| OLD | NEW |
| 1 // Copyright (c) 2012 The Chromium Authors. All rights reserved. | 1 // Copyright (c) 2012 The Chromium Authors. All rights reserved. |
| 2 // Use of this source code is governed by a BSD-style license that can be | 2 // Use of this source code is governed by a BSD-style license that can be |
| 3 // found in the LICENSE file. | 3 // found in the LICENSE file. |
| 4 | 4 |
| 5 #include "net/base/escape.h" | 5 #include "net/base/escape.h" |
| 6 | 6 |
| 7 #include <algorithm> | 7 #include <algorithm> |
| 8 | 8 |
| 9 #include "base/logging.h" | 9 #include "base/logging.h" |
| 10 #include "base/memory/scoped_ptr.h" | 10 #include "base/memory/scoped_ptr.h" |
| 11 #include "base/strings/string_piece.h" | 11 #include "base/strings/string_piece.h" |
| 12 #include "base/strings/string_util.h" | 12 #include "base/strings/string_util.h" |
| 13 #include "base/strings/utf_offset_string_conversions.h" | 13 #include "base/strings/utf_offset_string_conversions.h" |
| 14 #include "base/strings/utf_string_conversions.h" | 14 #include "base/strings/utf_string_conversions.h" |
| 15 | 15 |
| 16 namespace net { | 16 namespace net { |
| 17 | 17 |
| 18 namespace { | 18 namespace { |
| 19 | 19 |
| 20 const char kHexString[] = "0123456789ABCDEF"; | 20 const char kHexString[] = "0123456789ABCDEF"; |
| 21 inline char IntToHex(int i) { | 21 inline char IntToHex(int i) { |
| 22 DCHECK_GE(i, 0) << i << " not a hex value"; | 22 DCHECK_GE(i, 0) << i << " not a hex value"; |
| 23 DCHECK_LE(i, 15) << i << " not a hex value"; | 23 DCHECK_LE(i, 15) << i << " not a hex value"; |
| 24 return kHexString[i]; | 24 return kHexString[i]; |
| 25 } | 25 } |
| 26 | 26 |
| 27 inline bool IsHexChar(unsigned char c) { |
| 28 if ((c >= 'a' && c <= 'f') || (c >= 'A' && c <= 'F') || |
| 29 (c >= '0' && c <= '9')) |
| 30 return true; |
| 31 return false; |
| 32 } |
| 33 |
| 27 // A fast bit-vector map for ascii characters. | 34 // A fast bit-vector map for ascii characters. |
| 28 // | 35 // |
| 29 // Internally stores 256 bits in an array of 8 ints. | 36 // Internally stores 256 bits in an array of 8 ints. |
| 30 // Does quick bit-flicking to lookup needed characters. | 37 // Does quick bit-flicking to lookup needed characters. |
| 31 struct Charmap { | 38 struct Charmap { |
| 32 bool Contains(unsigned char c) const { | 39 bool Contains(unsigned char c) const { |
| 33 return ((map[c >> 5] & (1 << (c & 31))) != 0); | 40 return ((map[c >> 5] & (1 << (c & 31))) != 0); |
| 34 } | 41 } |
| 35 | 42 |
| 36 uint32 map[8]; | 43 uint32 map[8]; |
| 37 }; | 44 }; |
| 38 | 45 |
| 39 // Given text to escape and a Charmap defining which values to escape, | 46 // Given text to escape and a Charmap defining which values to escape, |
| 40 // return an escaped string. If use_plus is true, spaces are converted | 47 // return an escaped string. If use_plus is true, spaces are converted |
| 41 // to +, otherwise, if spaces are in the charmap, they are converted to | 48 // to +, otherwise, if spaces are in the charmap, they are converted to |
| 42 // %20. | 49 // %20. |
| 43 std::string Escape(const std::string& text, const Charmap& charmap, | 50 std::string Escape(const std::string& text, const Charmap& charmap, |
| 44 bool use_plus) { | 51 bool use_plus, bool keep_escaped = false) { |
| 45 std::string escaped; | 52 std::string escaped; |
| 46 escaped.reserve(text.length() * 3); | 53 escaped.reserve(text.length() * 3); |
| 47 for (unsigned int i = 0; i < text.length(); ++i) { | 54 for (unsigned int i = 0; i < text.length(); ++i) { |
| 48 unsigned char c = static_cast<unsigned char>(text[i]); | 55 unsigned char c = static_cast<unsigned char>(text[i]); |
| 49 if (use_plus && ' ' == c) { | 56 if (use_plus && ' ' == c) { |
| 50 escaped.push_back('+'); | 57 escaped.push_back('+'); |
| 58 } else if (keep_escaped && '%' == c && i + 2 < text.length() && |
| 59 IsHexChar(static_cast<unsigned char>(text[i + 1])) && |
| 60 IsHexChar(static_cast<unsigned char>(text[i + 2]))) { |
| 61 escaped.push_back('%'); |
| 51 } else if (charmap.Contains(c)) { | 62 } else if (charmap.Contains(c)) { |
| 52 escaped.push_back('%'); | 63 escaped.push_back('%'); |
| 53 escaped.push_back(IntToHex(c >> 4)); | 64 escaped.push_back(IntToHex(c >> 4)); |
| 54 escaped.push_back(IntToHex(c & 0xf)); | 65 escaped.push_back(IntToHex(c & 0xf)); |
| 55 } else { | 66 } else { |
| 56 escaped.push_back(c); | 67 escaped.push_back(c); |
| 57 } | 68 } |
| 58 } | 69 } |
| 59 return escaped; | 70 return escaped; |
| 60 } | 71 } |
| (...skipping 231 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
| 292 0xffffffffL, 0xffffffffL, 0xffffffffL, 0xffffffffL | 303 0xffffffffL, 0xffffffffL, 0xffffffffL, 0xffffffffL |
| 293 }}; | 304 }}; |
| 294 | 305 |
| 295 // Everything except alphanumerics, the reserved characters(;/?:@&=+$,) and | 306 // Everything except alphanumerics, the reserved characters(;/?:@&=+$,) and |
| 296 // !'()*-._~% | 307 // !'()*-._~% |
| 297 static const Charmap kExternalHandlerCharmap = {{ | 308 static const Charmap kExternalHandlerCharmap = {{ |
| 298 0xffffffffL, 0x5000080dL, 0x68000000L, 0xb8000001L, | 309 0xffffffffL, 0x5000080dL, 0x68000000L, 0xb8000001L, |
| 299 0xffffffffL, 0xffffffffL, 0xffffffffL, 0xffffffffL | 310 0xffffffffL, 0xffffffffL, 0xffffffffL, 0xffffffffL |
| 300 }}; | 311 }}; |
| 301 | 312 |
| 313 // Everything except alphanumerics, the reserved characters(;/?:@&=+$,), the |
| 314 // mark characters(-_.!~*'()). |
| 315 static const Charmap kQueryPartCharmap = {{ |
| 316 0xffffffffL, 0x5000002dL, 0x78000000L, 0xb8000001L, |
| 317 0xffffffffL, 0xffffffffL, 0xffffffffL, 0xffffffffL |
| 318 }}; |
| 319 |
| 302 } // namespace | 320 } // namespace |
| 303 | 321 |
| 304 std::string EscapeQueryParamValue(const std::string& text, bool use_plus) { | 322 std::string EscapeQueryParamValue(const std::string& text, bool use_plus) { |
| 305 return Escape(text, kQueryCharmap, use_plus); | 323 return Escape(text, kQueryCharmap, use_plus); |
| 306 } | 324 } |
| 307 | 325 |
| 308 std::string EscapePath(const std::string& path) { | 326 std::string EscapePath(const std::string& path) { |
| 309 return Escape(path, kPathCharmap, false); | 327 return Escape(path, kPathCharmap, false); |
| 310 } | 328 } |
| 311 | 329 |
| 312 std::string EscapeUrlEncodedData(const std::string& path, bool use_plus) { | 330 std::string EscapeUrlEncodedData(const std::string& path, bool use_plus) { |
| 313 return Escape(path, kUrlEscape, use_plus); | 331 return Escape(path, kUrlEscape, use_plus); |
| 314 } | 332 } |
| 315 | 333 |
| 316 std::string EscapeNonASCII(const std::string& input) { | 334 std::string EscapeNonASCII(const std::string& input) { |
| 317 return Escape(input, kNonASCIICharmap, false); | 335 return Escape(input, kNonASCIICharmap, false); |
| 318 } | 336 } |
| 319 | 337 |
| 320 std::string EscapeExternalHandlerValue(const std::string& text) { | 338 std::string EscapeExternalHandlerValue(const std::string& text) { |
| 321 return Escape(text, kExternalHandlerCharmap, false); | 339 return Escape(text, kExternalHandlerCharmap, false); |
| 322 } | 340 } |
| 323 | 341 |
| 342 std::string EscapeQueryPart(const std::string& query) { |
| 343 return Escape(query, kQueryPartCharmap, false, true); |
| 344 } |
| 345 |
| 324 void AppendEscapedCharForHTML(char c, std::string* output) { | 346 void AppendEscapedCharForHTML(char c, std::string* output) { |
| 325 AppendEscapedCharForHTMLImpl(c, output); | 347 AppendEscapedCharForHTMLImpl(c, output); |
| 326 } | 348 } |
| 327 | 349 |
| 328 std::string EscapeForHTML(const std::string& input) { | 350 std::string EscapeForHTML(const std::string& input) { |
| 329 return EscapeForHTMLImpl(input); | 351 return EscapeForHTMLImpl(input); |
| 330 } | 352 } |
| 331 | 353 |
| 332 base::string16 EscapeForHTML(const base::string16& input) { | 354 base::string16 EscapeForHTML(const base::string16& input) { |
| 333 return EscapeForHTMLImpl(input); | 355 return EscapeForHTMLImpl(input); |
| (...skipping 68 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
| 402 1, kEscapeToChars[i].replacement); | 424 1, kEscapeToChars[i].replacement); |
| 403 break; | 425 break; |
| 404 } | 426 } |
| 405 } | 427 } |
| 406 } | 428 } |
| 407 } | 429 } |
| 408 return text; | 430 return text; |
| 409 } | 431 } |
| 410 | 432 |
| 411 } // namespace net | 433 } // namespace net |
| OLD | NEW |