OLD | NEW |
---|---|
1 // Copyright (c) 2012, the Dart project authors. Please see the AUTHORS file | 1 // Copyright (c) 2012, the Dart project authors. Please see the AUTHORS file |
2 // for details. All rights reserved. Use of this source code is governed by a | 2 // for details. All rights reserved. Use of this source code is governed by a |
3 // BSD-style license that can be found in the LICENSE file. | 3 // BSD-style license that can be found in the LICENSE file. |
4 | 4 |
5 part of dart.core; | 5 part of dart.core; |
6 | 6 |
7 /** | 7 /** |
8 * A parsed URI, such as a URL. | 8 * A parsed URI, such as a URL. |
9 * | 9 * |
10 * **See also:** | 10 * **See also:** |
(...skipping 158 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
169 // | 169 // |
170 // pchar = unreserved / pct-encoded / sub-delims / ":" / "@" | 170 // pchar = unreserved / pct-encoded / sub-delims / ":" / "@" |
171 // | 171 // |
172 // query = *( pchar / "/" / "?" ) | 172 // query = *( pchar / "/" / "?" ) |
173 // | 173 // |
174 // fragment = *( pchar / "/" / "?" ) | 174 // fragment = *( pchar / "/" / "?" ) |
175 bool isRegName(int ch) { | 175 bool isRegName(int ch) { |
176 return ch < 128 && ((_regNameTable[ch >> 4] & (1 << (ch & 0x0f))) != 0); | 176 return ch < 128 && ((_regNameTable[ch >> 4] & (1 << (ch & 0x0f))) != 0); |
177 } | 177 } |
178 | 178 |
179 int ipV6Address(int index) { | 179 String scheme = ""; |
180 // IPv6. Skip to ']'. | 180 String path; |
181 index = uri.indexOf(']', index); | 181 String userinfo = ""; |
182 if (index == -1) { | 182 String host = ""; |
183 throw new FormatException("Bad end of IPv6 host"); | 183 int port = 0; |
184 } | 184 String query = ""; |
185 return index + 1; | 185 String fragment = ""; |
186 } | 186 |
187 | |
188 int length = uri.length; | |
189 int index = 0; | 187 int index = 0; |
190 | 188 int pathStart = 0; |
191 int schemeEndIndex = 0; | 189 int char = -1; |
192 | 190 |
193 if (length == 0) { | 191 void parseAuth() { |
194 return new Uri(); | 192 if (index == uri.length) { |
195 } | 193 char = -1; |
196 | 194 return; |
197 if (uri.codeUnitAt(0) != _SLASH) { | 195 } |
198 // Can be scheme. | 196 int authStart = index; |
199 while (index < length) { | 197 int lastColon = -1; |
200 // Look for ':'. If found, continue from the post of ':'. If not (end | 198 int lastAt = -1; |
201 // reached or invalid scheme char found) back up one char, and continue | 199 char = uri.codeUnitAt(index); |
202 // to path. | 200 authEnd: { |
203 // Note that scheme-chars is contained in path-chars. | 201 while (index < uri.length) { |
204 int codeUnit = uri.codeUnitAt(index++); | 202 char = uri.codeUnitAt(index); |
205 if (!_isSchemeCharacter(codeUnit)) { | 203 if (char == _SLASH || char == _QUESTION || char == _NUMBER_SIGN) { |
206 if (codeUnit == _COLON) { | 204 break authEnd; |
Anders Johnsen
2014/06/17 13:07:03
Consider alternative without authEnd block.
Lasse Reichstein Nielsen
2014/06/17 14:31:45
Considered. See suggestion.
| |
207 schemeEndIndex = index; | 205 } |
208 } else { | 206 if (char == _AT_SIGN) { |
209 // Back up one char, since we met an invalid scheme char. | 207 lastAt = index; |
210 index--; | 208 lastColon = -1; |
211 } | 209 } else if (char == _COLON) { |
212 break; | 210 lastColon = index; |
213 } | 211 } else if (char == _LEFT_BRACKET) { |
214 } | 212 lastColon = -1; |
215 } | 213 int endBracket = uri.indexOf(']', index + 1); |
216 | 214 if (endBracket == -1) { |
217 int userInfoEndIndex = -1; | 215 index = uri.length; |
218 int portIndex = -1; | |
219 int authorityEndIndex = schemeEndIndex; | |
220 // If we see '//', there must be an authority. | |
221 if (authorityEndIndex == index && | |
222 authorityEndIndex + 1 < length && | |
223 uri.codeUnitAt(authorityEndIndex) == _SLASH && | |
224 uri.codeUnitAt(authorityEndIndex + 1) == _SLASH) { | |
225 // Skip '//'. | |
226 authorityEndIndex += 2; | |
227 // It can both be host and userInfo. | |
228 while (authorityEndIndex < length) { | |
229 int codeUnit = uri.codeUnitAt(authorityEndIndex++); | |
230 if (!isRegName(codeUnit)) { | |
231 if (codeUnit == _LEFT_BRACKET) { | |
232 authorityEndIndex = ipV6Address(authorityEndIndex); | |
233 } else if (portIndex == -1 && codeUnit == _COLON) { | |
234 // First time ':'. | |
235 portIndex = authorityEndIndex; | |
236 } else if (codeUnit == _AT_SIGN || codeUnit == _COLON) { | |
237 // Second time ':' or first '@'. Must be userInfo. | |
238 userInfoEndIndex = uri.indexOf('@', authorityEndIndex - 1); | |
239 // Not found. Must be path then. | |
240 if (userInfoEndIndex == -1) { | |
241 authorityEndIndex = index; | |
242 break; | 216 break; |
217 } else { | |
218 index = endBracket; | |
243 } | 219 } |
244 portIndex = -1; | 220 } |
245 authorityEndIndex = userInfoEndIndex + 1; | 221 index++; |
246 // Now it can only be host:port. | 222 } |
247 while (authorityEndIndex < length) { | 223 char = -1; |
248 int codeUnit = uri.codeUnitAt(authorityEndIndex++); | 224 } |
249 if (!isRegName(codeUnit)) { | 225 int hostStart = authStart; |
250 if (codeUnit == _LEFT_BRACKET) { | 226 int hostEnd = index; |
251 authorityEndIndex = ipV6Address(authorityEndIndex); | 227 if (lastAt >= 0) { |
252 } else if (codeUnit == _COLON) { | 228 userinfo = _makeUserInfo(uri, authStart, lastAt); |
253 if (portIndex != -1) { | 229 hostStart = lastAt + 1; |
254 throw new FormatException("Double port in host"); | 230 } |
255 } | 231 if (lastColon >= 0) { |
256 portIndex = authorityEndIndex; | 232 if (lastColon + 1 == index) { |
257 } else { | 233 _fail(uri, index, "Invalid port number"); |
258 authorityEndIndex--; | 234 } |
259 break; | 235 int portNumber = 0; |
260 } | 236 for (int i = lastColon + 1; i < index; i++) { |
237 int digit = uri.codeUnitAt(i); | |
238 if (_ZERO > digit || _NINE < digit) { | |
239 _fail(uri, i, "Invalid port number"); | |
240 } | |
241 portNumber = portNumber * 10 + (digit - _ZERO); | |
242 } | |
243 port = _makePort(portNumber, scheme); | |
244 hostEnd = lastColon; | |
245 } | |
246 host = _makeHost(uri, hostStart, hostEnd); | |
247 if (index < uri.length) { | |
248 char = uri.codeUnitAt(index); | |
249 } | |
250 } | |
251 | |
252 endPath: { | |
253 toPath: { | |
Anders Johnsen
2014/06/17 13:07:03
Consider doing:
bool hasPath = true;
while(...) {
Lasse Reichstein Nielsen
2014/06/17 14:31:44
Ack, my precious flow control!!!
See alternative
| |
254 toAuthOpt: { | |
255 while (index < uri.length) { | |
256 char = uri.codeUnitAt(index); | |
257 if (char == _QUESTION || char == _NUMBER_SIGN) break endPath; | |
258 if (char == _SLASH) { | |
259 if (index == 0) break toAuthOpt; | |
260 break toPath; | |
261 } | |
262 if (char == _COLON) { | |
263 if (index == 0) _fail(uri, 0, "Invalid empty scheme"); | |
264 scheme = _makeScheme(uri, 0, index); | |
265 index++; | |
266 pathStart = index; | |
267 if (index == uri.length) { | |
268 char = -1; | |
269 break endPath; | |
261 } | 270 } |
271 char = uri.codeUnitAt(index); | |
272 if (char == _QUESTION || char == _NUMBER_SIGN) break endPath; | |
273 if (char == _SLASH) break toAuthOpt; | |
274 break toPath; | |
262 } | 275 } |
263 break; | 276 index++; |
264 } else { | 277 } |
265 authorityEndIndex--; | 278 char = -1; |
266 break; | 279 break endPath; |
267 } | 280 } // toAuthOpt. |
268 } | 281 // Have seen one slash either at start or right after scheme. |
269 } | 282 // If two slashes, it's an authority, otherwise it's just the path. |
283 assert(char == _SLASH); | |
284 index++; | |
285 if (index == uri.length) { | |
286 char = -1; | |
287 break endPath; | |
288 } | |
289 char = uri.codeUnitAt(index); | |
290 if (char == _QUESTION || char == _NUMBER_SIGN) break endPath; | |
291 if (char == _SLASH) { | |
292 index++; | |
293 parseAuth(); | |
294 pathStart = index; | |
295 if (char == _QUESTION || char == _NUMBER_SIGN) break endPath; | |
296 } else { | |
297 break toPath; | |
298 } | |
299 } // toPath | |
300 | |
301 // Characters from pathStart to index (inclusive) are known | |
302 // to be part of the path. | |
303 if (index < uri.length) { | |
304 while (++index < uri.length) { | |
305 char = uri.codeUnitAt(index); | |
306 if (char== _QUESTION || char == _NUMBER_SIGN) break endPath; | |
Anders Johnsen
2014/06/17 13:07:03
Add missing space.
Lasse Reichstein Nielsen
2014/06/17 14:31:45
Done.
| |
307 } | |
308 char = -1; | |
309 } | |
310 } // endPath | |
311 bool ensureLeadingSlash = (host != "" || scheme == "file"); | |
312 path = _makePath(uri, pathStart, index, null, ensureLeadingSlash); | |
313 | |
314 if (char == _QUESTION) { | |
315 int numberSignIndex = uri.indexOf('#', index + 1); | |
316 if (numberSignIndex < 0) { | |
317 query = _makeQuery(uri, index + 1, uri.length, null); | |
318 } else { | |
319 query = _makeQuery(uri, index + 1, numberSignIndex, null); | |
320 fragment = _makeFragment(uri, numberSignIndex + 1, uri.length); | |
321 } | |
322 } else if (char == _NUMBER_SIGN) { | |
323 fragment = _makeFragment(uri, index + 1, uri.length); | |
324 } | |
325 return new Uri._internal(scheme, | |
326 userinfo, | |
327 host, | |
328 port, | |
329 path, | |
330 query, | |
331 fragment); | |
332 } | |
333 | |
334 // Report a parse failure. | |
335 static void _fail(String uri, int index, String message) { | |
336 // TODO(lrn): Consider adding this to FormatException. | |
337 if (index == uri.length) { | |
338 message += " at end of input."; | |
270 } else { | 339 } else { |
271 authorityEndIndex = schemeEndIndex; | 340 message += " at position $index.\n"; |
272 } | 341 // Pick a slice of uri containing index and, if |
273 | 342 // necessary, truncate the ends to ensure the entire |
274 // At path now. | 343 // slice fits on one line. |
275 int pathEndIndex = authorityEndIndex; | 344 int min = 0; |
276 while (pathEndIndex < length) { | 345 int max = uri.length; |
277 int codeUnit = uri.codeUnitAt(pathEndIndex++); | 346 String pre = ""; |
278 if (codeUnit == _QUESTION || codeUnit == _NUMBER_SIGN) { | 347 String post = ""; |
279 pathEndIndex--; | 348 if (uri.length > 78) { |
280 break; | 349 min = index - 10; |
281 } | 350 if (min < 0) min = 0; |
282 } | 351 int max = min + 72; |
283 | 352 if (max > uri.length) { |
284 // Maybe query. | 353 max = uri.length; |
285 int queryEndIndex = pathEndIndex; | 354 min = max - 72; |
286 if (queryEndIndex < length && uri.codeUnitAt(queryEndIndex) == _QUESTION) { | 355 } |
287 while (queryEndIndex < length) { | 356 if (min != 0) pre = "..."; |
288 int codeUnit = uri.codeUnitAt(queryEndIndex++); | 357 if (max != uri.length) post = "..."; |
289 if (codeUnit == _NUMBER_SIGN) { | 358 } |
290 queryEndIndex--; | 359 // Combine message, slice and a caret pointing to the error index. |
291 break; | 360 message = "$message$pre${uri.substring(min, max)}$post\n" |
292 } | 361 "${' ' * (pre.length + index - min)}^"; |
293 } | 362 } |
294 } | 363 throw new FormatException(message); |
295 | |
296 var scheme = null; | |
297 if (schemeEndIndex > 0) { | |
298 scheme = uri.substring(0, schemeEndIndex - 1); | |
299 } | |
300 | |
301 var host = ""; | |
302 var userInfo = ""; | |
303 var port = 0; | |
304 if (schemeEndIndex != authorityEndIndex) { | |
305 int startIndex = schemeEndIndex + 2; | |
306 if (userInfoEndIndex > 0) { | |
307 userInfo = uri.substring(startIndex, userInfoEndIndex); | |
308 startIndex = userInfoEndIndex + 1; | |
309 } | |
310 if (portIndex > 0) { | |
311 var portStr = uri.substring(portIndex, authorityEndIndex); | |
312 try { | |
313 port = int.parse(portStr); | |
314 } catch (_) { | |
315 throw new FormatException("Invalid port: '$portStr'"); | |
316 } | |
317 host = uri.substring(startIndex, portIndex - 1); | |
318 } else { | |
319 host = uri.substring(startIndex, authorityEndIndex); | |
320 } | |
321 } | |
322 | |
323 var path = uri.substring(authorityEndIndex, pathEndIndex); | |
324 var query = ""; | |
325 if (pathEndIndex < queryEndIndex) { | |
326 query = uri.substring(pathEndIndex + 1, queryEndIndex); | |
327 } | |
328 var fragment = ""; | |
329 // If queryEndIndex is not at end (length), there is a fragment. | |
330 if (queryEndIndex < length) { | |
331 fragment = uri.substring(queryEndIndex + 1, length); | |
332 } | |
333 | |
334 return new Uri(scheme: scheme, | |
335 userInfo: userInfo, | |
336 host: host, | |
337 port: port, | |
338 path: path, | |
339 query: query, | |
340 fragment: fragment); | |
341 } | 364 } |
342 | 365 |
366 /// Internal non-verifying constructor. Only call with validated arguments. | |
367 Uri._internal(this.scheme, | |
368 this.userInfo, | |
369 this._host, | |
370 this._port, | |
371 this._path, | |
372 this.query, | |
373 this.fragment); | |
374 | |
343 /** | 375 /** |
344 * Creates a new URI from its components. | 376 * Creates a new URI from its components. |
345 * | 377 * |
346 * Each component is set through a named argument. Any number of | 378 * Each component is set through a named argument. Any number of |
347 * components can be provided. The default value for the components | 379 * components can be provided. The default value for the components |
348 * not provided is the empry string, except for [port] which has a | 380 * not provided is the empry string, except for [port] which has a |
349 * default value of 0. The [path] and [query] components can be set | 381 * default value of 0. The [path] and [query] components can be set |
350 * using two different named arguments. | 382 * using two different named arguments. |
351 * | 383 * |
352 * The scheme component is set through [scheme]. The scheme is | 384 * The scheme component is set through [scheme]. The scheme is |
(...skipping 26 matching lines...) Expand all Loading... | |
379 * [queryParameters]. When [query] is used the provided string is | 411 * [queryParameters]. When [query] is used the provided string is |
380 * expected to be fully percent-encoded and is used in its literal | 412 * expected to be fully percent-encoded and is used in its literal |
381 * form. When [queryParameters] is used the query is built from the | 413 * form. When [queryParameters] is used the query is built from the |
382 * provided map. Each key and value in the map is percent-encoded | 414 * provided map. Each key and value in the map is percent-encoded |
383 * and joined using equal and ampersand characters. The | 415 * and joined using equal and ampersand characters. The |
384 * percent-encoding of the keys and values encodes all characters | 416 * percent-encoding of the keys and values encodes all characters |
385 * except for the unreserved characters. | 417 * except for the unreserved characters. |
386 * | 418 * |
387 * The fragment component is set through [fragment]. | 419 * The fragment component is set through [fragment]. |
388 */ | 420 */ |
389 Uri({String scheme, | 421 factory Uri({String scheme, |
390 this.userInfo: "", | 422 String userInfo: "", |
391 String host: "", | 423 String host: "", |
392 port: 0, | 424 port: 0, |
393 String path, | 425 String path, |
394 Iterable<String> pathSegments, | 426 Iterable<String> pathSegments, |
395 String query, | 427 String query, |
396 Map<String, String> queryParameters, | 428 Map<String, String> queryParameters, |
397 fragment: ""}) : | 429 fragment: ""}) { |
398 scheme = _makeScheme(scheme), | 430 scheme = _makeScheme(scheme, 0, _stringOrNullLength(scheme)); |
399 _host = _makeHost(host), | 431 userInfo = _makeUserInfo(userInfo, 0, _stringOrNullLength(userInfo)); |
400 query = _makeQuery(query, queryParameters), | 432 host = _makeHost(host, 0, _stringOrNullLength(host)); |
401 fragment = _makeFragment(fragment) { | 433 query = _makeQuery(query, 0, _stringOrNullLength(query), queryParameters); |
402 // Perform scheme specific normalization. | 434 fragment = _makeFragment(fragment, 0, _stringOrNullLength(fragment)); |
403 if (scheme == "http" && port == 80) { | 435 port = _makePort(port, scheme); |
404 _port = 0; | 436 bool ensureLeadingSlash = (host != "" || scheme == "file"); |
405 } else if (scheme == "https" && port == 443) { | 437 path = _makePath(path, 0, _stringOrNullLength(path), pathSegments, |
406 _port = 0; | 438 ensureLeadingSlash); |
407 } else { | 439 |
408 _port = port; | 440 return new Uri._internal(scheme, userInfo, host, port, |
409 } | 441 path, query, fragment); |
410 // Fill the path. | |
411 _path = _makePath(path, pathSegments); | |
412 } | 442 } |
413 | 443 |
414 /** | 444 /** |
415 * Creates a new `http` URI from authority, path and query. | 445 * Creates a new `http` URI from authority, path and query. |
416 * | 446 * |
417 * Examples: | 447 * Examples: |
418 * | 448 * |
419 * ``` | 449 * ``` |
420 * // http://example.org/path?q=dart. | 450 * // http://example.org/path?q=dart. |
421 * new Uri.http("google.com", "/search", { "q" : "dart" }); | 451 * new Uri.http("google.com", "/search", { "q" : "dart" }); |
(...skipping 60 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
482 var hostEnd = hostStart; | 512 var hostEnd = hostStart; |
483 if (hostStart < authority.length && | 513 if (hostStart < authority.length && |
484 authority.codeUnitAt(hostStart) == _LEFT_BRACKET) { | 514 authority.codeUnitAt(hostStart) == _LEFT_BRACKET) { |
485 // IPv6 host. | 515 // IPv6 host. |
486 for (; hostEnd < authority.length; hostEnd++) { | 516 for (; hostEnd < authority.length; hostEnd++) { |
487 if (authority.codeUnitAt(hostEnd) == _RIGHT_BRACKET) break; | 517 if (authority.codeUnitAt(hostEnd) == _RIGHT_BRACKET) break; |
488 } | 518 } |
489 if (hostEnd == authority.length) { | 519 if (hostEnd == authority.length) { |
490 throw new FormatException("Invalid IPv6 host entry."); | 520 throw new FormatException("Invalid IPv6 host entry."); |
491 } | 521 } |
492 parseIPv6Address(authority.substring(hostStart + 1, hostEnd)); | 522 parseIPv6Address(authority, hostStart + 1, hostEnd); |
493 hostEnd++; // Skip the closing bracket. | 523 hostEnd++; // Skip the closing bracket. |
494 if (hostEnd != authority.length && | 524 if (hostEnd != authority.length && |
495 authority.codeUnitAt(hostEnd) != _COLON) { | 525 authority.codeUnitAt(hostEnd) != _COLON) { |
496 throw new FormatException("Invalid end of authority"); | 526 throw new FormatException("Invalid end of authority"); |
497 } | 527 } |
498 } | 528 } |
499 // Split host and port. | 529 // Split host and port. |
500 bool hasPort = false; | 530 bool hasPort = false; |
501 for (; hostEnd < authority.length; hostEnd++) { | 531 for (; hostEnd < authority.length; hostEnd++) { |
502 if (authority.codeUnitAt(hostEnd) == _COLON) { | 532 if (authority.codeUnitAt(hostEnd) == _COLON) { |
(...skipping 255 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
758 * The returned map is unmodifiable and will throw [UnsupportedError] on any | 788 * The returned map is unmodifiable and will throw [UnsupportedError] on any |
759 * calls that would mutate it. | 789 * calls that would mutate it. |
760 */ | 790 */ |
761 Map<String, String> get queryParameters { | 791 Map<String, String> get queryParameters { |
762 if (_queryParameters == null) { | 792 if (_queryParameters == null) { |
763 _queryParameters = new UnmodifiableMapView(splitQueryString(query)); | 793 _queryParameters = new UnmodifiableMapView(splitQueryString(query)); |
764 } | 794 } |
765 return _queryParameters; | 795 return _queryParameters; |
766 } | 796 } |
767 | 797 |
768 static String _makeHost(String host) { | 798 static int _makePort(int port, String scheme) { |
769 if (host == null || host.isEmpty) return host; | 799 // Perform scheme specific normalization. |
770 if (host.codeUnitAt(0) == _LEFT_BRACKET) { | 800 if (port == 80 && scheme == "http") { |
771 if (host.codeUnitAt(host.length - 1) != _RIGHT_BRACKET) { | 801 return 0; |
772 throw new FormatException('Missing end `]` to match `[` in host'); | 802 } |
803 if (port == 443 && scheme == "https") { | |
804 return 0; | |
805 } | |
806 return port; | |
807 } | |
808 | |
809 static String _makeHost(String host, int start, int end) { | |
810 if (host == null) return null; | |
811 if (start == end) return ""; | |
812 // Host is an IPv6 address if it starts with '[' or contains a colon. | |
813 if (host.codeUnitAt(start) == _LEFT_BRACKET) { | |
814 if (host.codeUnitAt(end - 1) != _RIGHT_BRACKET) { | |
815 _fail(host, start, 'Missing end `]` to match `[` in host'); | |
773 } | 816 } |
774 parseIPv6Address(host.substring(1, host.length - 1)); | 817 parseIPv6Address(host, start + 1, end - 1); |
775 return host; | 818 return host.substring(start, end); |
776 } | 819 } |
777 for (int i = 0; i < host.length; i++) { | 820 // TODO(lrn): skip if too short to be a valid IPv6 address. |
821 for (int i = start; i < end; i++) { | |
778 if (host.codeUnitAt(i) == _COLON) { | 822 if (host.codeUnitAt(i) == _COLON) { |
779 parseIPv6Address(host); | 823 parseIPv6Address(host, start, end); |
780 return '[$host]'; | 824 return '[$host]'; |
781 } | 825 } |
782 } | 826 } |
783 return host; | 827 return _normalizeRegName(host, start, end); |
784 } | 828 } |
785 | 829 |
786 static String _makeScheme(String scheme) { | 830 static bool _isRegNameChar(int char) { |
787 bool isSchemeLowerCharacter(int ch) { | 831 return char < 127 && (_regNameTable[char >> 4] & (1 << (char & 0xf))) != 0; |
788 return ch < 128 && | 832 } |
789 ((_schemeLowerTable[ch >> 4] & (1 << (ch & 0x0f))) != 0); | |
790 } | |
791 | 833 |
792 if (scheme == null) return ""; | 834 /** |
793 bool allLowercase = true; | 835 * Validates and does case- and percent-encoding normalization. |
794 int length = scheme.length; | 836 * |
795 for (int i = 0; i < length; i++) { | 837 * The [host] must be an RFC3986 "reg-name". It is converted |
796 int codeUnit = scheme.codeUnitAt(i); | 838 * to lower case, and percent escapes are converted to either |
797 if (i == 0 && !_isAlphabeticCharacter(codeUnit)) { | 839 * lower case unreserved characters or upper case escapes. |
798 // First code unit must be an alphabetic character. | 840 */ |
799 throw new ArgumentError('Illegal scheme: $scheme'); | 841 static String _normalizeRegName(String host, int start, int end) { |
800 } | 842 StringBuffer buffer; |
801 if (!isSchemeLowerCharacter(codeUnit)) { | 843 int sectionStart = start; |
802 if (_isSchemeCharacter(codeUnit)) { | 844 int index = start; |
803 allLowercase = false; | 845 // Whether all characters between sectionStart and index are normalized, |
804 } else { | 846 bool isNormalized = true; |
805 throw new ArgumentError('Illegal scheme: $scheme'); | 847 |
848 while (index < end) { | |
849 int char = host.codeUnitAt(index); | |
850 if (char == _PERCENT) { | |
851 // The _regNameTable contains "%", so we check that first. | |
852 String replacement = _normalizeEscape(host, index, true); | |
853 if (replacement == null && isNormalized) { | |
854 index += 3; | |
855 continue; | |
806 } | 856 } |
857 if (buffer == null) buffer = new StringBuffer(); | |
858 String slice = host.substring(sectionStart, index); | |
859 if (!isNormalized) slice = slice.toLowerCase(); | |
860 buffer.write(slice); | |
861 int sourceLength = 3; | |
862 if (replacement == null) { | |
863 replacement = host.substring(index, index + 3); | |
864 } else if (replacement == "%") { | |
865 replacement = "%25"; | |
866 sourceLength = 1; | |
867 } | |
868 buffer.write(replacement); | |
869 index += sourceLength; | |
870 sectionStart = index; | |
871 isNormalized = true; | |
872 } else if (_isRegNameChar(char)) { | |
873 if (isNormalized && _UPPER_CASE_A <= char && _UPPER_CASE_Z >= char) { | |
874 // Put initial slice in buffer and continue in non-normalized mode | |
875 if (buffer == null) buffer = new StringBuffer(); | |
876 if (sectionStart < index) { | |
877 buffer.write(host.substring(sectionStart, index)); | |
878 sectionStart = index; | |
879 } | |
880 isNormalized = false; | |
881 } | |
882 index++; | |
883 } else if (_isGeneralDelimiter(char)) { | |
884 _fail(host, index, "Invalid character"); | |
885 } else { | |
886 int sourceLength = 1; | |
887 if ((char & 0xFC00) == 0xD800 && (index + 1) < end) { | |
888 int tail = host.codeUnitAt(index + 1); | |
889 if ((tail & 0xFC00) == 0xDC00) { | |
890 char = 0x10000 | ((char & 0x3ff) << 10) | (tail & 0x3ff); | |
891 sourceLength = 2; | |
892 } | |
893 } | |
894 if (buffer == null) buffer = new StringBuffer(); | |
895 String slice = host.substring(sectionStart, index); | |
896 if (!isNormalized) slice = slice.toLowerCase(); | |
897 buffer.write(slice); | |
898 buffer.write(_escapeChar(char)); | |
899 index += sourceLength; | |
900 sectionStart = index; | |
807 } | 901 } |
808 } | 902 } |
809 | 903 if (buffer == null) return host.substring(start, end); |
810 return allLowercase ? scheme : scheme.toLowerCase(); | 904 if (sectionStart < end) { |
905 String slice = host.substring(sectionStart, end); | |
906 if (!isNormalized) slice = slice.toLowerCase(); | |
907 buffer.write(slice); | |
908 } | |
909 return buffer.toString(); | |
811 } | 910 } |
812 | 911 |
813 String _makePath(String path, Iterable<String> pathSegments) { | 912 /** |
913 * Validates scheme characters and does case-normalization. | |
914 * | |
915 * Schemes are converted to lower case. They cannot contain escapes. | |
916 */ | |
917 static String _makeScheme(String scheme, int start, int end) { | |
Anders Johnsen
2014/06/17 13:07:03
start is redundant, as it's always 0.
Lasse Reichstein Nielsen
2014/06/17 14:31:45
True. And I know.
I have a dream of making Uri.pa
| |
918 if (start == end) return ""; | |
919 int char = scheme.codeUnitAt(start); | |
920 if (!_isAlphabeticCharacter(char)) { | |
921 _fail(scheme, start, "Scheme not starting with alphabetic character"); | |
922 } | |
923 bool allLowercase = char >= _LOWER_CASE_A; | |
924 for (int i = start; i < end; i++) { | |
925 int codeUnit = scheme.codeUnitAt(i); | |
926 if (!_isSchemeCharacter(codeUnit)) { | |
927 _fail(scheme, i, "Illegal scheme character"); | |
928 } | |
929 if (_LOWER_CASE_A <= char && _LOWER_CASE_Z >= char) { | |
930 allLowercase = false; | |
931 } | |
932 } | |
933 scheme = scheme.substring(start, end); | |
934 if (!allLowercase) scheme = scheme.toLowerCase(); | |
935 return scheme; | |
936 } | |
937 | |
938 static String _makeUserInfo(String userInfo, int start, int end) { | |
939 if (userInfo == null) return "null"; | |
940 return _normalize(userInfo, start, end, _userinfoTable); | |
941 } | |
942 | |
943 static String _makePath(String path, int start, int end, | |
944 Iterable<String> pathSegments, | |
945 bool ensureLeadingSlash) { | |
814 if (path == null && pathSegments == null) return ""; | 946 if (path == null && pathSegments == null) return ""; |
815 if (path != null && pathSegments != null) { | 947 if (path != null && pathSegments != null) { |
816 throw new ArgumentError('Both path and pathSegments specified'); | 948 throw new ArgumentError('Both path and pathSegments specified'); |
817 } | 949 } |
818 var result; | 950 var result; |
819 if (path != null) { | 951 if (path != null) { |
820 result = _normalize(path); | 952 result = _normalize(path, start, end, _pathCharOrSlashTable); |
821 } else { | 953 } else { |
822 result = pathSegments.map((s) => _uriEncode(_pathCharTable, s)).join("/"); | 954 result = pathSegments.map((s) => _uriEncode(_pathCharTable, s)).join("/"); |
823 } | 955 } |
824 if ((hasAuthority || (scheme == "file")) && | 956 if (ensureLeadingSlash && result.isNotEmpty && !result.startsWith("/")) { |
825 result.isNotEmpty && !result.startsWith("/")) { | |
826 return "/$result"; | 957 return "/$result"; |
827 } | 958 } |
828 return result; | 959 return result; |
829 } | 960 } |
830 | 961 |
831 static String _makeQuery(String query, Map<String, String> queryParameters) { | 962 static String _makeQuery(String query, int start, int end, |
963 Map<String, String> queryParameters) { | |
832 if (query == null && queryParameters == null) return ""; | 964 if (query == null && queryParameters == null) return ""; |
833 if (query != null && queryParameters != null) { | 965 if (query != null && queryParameters != null) { |
834 throw new ArgumentError('Both query and queryParameters specified'); | 966 throw new ArgumentError('Both query and queryParameters specified'); |
835 } | 967 } |
836 if (query != null) return _normalize(query); | 968 if (query != null) return _normalize(query, start, end, _queryCharTable); |
837 | 969 |
838 var result = new StringBuffer(); | 970 var result = new StringBuffer(); |
839 var first = true; | 971 var first = true; |
840 queryParameters.forEach((key, value) { | 972 queryParameters.forEach((key, value) { |
841 if (!first) { | 973 if (!first) { |
842 result.write("&"); | 974 result.write("&"); |
843 } | 975 } |
844 first = false; | 976 first = false; |
845 result.write(Uri.encodeQueryComponent(key)); | 977 result.write(Uri.encodeQueryComponent(key)); |
846 if (value != null && !value.isEmpty) { | 978 if (value != null && !value.isEmpty) { |
847 result.write("="); | 979 result.write("="); |
848 result.write(Uri.encodeQueryComponent(value)); | 980 result.write(Uri.encodeQueryComponent(value)); |
849 } | 981 } |
850 }); | 982 }); |
851 return result.toString(); | 983 return result.toString(); |
852 } | 984 } |
853 | 985 |
854 static String _makeFragment(String fragment) { | 986 static String _makeFragment(String fragment, int start, int end) { |
855 if (fragment == null) return ""; | 987 if (fragment == null) return ""; |
856 return _normalize(fragment); | 988 return _normalize(fragment, start, end, _queryCharTable); |
857 } | 989 } |
858 | 990 |
859 static String _normalize(String component) { | 991 static int _stringOrNullLength(String s) => (s == null) ? 0 : s.length; |
860 int index = component.indexOf('%'); | |
861 if (index < 0) return component; | |
862 | 992 |
863 bool isNormalizedHexDigit(int digit) { | 993 static bool _isHexDigit(int char) { |
864 return (_ZERO <= digit && digit <= _NINE) || | 994 if (_NINE >= char) return _ZERO <= char; |
865 (_UPPER_CASE_A <= digit && digit <= _UPPER_CASE_F); | 995 char |= 0x20; |
996 return _LOWER_CASE_A <= char && _LOWER_CASE_F >= char; | |
997 } | |
998 | |
999 static int _hexValue(int char) { | |
1000 assert(_isHexDigit(char)); | |
1001 if (_NINE >= char) return char - _ZERO; | |
1002 char |= 0x20; | |
1003 return char - (_LOWER_CASE_A - 10); | |
1004 } | |
1005 | |
1006 /** | |
1007 * Performs RFC 3986 Percent-Encoding Normalization. | |
1008 * | |
1009 * Returns a replacement string that should be replace the original escape. | |
1010 * Returns null if no replacement is necessary because the escape is | |
1011 * not for an unreserved character and is already non-lower-case. | |
1012 * | |
1013 * Returns "%" if the escape is invalid (not two valid hex digits following | |
1014 * the percent sign). The calling code should replace the percent | |
1015 * sign with "%25", but leave the following two characters unmodified. | |
1016 * | |
1017 * If [lowerCase] is true, a single character returned is always lower case, | |
1018 */ | |
1019 static String _normalizeEscape(String source, int index, bool lowerCase) { | |
1020 assert(source.codeUnitAt(index) == _PERCENT); | |
1021 if (index + 2 >= source.length) { | |
1022 return "%"; // Marks the escape as invalid. | |
866 } | 1023 } |
1024 int firstDigit = source.codeUnitAt(index + 1); | |
1025 int secondDigit = source.codeUnitAt(index + 2); | |
1026 if (!_isHexDigit(firstDigit) || !_isHexDigit(secondDigit)) { | |
1027 return "%"; // Marks the escape as invalid. | |
1028 } | |
1029 int value = _hexValue(firstDigit) * 16 + _hexValue(secondDigit); | |
1030 if (_isUnreservedChar(value)) { | |
1031 if (lowerCase && _UPPER_CASE_A <= value && _UPPER_CASE_Z >= value) { | |
1032 value |= 0x20; | |
1033 } | |
1034 return new String.fromCharCode(value); | |
1035 } | |
1036 if (firstDigit >= _LOWER_CASE_A || secondDigit >= _LOWER_CASE_A) { | |
1037 // Either digit is lower case. | |
1038 return source.substring(index, index + 3).toUpperCase(); | |
1039 } | |
1040 // Escape is retained, and is already non-lower case, so return null to | |
1041 // represent "no replacement necessary". | |
1042 return null; | |
1043 } | |
867 | 1044 |
868 bool isLowerCaseHexDigit(int digit) { | 1045 static bool _isUnreservedChar(int ch) { |
869 return _LOWER_CASE_A <= digit && digit <= _LOWER_CASE_F; | 1046 return ch < 127 && |
870 } | 1047 ((_unreservedTable[ch >> 4] & (1 << (ch & 0x0f))) != 0); |
1048 } | |
871 | 1049 |
872 bool isUnreserved(int ch) { | 1050 static String _escapeChar(char) { |
873 return ch < 128 && | 1051 const hexDigits = "0123456789ABCDEF"; |
874 ((_unreservedTable[ch >> 4] & (1 << (ch & 0x0f))) != 0); | 1052 List codeUnits; |
875 } | 1053 if (char < 0x80) { |
876 | 1054 // ASCII, a single percent encoded sequence. |
877 int normalizeHexDigit(int index) { | 1055 codeUnits = new List(3); |
878 var codeUnit = component.codeUnitAt(index); | 1056 codeUnits[0] = _PERCENT; |
879 if (isLowerCaseHexDigit(codeUnit)) { | 1057 codeUnits[1] = hexDigits.codeUnitAt(char >> 4); |
880 return codeUnit - 0x20; | 1058 codeUnits[2] = hexDigits.codeUnitAt(char & 0xf); |
881 } else if (!isNormalizedHexDigit(codeUnit)) { | 1059 } else { |
882 throw new ArgumentError("Invalid URI component: $component"); | 1060 // Do UTF-8 encoding of character, then percent encode bytes. |
883 } else { | 1061 int flag = 0xc0; // The high-bit markers on the first byte of UTF-8. |
884 return codeUnit; | 1062 int encodedBytes = 2; |
1063 if (char > 0x7ff) { | |
1064 flag = 0xe0; | |
1065 encodedBytes = 3; | |
1066 if (char > 0xffff) { | |
1067 encodedBytes = 4; | |
1068 flag = 0xf0; | |
1069 } | |
1070 } | |
1071 codeUnits = new List(3 * encodedBytes); | |
1072 int index = 0; | |
1073 while (--encodedBytes >= 0) { | |
1074 int byte = ((char >> (6 * encodedBytes)) & 0x3f) | flag; | |
1075 codeUnits[index] = _PERCENT; | |
1076 codeUnits[index + 1] = hexDigits.codeUnitAt(byte >> 4); | |
1077 codeUnits[index + 2] = hexDigits.codeUnitAt(byte & 0xf); | |
1078 index += 3; | |
1079 flag = 0x80; // Following bytes have only high bit set. | |
885 } | 1080 } |
886 } | 1081 } |
1082 return new String.fromCharCodes(codeUnits); | |
1083 } | |
887 | 1084 |
888 int decodeHexDigitPair(int index) { | 1085 /** |
889 int byte = 0; | 1086 * Runs through component checking that each character is valid and |
890 for (int i = 0; i < 2; i++) { | 1087 * normalize percent escapes. |
891 var codeUnit = component.codeUnitAt(index + i); | 1088 * |
892 if (_ZERO <= codeUnit && codeUnit <= _NINE) { | 1089 * Uses [charTable] to check if a non-`%` character is allowed. |
893 byte = byte * 16 + codeUnit - _ZERO; | 1090 * Each `%` character must be followed by two hex digits. |
1091 * If the hex-digits are lower case letters, they are converted to | |
1092 * upper case. | |
1093 */ | |
1094 static String _normalize(String component, int start, int end, | |
1095 List<int> charTable) { | |
1096 StringBuffer buffer; | |
1097 int sectionStart = start; | |
1098 int index = start; | |
1099 // Loop while characters are valid and escapes correct and upper-case. | |
1100 while (index < end) { | |
1101 int char = component.codeUnitAt(index); | |
1102 if (char < 127 && (charTable[char >> 4] & (1 << (char & 0x0f))) != 0) { | |
1103 index++; | |
1104 } else { | |
1105 String replacement; | |
1106 int sourceLength; | |
1107 if (char == _PERCENT) { | |
1108 replacement = _normalizeEscape(component, index, false); | |
1109 // Returns null if we should keep the existing escape. | |
1110 if (replacement == null) { | |
1111 index += 3; | |
1112 continue; | |
1113 } | |
1114 // Returns "%" if we should escape the existing percent. | |
1115 if ("%" == replacement) { | |
1116 replacement = "%25"; | |
1117 sourceLength = 1; | |
1118 } else { | |
1119 sourceLength = 3; | |
1120 } | |
1121 } else if (_isGeneralDelimiter(char)) { | |
1122 _fail(component, index, "Invalid character"); | |
894 } else { | 1123 } else { |
895 // Check ranges A-F (0x41-0x46) and a-f (0x61-0x66). | 1124 sourceLength = 1; |
896 codeUnit |= 0x20; | 1125 if ((char & 0xFC00) == 0xD800) { |
897 if (_LOWER_CASE_A <= codeUnit && | 1126 // Possible lead surrogate. |
898 codeUnit <= _LOWER_CASE_F) { | 1127 if (index + 1 < end) { |
899 byte = byte * 16 + codeUnit - _LOWER_CASE_A + 10; | 1128 int tail = component.codeUnitAt(index + 1); |
900 } else { | 1129 if ((tail & 0xFC00) == 0xDC00) { |
901 throw new ArgumentError( | 1130 // Tail surrogat. |
902 "Invalid percent-encoding in URI component: $component"); | 1131 sourceLength = 2; |
1132 char = 0x10000 | ((char & 0x3ff) << 10) | (tail & 0x3ff); | |
1133 } | |
1134 } | |
903 } | 1135 } |
1136 replacement = _escapeChar(char); | |
904 } | 1137 } |
905 } | 1138 if (buffer == null) buffer = new StringBuffer(); |
906 return byte; | 1139 buffer.write(component.substring(sectionStart, index)); |
907 } | 1140 buffer.write(replacement); |
908 | 1141 index += sourceLength; |
909 // Start building the normalized component string. | 1142 sectionStart = index; |
910 StringBuffer result; | |
911 int length = component.length; | |
912 int prevIndex = 0; | |
913 | |
914 // Copy a part of the component string to the result. | |
915 void fillResult() { | |
916 if (result == null) { | |
917 assert(prevIndex == 0); | |
918 result = new StringBuffer(component.substring(prevIndex, index)); | |
919 } else { | |
920 result.write(component.substring(prevIndex, index)); | |
921 } | 1143 } |
922 } | 1144 } |
923 | 1145 if (buffer == null) { |
924 while (index < length) { | 1146 // Makes no copy if start == 0 and end == component.length. |
925 // Normalize percent-encoding to uppercase and don't encode | 1147 return component.substring(start, end); |
926 // unreserved characters. | |
927 assert(component.codeUnitAt(index) == _PERCENT); | |
928 if (length < index + 2) { | |
929 throw new ArgumentError( | |
930 "Invalid percent-encoding in URI component: $component"); | |
931 } | |
932 | |
933 var codeUnit1 = component.codeUnitAt(index + 1); | |
934 var codeUnit2 = component.codeUnitAt(index + 2); | |
935 var decodedCodeUnit = decodeHexDigitPair(index + 1); | |
936 if (isNormalizedHexDigit(codeUnit1) && | |
937 isNormalizedHexDigit(codeUnit2) && | |
938 !isUnreserved(decodedCodeUnit)) { | |
939 index += 3; | |
940 } else { | |
941 fillResult(); | |
942 if (isUnreserved(decodedCodeUnit)) { | |
943 result.writeCharCode(decodedCodeUnit); | |
944 } else { | |
945 result.write("%"); | |
946 result.writeCharCode(normalizeHexDigit(index + 1)); | |
947 result.writeCharCode(normalizeHexDigit(index + 2)); | |
948 } | |
949 index += 3; | |
950 prevIndex = index; | |
951 } | |
952 int next = component.indexOf('%', index); | |
953 if (next >= index) { | |
954 index = next; | |
955 } else { | |
956 index = length; | |
957 } | |
958 } | 1148 } |
959 if (result == null) return component; | 1149 if (sectionStart < end) { |
960 | 1150 buffer.write(component.substring(sectionStart, end)); |
961 if (result != null && prevIndex != index) fillResult(); | 1151 } |
962 assert(index == length); | 1152 return buffer.toString(); |
963 | |
964 return result.toString(); | |
965 } | 1153 } |
966 | 1154 |
967 static bool _isSchemeCharacter(int ch) { | 1155 static bool _isSchemeCharacter(int ch) { |
968 return ch < 128 && ((_schemeTable[ch >> 4] & (1 << (ch & 0x0f))) != 0); | 1156 return ch < 128 && ((_schemeTable[ch >> 4] & (1 << (ch & 0x0f))) != 0); |
969 } | 1157 } |
970 | 1158 |
1159 static bool _isGeneralDelimiter(int ch) { | |
1160 return ch <= 64 && | |
1161 ((_genDelimitersTable[ch >> 4] & (1 << (ch & 0x0f))) != 0); | |
1162 } | |
971 | 1163 |
972 /** | 1164 /** |
973 * Returns whether the URI is absolute. | 1165 * Returns whether the URI is absolute. |
974 */ | 1166 */ |
975 bool get isAbsolute => scheme != "" && fragment == ""; | 1167 bool get isAbsolute => scheme != "" && fragment == ""; |
976 | 1168 |
977 String _merge(String base, String reference) { | 1169 String _merge(String base, String reference) { |
978 if (base == "") return "/$reference"; | 1170 if (base == "") return "/$reference"; |
979 return "${base.substring(0, base.lastIndexOf("/") + 1)}$reference"; | 1171 return "${base.substring(0, base.lastIndexOf("/") + 1)}$reference"; |
980 } | 1172 } |
(...skipping 477 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
1458 .toList(); | 1650 .toList(); |
1459 } | 1651 } |
1460 | 1652 |
1461 /** | 1653 /** |
1462 * Parse the [host] as an IP version 6 (IPv6) address, returning the address | 1654 * Parse the [host] as an IP version 6 (IPv6) address, returning the address |
1463 * as a list of 16 bytes in network byte order (big endian). | 1655 * as a list of 16 bytes in network byte order (big endian). |
1464 * | 1656 * |
1465 * Throws a [FormatException] if [host] is not a valid IPv6 address | 1657 * Throws a [FormatException] if [host] is not a valid IPv6 address |
1466 * representation. | 1658 * representation. |
1467 * | 1659 * |
1660 * Acts on the substring from [start] to [end]. If [end] is omitted, it | |
1661 * defaults ot the end of the string. | |
1662 * | |
1468 * Some examples of IPv6 addresses: | 1663 * Some examples of IPv6 addresses: |
1469 * * ::1 | 1664 * * ::1 |
1470 * * FEDC:BA98:7654:3210:FEDC:BA98:7654:3210 | 1665 * * FEDC:BA98:7654:3210:FEDC:BA98:7654:3210 |
1471 * * 3ffe:2a00:100:7031::1 | 1666 * * 3ffe:2a00:100:7031::1 |
1472 * * ::FFFF:129.144.52.38 | 1667 * * ::FFFF:129.144.52.38 |
1473 * * 2010:836B:4179::836B:4179 | 1668 * * 2010:836B:4179::836B:4179 |
1474 */ | 1669 */ |
1475 static List<int> parseIPv6Address(String host) { | 1670 static List<int> parseIPv6Address(String host, [int start = 0, int end]) { |
Lasse Reichstein Nielsen
2014/06/17 14:31:44
This is a public method that I added start/end to.
| |
1671 if (end == null) end = host.length; | |
1476 // An IPv6 address consists of exactly 8 parts of 1-4 hex digits, seperated | 1672 // An IPv6 address consists of exactly 8 parts of 1-4 hex digits, seperated |
1477 // by `:`'s, with the following exceptions: | 1673 // by `:`'s, with the following exceptions: |
1478 // | 1674 // |
1479 // - One (and only one) wildcard (`::`) may be present, representing a fill | 1675 // - One (and only one) wildcard (`::`) may be present, representing a fill |
1480 // of 0's. The IPv6 `::` is thus 16 bytes of `0`. | 1676 // of 0's. The IPv6 `::` is thus 16 bytes of `0`. |
1481 // - The last two parts may be replaced by an IPv4 address. | 1677 // - The last two parts may be replaced by an IPv4 address. |
1482 void error(String msg) { | 1678 void error(String msg) { |
1483 throw new FormatException('Illegal IPv6 address, $msg'); | 1679 throw new FormatException('Illegal IPv6 address, $msg'); |
1484 } | 1680 } |
1485 int parseHex(int start, int end) { | 1681 int parseHex(int start, int end) { |
1486 if (end - start > 4) { | 1682 if (end - start > 4) { |
1487 error('an IPv6 part can only contain a maximum of 4 hex digits'); | 1683 error('an IPv6 part can only contain a maximum of 4 hex digits'); |
1488 } | 1684 } |
1489 int value = int.parse(host.substring(start, end), radix: 16); | 1685 int value = int.parse(host.substring(start, end), radix: 16); |
1490 if (value < 0 || value > (1 << 16) - 1) { | 1686 if (value < 0 || value > (1 << 16) - 1) { |
1491 error('each part must be in the range of `0x0..0xFFFF`'); | 1687 error('each part must be in the range of `0x0..0xFFFF`'); |
1492 } | 1688 } |
1493 return value; | 1689 return value; |
1494 } | 1690 } |
1495 if (host.length < 2) error('address is too short'); | 1691 if (host.length < 2) error('address is too short'); |
1496 List<int> parts = []; | 1692 List<int> parts = []; |
1497 bool wildcardSeen = false; | 1693 bool wildcardSeen = false; |
1498 int partStart = 0; | 1694 int partStart = start; |
1499 // Parse all parts, except a potential last one. | 1695 // Parse all parts, except a potential last one. |
1500 for (int i = 0; i < host.length; i++) { | 1696 for (int i = start; i < end; i++) { |
1501 if (host.codeUnitAt(i) == _COLON) { | 1697 if (host.codeUnitAt(i) == _COLON) { |
1502 if (i == 0) { | 1698 if (i == start) { |
1503 // If we see a `:` in the beginning, expect wildcard. | 1699 // If we see a `:` in the beginning, expect wildcard. |
1504 i++; | 1700 i++; |
1505 if (host.codeUnitAt(i) != _COLON) { | 1701 if (host.codeUnitAt(i) != _COLON) { |
1506 error('invalid start colon.'); | 1702 error('invalid start colon.'); |
1507 } | 1703 } |
1508 partStart = i; | 1704 partStart = i; |
1509 } | 1705 } |
1510 if (i == partStart) { | 1706 if (i == partStart) { |
1511 // Wildcard. We only allow one. | 1707 // Wildcard. We only allow one. |
1512 if (wildcardSeen) { | 1708 if (wildcardSeen) { |
1513 error('only one wildcard `::` is allowed'); | 1709 error('only one wildcard `::` is allowed'); |
1514 } | 1710 } |
1515 wildcardSeen = true; | 1711 wildcardSeen = true; |
1516 parts.add(-1); | 1712 parts.add(-1); |
1517 } else { | 1713 } else { |
1518 // Found a single colon. Parse [partStart..i] as a hex entry. | 1714 // Found a single colon. Parse [partStart..i] as a hex entry. |
1519 parts.add(parseHex(partStart, i)); | 1715 parts.add(parseHex(partStart, i)); |
1520 } | 1716 } |
1521 partStart = i + 1; | 1717 partStart = i + 1; |
1522 } | 1718 } |
1523 } | 1719 } |
1524 if (parts.length == 0) error('too few parts'); | 1720 if (parts.length == 0) error('too few parts'); |
1525 bool atEnd = partStart == host.length; | 1721 bool atEnd = (partStart == end); |
1526 bool isLastWildcard = parts.last == -1; | 1722 bool isLastWildcard = (parts.last == -1); |
1527 if (atEnd && !isLastWildcard) { | 1723 if (atEnd && !isLastWildcard) { |
1528 error('expected a part after last `:`'); | 1724 error('expected a part after last `:`'); |
1529 } | 1725 } |
1530 if (!atEnd) { | 1726 if (!atEnd) { |
1531 try { | 1727 try { |
1532 parts.add(parseHex(partStart, host.length)); | 1728 parts.add(parseHex(partStart, end)); |
1533 } catch (e) { | 1729 } catch (e) { |
1534 // Failed to parse the last chunk as hex. Try IPv4. | 1730 // Failed to parse the last chunk as hex. Try IPv4. |
1535 try { | 1731 try { |
1536 List<int> last = parseIPv4Address(host.substring(partStart)); | 1732 List<int> last = parseIPv4Address(host.substring(partStart, end)); |
1537 parts.add(last[0] << 8 | last[1]); | 1733 parts.add(last[0] << 8 | last[1]); |
1538 parts.add(last[2] << 8 | last[3]); | 1734 parts.add(last[2] << 8 | last[3]); |
1539 } catch (e) { | 1735 } catch (e) { |
1540 error('invalid end of IPv6 address.'); | 1736 error('invalid end of IPv6 address.'); |
1541 } | 1737 } |
1542 } | 1738 } |
1543 } | 1739 } |
1544 if (wildcardSeen) { | 1740 if (wildcardSeen) { |
1545 if (parts.length > 7) { | 1741 if (parts.length > 7) { |
1546 error('an address with a wildcard must have less than 7 parts'); | 1742 error('an address with a wildcard must have less than 7 parts'); |
1547 } | 1743 } |
1548 } else if (parts.length != 8) { | 1744 } else if (parts.length != 8) { |
1549 error('an address without a wildcard must contain exactly 8 parts'); | 1745 error('an address without a wildcard must contain exactly 8 parts'); |
1550 } | 1746 } |
1551 // TODO(ajohnsen): Consider using Uint8List. | 1747 // TODO(ajohnsen): Consider using Uint8List. |
1552 return parts | 1748 List bytes = new List<int>(16); |
1553 .expand((value) { | 1749 for (int i = 0, index = 0; i < parts.length; i++) { |
1554 if (value == -1) { | 1750 int value = parts[i]; |
1555 return new List.filled((9 - parts.length) * 2, 0); | 1751 if (value == -1) { |
1556 } else { | 1752 int wildCardLength = 9 - parts.length; |
1557 return [(value >> 8) & 0xFF, value & 0xFF]; | 1753 for (int j = 0; j < wildCardLength; j++) { |
1558 } | 1754 bytes[index] = 0; |
1559 }) | 1755 bytes[index + 1] = 0; |
1560 .toList(); | 1756 index += 2; |
1757 } | |
1758 } else { | |
1759 bytes[index] = value >> 8; | |
1760 bytes[index + 1] = value & 0xff; | |
1761 index += 2; | |
1762 } | |
1763 } | |
1764 return bytes; | |
1561 } | 1765 } |
1562 | 1766 |
1563 // Frequently used character codes. | 1767 // Frequently used character codes. |
1564 static const int _SPACE = 0x20; | 1768 static const int _SPACE = 0x20; |
1565 static const int _DOUBLE_QUOTE = 0x22; | 1769 static const int _DOUBLE_QUOTE = 0x22; |
1566 static const int _NUMBER_SIGN = 0x23; | 1770 static const int _NUMBER_SIGN = 0x23; |
1567 static const int _PERCENT = 0x25; | 1771 static const int _PERCENT = 0x25; |
1568 static const int _ASTERISK = 0x2A; | 1772 static const int _ASTERISK = 0x2A; |
1569 static const int _PLUS = 0x2B; | 1773 static const int _PLUS = 0x2B; |
1570 static const int _SLASH = 0x2F; | 1774 static const int _SLASH = 0x2F; |
(...skipping 242 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
1813 0x2bff, // 0x30 - 0x3f 1111111111010100 | 2017 0x2bff, // 0x30 - 0x3f 1111111111010100 |
1814 // ABCDEFGHIJKLMNO | 2018 // ABCDEFGHIJKLMNO |
1815 0xfffe, // 0x40 - 0x4f 0111111111111111 | 2019 0xfffe, // 0x40 - 0x4f 0111111111111111 |
1816 // PQRSTUVWXYZ _ | 2020 // PQRSTUVWXYZ _ |
1817 0x87ff, // 0x50 - 0x5f 1111111111100001 | 2021 0x87ff, // 0x50 - 0x5f 1111111111100001 |
1818 // abcdefghijklmno | 2022 // abcdefghijklmno |
1819 0xfffe, // 0x60 - 0x6f 0111111111111111 | 2023 0xfffe, // 0x60 - 0x6f 0111111111111111 |
1820 // pqrstuvwxyz ~ | 2024 // pqrstuvwxyz ~ |
1821 0x47ff]; // 0x70 - 0x7f 1111111111100010 | 2025 0x47ff]; // 0x70 - 0x7f 1111111111100010 |
1822 | 2026 |
2027 // General delimiter characters, RFC 3986 section 2.2. | |
2028 // gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@" | |
2029 // | |
2030 static const _genDelimitersTable = const [ | |
2031 // LSB MSB | |
2032 // | | | |
2033 0x0000, // 0x00 - 0x0f 0000000000000000 | |
2034 0x0000, // 0x10 - 0x1f 0000000000000000 | |
2035 // # / | |
2036 0x8008, // 0x20 - 0x2f 0001000000000001 | |
2037 // : [ ]? | |
2038 0xe400, // 0x30 - 0x3f 0000000000101011 | |
2039 // @ | |
2040 0x0001, // 0x40 - 0x4f 1000000000000000 | |
2041 // | |
2042 0x0000, // 0x50 - 0x5f 0000000000000000 | |
2043 // | |
2044 0x0000, // 0x60 - 0x6f 0000000000000000 | |
2045 // | |
2046 0x0000]; // 0x70 - 0x7f 0000000000000000 | |
2047 | |
2048 // Characters allowed in the userinfo as of RFC 3986. | |
2049 // RFC 3986 Apendix A | |
2050 // userinfo = *( unreserved / pct-encoded / sub-delims / ':') | |
2051 static const _userinfoTable = const [ | |
2052 // LSB MSB | |
2053 // | | | |
2054 0x0000, // 0x00 - 0x0f 0000000000000000 | |
2055 0x0000, // 0x10 - 0x1f 0000000000000000 | |
2056 // ! $ &'()*+,-. | |
2057 0x7fd2, // 0x20 - 0x2f 0100101111111110 | |
2058 // 0123456789:; = | |
2059 0x2fff, // 0x30 - 0x3f 1111111111110100 | |
2060 // ABCDEFGHIJKLMNO | |
2061 0xfffe, // 0x40 - 0x4f 0111111111111111 | |
2062 // PQRSTUVWXYZ _ | |
2063 0x87ff, // 0x50 - 0x5f 1111111111100001 | |
2064 // abcdefghijklmno | |
2065 0xfffe, // 0x60 - 0x6f 0111111111111111 | |
2066 // pqrstuvwxyz ~ | |
2067 0x47ff]; // 0x70 - 0x7f 1111111111100010 | |
2068 | |
1823 // Characters allowed in the reg-name as of RFC 3986. | 2069 // Characters allowed in the reg-name as of RFC 3986. |
1824 // RFC 3986 Apendix A | 2070 // RFC 3986 Apendix A |
1825 // reg-name = *( unreserved / pct-encoded / sub-delims ) | 2071 // reg-name = *( unreserved / pct-encoded / sub-delims ) |
1826 static const _regNameTable = const [ | 2072 static const _regNameTable = const [ |
1827 // LSB MSB | 2073 // LSB MSB |
1828 // | | | 2074 // | | |
1829 0x0000, // 0x00 - 0x0f 0000000000000000 | 2075 0x0000, // 0x00 - 0x0f 0000000000000000 |
1830 0x0000, // 0x10 - 0x1f 0000000000000000 | 2076 0x0000, // 0x10 - 0x1f 0000000000000000 |
1831 // ! $%&'()*+,-. | 2077 // ! $%&'()*+,-. |
1832 0x7ff2, // 0x20 - 0x2f 0100111111111110 | 2078 0x7ff2, // 0x20 - 0x2f 0100111111111110 |
(...skipping 22 matching lines...) Expand all Loading... | |
1855 0x2fff, // 0x30 - 0x3f 1111111111110100 | 2101 0x2fff, // 0x30 - 0x3f 1111111111110100 |
1856 // @ABCDEFGHIJKLMNO | 2102 // @ABCDEFGHIJKLMNO |
1857 0xffff, // 0x40 - 0x4f 1111111111111111 | 2103 0xffff, // 0x40 - 0x4f 1111111111111111 |
1858 // PQRSTUVWXYZ _ | 2104 // PQRSTUVWXYZ _ |
1859 0x87ff, // 0x50 - 0x5f 1111111111100001 | 2105 0x87ff, // 0x50 - 0x5f 1111111111100001 |
1860 // abcdefghijklmno | 2106 // abcdefghijklmno |
1861 0xfffe, // 0x60 - 0x6f 0111111111111111 | 2107 0xfffe, // 0x60 - 0x6f 0111111111111111 |
1862 // pqrstuvwxyz ~ | 2108 // pqrstuvwxyz ~ |
1863 0x47ff]; // 0x70 - 0x7f 1111111111100010 | 2109 0x47ff]; // 0x70 - 0x7f 1111111111100010 |
1864 | 2110 |
2111 // Characters allowed in the path as of RFC 3986. | |
2112 // RFC 3986 section 3.3 *and* slash. | |
2113 static const _pathCharOrSlashTable = const [ | |
2114 // LSB MSB | |
2115 // | | | |
2116 0x0000, // 0x00 - 0x0f 0000000000000000 | |
2117 0x0000, // 0x10 - 0x1f 0000000000000000 | |
2118 // ! $ &'()*+,-./ | |
2119 0xffd2, // 0x20 - 0x2f 0100101111111111 | |
2120 // 0123456789:; = | |
2121 0x2fff, // 0x30 - 0x3f 1111111111110100 | |
2122 // @ABCDEFGHIJKLMNO | |
2123 0xffff, // 0x40 - 0x4f 1111111111111111 | |
2124 | |
2125 // PQRSTUVWXYZ _ | |
2126 0x87ff, // 0x50 - 0x5f 1111111111100001 | |
2127 // abcdefghijklmno | |
2128 0xfffe, // 0x60 - 0x6f 0111111111111111 | |
2129 // pqrstuvwxyz ~ | |
2130 0x47ff]; // 0x70 - 0x7f 1111111111100010 | |
2131 | |
1865 // Characters allowed in the query as of RFC 3986. | 2132 // Characters allowed in the query as of RFC 3986. |
1866 // RFC 3986 section 3.4. | 2133 // RFC 3986 section 3.4. |
1867 // query = *( pchar / "/" / "?" ) | 2134 // query = *( pchar / "/" / "?" ) |
1868 static const _queryCharTable = const [ | 2135 static const _queryCharTable = const [ |
1869 // LSB MSB | 2136 // LSB MSB |
1870 // | | | 2137 // | | |
1871 0x0000, // 0x00 - 0x0f 0000000000000000 | 2138 0x0000, // 0x00 - 0x0f 0000000000000000 |
1872 0x0000, // 0x10 - 0x1f 0000000000000000 | 2139 0x0000, // 0x10 - 0x1f 0000000000000000 |
1873 // ! $ &'()*+,-./ | 2140 // ! $ &'()*+,-./ |
1874 0xffd2, // 0x20 - 0x2f 0100101111111111 | 2141 0xffd2, // 0x20 - 0x2f 0100101111111111 |
1875 // 0123456789:; = ? | 2142 // 0123456789:; = ? |
1876 0xafff, // 0x30 - 0x3f 1111111111110101 | 2143 0xafff, // 0x30 - 0x3f 1111111111110101 |
1877 // @ABCDEFGHIJKLMNO | 2144 // @ABCDEFGHIJKLMNO |
1878 0xffff, // 0x40 - 0x4f 1111111111111111 | 2145 0xffff, // 0x40 - 0x4f 1111111111111111 |
1879 // PQRSTUVWXYZ _ | 2146 // PQRSTUVWXYZ _ |
1880 0x87ff, // 0x50 - 0x5f 1111111111100001 | 2147 0x87ff, // 0x50 - 0x5f 1111111111100001 |
1881 // abcdefghijklmno | 2148 // abcdefghijklmno |
1882 0xfffe, // 0x60 - 0x6f 0111111111111111 | 2149 0xfffe, // 0x60 - 0x6f 0111111111111111 |
1883 // pqrstuvwxyz ~ | 2150 // pqrstuvwxyz ~ |
1884 0x47ff]; // 0x70 - 0x7f 1111111111100010 | 2151 0x47ff]; // 0x70 - 0x7f 1111111111100010 |
1885 } | 2152 } |
OLD | NEW |