] | 202 | namespace { // anonymous namespace for EncodeURI helper functions |
| 203 | bool IsUnescapePredicateInUriComponent(uc16 c) { |
| 204 | if (IsAlphaNumeric(c)) { |
| 205 | return true; |
| 206 | } |
| 207 | |
| 208 | switch (c) { |
| 209 | case '!': |
| 210 | case '\'': |
| 211 | case '(': |
| 212 | case ')': |
| 213 | case '*': |
| 214 | case '-': |
| 215 | case '.': |
| 216 | case '_': |
| 217 | case '~': |
| 218 | return true; |
| 219 | default: |
| 220 | return false; |
| 221 | } |
| 222 | } |
| 223 | |
| 224 | bool IsUriSeparator(uc16 c) { |
| 225 | switch (c) { |
| 226 | case '#': |
| 227 | case ':': |
| 228 | case ';': |
| 229 | case '/': |
| 230 | case '?': |
| 231 | case '$': |
| 232 | case '&': |
| 233 | case '+': |
| 234 | case ',': |
| 235 | case '@': |
| 236 | case '=': |
| 237 | return true; |
| 238 | default: |
| 239 | return false; |
| 240 | } |
| 241 | } |
| 242 | |
Ben Murdoch | 61f157c | 2016-09-16 13:49:30 +0100 | [diff] [blame] | 243 | void AddEncodedOctetToBuffer(uint8_t octet, List<uint8_t>* buffer) { |
Ben Murdoch | c561043 | 2016-08-08 18:44:38 +0100 | [diff] [blame] | 244 | buffer->Add('%'); |
| 245 | buffer->Add(HexCharOfValue(octet >> 4)); |
| 246 | buffer->Add(HexCharOfValue(octet & 0x0F)); |
| 247 | } |
| 248 | |
| 249 | void EncodeSingle(uc16 c, List<uint8_t>* buffer) { |
Ben Murdoch | 61f157c | 2016-09-16 13:49:30 +0100 | [diff] [blame] | 250 | char s[4] = {}; |
| 251 | int number_of_bytes; |
| 252 | number_of_bytes = |
| 253 | unibrow::Utf8::Encode(s, c, unibrow::Utf16::kNoPreviousCharacter, false); |
| 254 | for (int k = 0; k < number_of_bytes; k++) { |
| 255 | AddEncodedOctetToBuffer(s[k], buffer); |
Ben Murdoch | c561043 | 2016-08-08 18:44:38 +0100 | [diff] [blame] | 256 | } |
| 257 | } |
| 258 | |
| 259 | void EncodePair(uc16 cc1, uc16 cc2, List<uint8_t>* buffer) { |
Ben Murdoch | 61f157c | 2016-09-16 13:49:30 +0100 | [diff] [blame] | 260 | char s[4] = {}; |
| 261 | int number_of_bytes = |
| 262 | unibrow::Utf8::Encode(s, unibrow::Utf16::CombineSurrogatePair(cc1, cc2), |
| 263 | unibrow::Utf16::kNoPreviousCharacter, false); |
| 264 | for (int k = 0; k < number_of_bytes; k++) { |
| 265 | AddEncodedOctetToBuffer(s[k], buffer); |
| 266 | } |
Ben Murdoch | c561043 | 2016-08-08 18:44:38 +0100 | [diff] [blame] | 267 | } |
| 268 | |
| 269 | } // anonymous namespace |
| 270 | |
Ben Murdoch | 61f157c | 2016-09-16 13:49:30 +0100 | [diff] [blame] | 271 | MaybeHandle<String> Uri::Encode(Isolate* isolate, Handle<String> uri, |
| 272 | bool is_uri) { |
Ben Murdoch | c561043 | 2016-08-08 18:44:38 +0100 | [diff] [blame] | 273 | uri = String::Flatten(uri); |
| 274 | int uri_length = uri->length(); |
| 275 | List<uint8_t> buffer(uri_length); |
| 276 | |
| 277 | { |
| 278 | DisallowHeapAllocation no_gc; |
| 279 | String::FlatContent uri_content = uri->GetFlatContent(); |
| 280 | |
| 281 | for (int k = 0; k < uri_length; k++) { |
| 282 | uc16 cc1 = uri_content.Get(k); |
| 283 | if (unibrow::Utf16::IsLeadSurrogate(cc1)) { |
| 284 | k++; |
| 285 | if (k < uri_length) { |
| 286 | uc16 cc2 = uri->Get(k); |
| 287 | if (unibrow::Utf16::IsTrailSurrogate(cc2)) { |
| 288 | EncodePair(cc1, cc2, &buffer); |
| 289 | continue; |
| 290 | } |
| 291 | } |
| 292 | } else if (!unibrow::Utf16::IsTrailSurrogate(cc1)) { |
| 293 | if (IsUnescapePredicateInUriComponent(cc1) || |
| 294 | (is_uri && IsUriSeparator(cc1))) { |
| 295 | buffer.Add(cc1); |
| 296 | } else { |
| 297 | EncodeSingle(cc1, &buffer); |
| 298 | } |
| 299 | continue; |
| 300 | } |
| 301 | |
| 302 | AllowHeapAllocation allocate_error_and_return; |
Ben Murdoch | 61f157c | 2016-09-16 13:49:30 +0100 | [diff] [blame] | 303 | THROW_NEW_ERROR(isolate, NewURIError(), String); |
Ben Murdoch | c561043 | 2016-08-08 18:44:38 +0100 | [diff] [blame] | 304 | } |
| 305 | } |
| 306 | |
Ben Murdoch | 61f157c | 2016-09-16 13:49:30 +0100 | [diff] [blame] | 307 | return isolate->factory()->NewStringFromOneByte(buffer.ToConstVector()); |
| 308 | } |
| 309 | |
| 310 | namespace { // Anonymous namespace for Escape and Unescape |
| 311 | |
| 312 | template <typename Char> |
| 313 | int UnescapeChar(Vector<const Char> vector, int i, int length, int* step) { |
| 314 | uint16_t character = vector[i]; |
| 315 | int32_t hi = 0; |
| 316 | int32_t lo = 0; |
| 317 | if (character == '%' && i <= length - 6 && vector[i + 1] == 'u' && |
| 318 | (hi = TwoDigitHex(vector[i + 2], vector[i + 3])) > -1 && |
| 319 | (lo = TwoDigitHex(vector[i + 4], vector[i + 5])) > -1) { |
| 320 | *step = 6; |
| 321 | return (hi << 8) + lo; |
| 322 | } else if (character == '%' && i <= length - 3 && |
| 323 | (lo = TwoDigitHex(vector[i + 1], vector[i + 2])) > -1) { |
| 324 | *step = 3; |
| 325 | return lo; |
| 326 | } else { |
| 327 | *step = 1; |
| 328 | return character; |
| 329 | } |
| 330 | } |
| 331 | |
| 332 | template <typename Char> |
| 333 | MaybeHandle<String> UnescapeSlow(Isolate* isolate, Handle<String> string, |
| 334 | int start_index) { |
| 335 | bool one_byte = true; |
| 336 | int length = string->length(); |
| 337 | |
| 338 | int unescaped_length = 0; |
| 339 | { |
| 340 | DisallowHeapAllocation no_allocation; |
| 341 | Vector<const Char> vector = string->GetCharVector<Char>(); |
| 342 | for (int i = start_index; i < length; unescaped_length++) { |
| 343 | int step; |
| 344 | if (UnescapeChar(vector, i, length, &step) > |
| 345 | String::kMaxOneByteCharCode) { |
| 346 | one_byte = false; |
| 347 | } |
| 348 | i += step; |
| 349 | } |
| 350 | } |
| 351 | |
| 352 | DCHECK(start_index < length); |
| 353 | Handle<String> first_part = |
| 354 | isolate->factory()->NewProperSubString(string, 0, start_index); |
| 355 | |
| 356 | int dest_position = 0; |
| 357 | Handle<String> second_part; |
| 358 | DCHECK(unescaped_length <= String::kMaxLength); |
| 359 | if (one_byte) { |
| 360 | Handle<SeqOneByteString> dest = isolate->factory() |
| 361 | ->NewRawOneByteString(unescaped_length) |
| 362 | .ToHandleChecked(); |
| 363 | DisallowHeapAllocation no_allocation; |
| 364 | Vector<const Char> vector = string->GetCharVector<Char>(); |
| 365 | for (int i = start_index; i < length; dest_position++) { |
| 366 | int step; |
| 367 | dest->SeqOneByteStringSet(dest_position, |
| 368 | UnescapeChar(vector, i, length, &step)); |
| 369 | i += step; |
| 370 | } |
| 371 | second_part = dest; |
| 372 | } else { |
| 373 | Handle<SeqTwoByteString> dest = isolate->factory() |
| 374 | ->NewRawTwoByteString(unescaped_length) |
| 375 | .ToHandleChecked(); |
| 376 | DisallowHeapAllocation no_allocation; |
| 377 | Vector<const Char> vector = string->GetCharVector<Char>(); |
| 378 | for (int i = start_index; i < length; dest_position++) { |
| 379 | int step; |
| 380 | dest->SeqTwoByteStringSet(dest_position, |
| 381 | UnescapeChar(vector, i, length, &step)); |
| 382 | i += step; |
| 383 | } |
| 384 | second_part = dest; |
| 385 | } |
| 386 | return isolate->factory()->NewConsString(first_part, second_part); |
| 387 | } |
| 388 | |
| 389 | bool IsNotEscaped(uint16_t c) { |
| 390 | if (IsAlphaNumeric(c)) { |
| 391 | return true; |
| 392 | } |
| 393 | // @*_+-./ |
| 394 | switch (c) { |
| 395 | case '@': |
| 396 | case '*': |
| 397 | case '_': |
| 398 | case '+': |
| 399 | case '-': |
| 400 | case '.': |
| 401 | case '/': |
| 402 | return true; |
| 403 | default: |
| 404 | return false; |
| 405 | } |
| 406 | } |
| 407 | |
| 408 | template <typename Char> |
| 409 | static MaybeHandle<String> UnescapePrivate(Isolate* isolate, |
| 410 | Handle<String> source) { |
| 411 | int index; |
| 412 | { |
| 413 | DisallowHeapAllocation no_allocation; |
| 414 | StringSearch<uint8_t, Char> search(isolate, STATIC_CHAR_VECTOR("%")); |
| 415 | index = search.Search(source->GetCharVector<Char>(), 0); |
| 416 | if (index < 0) return source; |
| 417 | } |
| 418 | return UnescapeSlow<Char>(isolate, source, index); |
| 419 | } |
| 420 | |
| 421 | template <typename Char> |
| 422 | static MaybeHandle<String> EscapePrivate(Isolate* isolate, |
| 423 | Handle<String> string) { |
| 424 | DCHECK(string->IsFlat()); |
| 425 | int escaped_length = 0; |
| 426 | int length = string->length(); |
| 427 | |
| 428 | { |
| 429 | DisallowHeapAllocation no_allocation; |
| 430 | Vector<const Char> vector = string->GetCharVector<Char>(); |
| 431 | for (int i = 0; i < length; i++) { |
| 432 | uint16_t c = vector[i]; |
| 433 | if (c >= 256) { |
| 434 | escaped_length += 6; |
| 435 | } else if (IsNotEscaped(c)) { |
| 436 | escaped_length++; |
| 437 | } else { |
| 438 | escaped_length += 3; |
| 439 | } |
| 440 | |
| 441 | // We don't allow strings that are longer than a maximal length. |
| 442 | DCHECK(String::kMaxLength < 0x7fffffff - 6); // Cannot overflow. |
| 443 | if (escaped_length > String::kMaxLength) break; // Provoke exception. |
| 444 | } |
| 445 | } |
| 446 | |
| 447 | // No length change implies no change. Return original string if no change. |
| 448 | if (escaped_length == length) return string; |
| 449 | |
| 450 | Handle<SeqOneByteString> dest; |
| 451 | ASSIGN_RETURN_ON_EXCEPTION( |
| 452 | isolate, dest, isolate->factory()->NewRawOneByteString(escaped_length), |
| 453 | String); |
| 454 | int dest_position = 0; |
| 455 | |
| 456 | { |
| 457 | DisallowHeapAllocation no_allocation; |
| 458 | Vector<const Char> vector = string->GetCharVector<Char>(); |
| 459 | for (int i = 0; i < length; i++) { |
| 460 | uint16_t c = vector[i]; |
| 461 | if (c >= 256) { |
| 462 | dest->SeqOneByteStringSet(dest_position, '%'); |
| 463 | dest->SeqOneByteStringSet(dest_position + 1, 'u'); |
| 464 | dest->SeqOneByteStringSet(dest_position + 2, HexCharOfValue(c >> 12)); |
| 465 | dest->SeqOneByteStringSet(dest_position + 3, |
| 466 | HexCharOfValue((c >> 8) & 0xf)); |
| 467 | dest->SeqOneByteStringSet(dest_position + 4, |
| 468 | HexCharOfValue((c >> 4) & 0xf)); |
| 469 | dest->SeqOneByteStringSet(dest_position + 5, HexCharOfValue(c & 0xf)); |
| 470 | dest_position += 6; |
| 471 | } else if (IsNotEscaped(c)) { |
| 472 | dest->SeqOneByteStringSet(dest_position, c); |
| 473 | dest_position++; |
| 474 | } else { |
| 475 | dest->SeqOneByteStringSet(dest_position, '%'); |
| 476 | dest->SeqOneByteStringSet(dest_position + 1, HexCharOfValue(c >> 4)); |
| 477 | dest->SeqOneByteStringSet(dest_position + 2, HexCharOfValue(c & 0xf)); |
| 478 | dest_position += 3; |
| 479 | } |
| 480 | } |
| 481 | } |
| 482 | |
| 483 | return dest; |
| 484 | } |
| 485 | |
| 486 | } // Anonymous namespace |
| 487 | |
| 488 | MaybeHandle<String> Uri::Escape(Isolate* isolate, Handle<String> string) { |
Ben Murdoch | c561043 | 2016-08-08 18:44:38 +0100 | [diff] [blame] | 489 | Handle<String> result; |
Ben Murdoch | 61f157c | 2016-09-16 13:49:30 +0100 | [diff] [blame] | 490 | string = String::Flatten(string); |
| 491 | return string->IsOneByteRepresentationUnderneath() |
| 492 | ? EscapePrivate<uint8_t>(isolate, string) |
| 493 | : EscapePrivate<uc16>(isolate, string); |
| 494 | } |
| 495 | |
| 496 | MaybeHandle<String> Uri::Unescape(Isolate* isolate, Handle<String> string) { |
| 497 | Handle<String> result; |
| 498 | string = String::Flatten(string); |
| 499 | return string->IsOneByteRepresentationUnderneath() |
| 500 | ? UnescapePrivate<uint8_t>(isolate, string) |
| 501 | : UnescapePrivate<uc16>(isolate, string); |
Ben Murdoch | c561043 | 2016-08-08 18:44:38 +0100 | [diff] [blame] | 502 | } |
| 503 | |
| 504 | } // namespace internal |
| 505 | } // namespace v8 |