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/net_util.h" | 5 #include "net/base/net_util.h" |
6 | 6 |
7 #include <map> | |
8 #include <vector> | 7 #include <vector> |
9 | 8 |
10 #include "base/i18n/time_formatting.h" | 9 #include "base/i18n/time_formatting.h" |
11 #include "base/json/string_escape.h" | 10 #include "base/json/string_escape.h" |
12 #include "base/lazy_instance.h" | 11 #include "base/lazy_instance.h" |
13 #include "base/logging.h" | 12 #include "base/logging.h" |
14 #include "base/memory/singleton.h" | 13 #include "base/memory/scoped_ptr.h" |
15 #include "base/stl_util.h" | |
16 #include "base/strings/string_tokenizer.h" | |
17 #include "base/strings/string_util.h" | 14 #include "base/strings/string_util.h" |
18 #include "base/strings/utf_offset_string_conversions.h" | 15 #include "base/strings/utf_offset_string_conversions.h" |
19 #include "base/strings/utf_string_conversions.h" | 16 #include "base/strings/utf_string_conversions.h" |
20 #include "base/time/time.h" | 17 #include "base/time/time.h" |
21 #include "url/gurl.h" | 18 #include "url/gurl.h" |
22 #include "third_party/icu/source/common/unicode/uidna.h" | 19 #include "third_party/icu/source/common/unicode/uidna.h" |
23 #include "third_party/icu/source/common/unicode/uniset.h" | 20 #include "third_party/icu/source/common/unicode/uniset.h" |
24 #include "third_party/icu/source/common/unicode/uscript.h" | |
25 #include "third_party/icu/source/common/unicode/uset.h" | |
26 #include "third_party/icu/source/i18n/unicode/datefmt.h" | 21 #include "third_party/icu/source/i18n/unicode/datefmt.h" |
27 #include "third_party/icu/source/i18n/unicode/regex.h" | 22 #include "third_party/icu/source/i18n/unicode/regex.h" |
28 #include "third_party/icu/source/i18n/unicode/ulocdata.h" | 23 #include "third_party/icu/source/i18n/unicode/uspoof.h" |
29 | 24 |
30 using base::Time; | 25 using base::Time; |
31 | 26 |
32 namespace net { | 27 namespace net { |
33 | 28 |
34 namespace { | 29 namespace { |
35 | 30 |
36 typedef std::vector<size_t> Offsets; | 31 typedef std::vector<size_t> Offsets; |
37 | 32 |
38 // Does some simple normalization of scripts so we can allow certain scripts | 33 class IDNSpoofChecker { |
Ryan Sleevi
2015/08/28 00:35:15
Document
jungshik at Google
2015/09/01 19:47:07
Done.
| |
39 // to exist together. | 34 public: |
40 // TODO(brettw) bug 880223: we should allow some other languages to be | 35 IDNSpoofChecker(); |
41 // oombined such as Chinese and Latin. We will probably need a more | 36 bool check(const base::char16* label, int label_len); |
Ryan Sleevi
2015/08/28 00:35:14
naming nit: These should be called Check
DANGER: u
Ryan Sleevi
2015/08/28 00:35:15
Document
jungshik at Google
2015/09/01 19:47:07
Done.
jungshik at Google
2015/09/01 19:47:07
Done.
jungshik at Google
2015/09/01 19:47:07
I switched to StringPiece16 (I thought about it bu
| |
42 // complicated system of language pairs to have more fine-grained control. | 37 |
43 UScriptCode NormalizeScript(UScriptCode code) { | 38 private: |
44 switch (code) { | 39 USpoofChecker* checker_; |
45 case USCRIPT_KATAKANA: | 40 DISALLOW_COPY_AND_ASSIGN(IDNSpoofChecker); |
46 case USCRIPT_HIRAGANA: | 41 }; |
47 case USCRIPT_KATAKANA_OR_HIRAGANA: | 42 |
48 case USCRIPT_HANGUL: // This one is arguable. | 43 base::LazyInstance<IDNSpoofChecker>::Leaky g_idn_spoof_checker = |
49 return USCRIPT_HAN; | 44 LAZY_INSTANCE_INITIALIZER; |
50 default: | 45 |
51 return code; | 46 class IDNSpoofCheckerExtra { |
Ryan Sleevi
2015/08/28 00:35:15
Document
jungshik at Google
2015/09/01 19:47:07
Done.
| |
52 } | 47 public: |
48 IDNSpoofCheckerExtra(); | |
49 bool check(const base::char16* label, int label_len); | |
Ryan Sleevi
2015/08/28 00:35:15
Same comments as above
jungshik at Google
2015/09/01 19:47:07
Done.
| |
50 | |
51 private: | |
52 icu::UnicodeSet non_ascii_latin_; | |
53 icu::RegexPattern* dangerous_pattern_; | |
54 DISALLOW_COPY_AND_ASSIGN(IDNSpoofCheckerExtra); | |
55 }; | |
56 | |
57 base::LazyInstance<IDNSpoofCheckerExtra>::Leaky g_idn_spoof_checker_extra = | |
58 LAZY_INSTANCE_INITIALIZER; | |
59 | |
60 IDNSpoofChecker::IDNSpoofChecker() { | |
61 UErrorCode status = U_ZERO_ERROR; | |
62 checker_ = uspoof_open(&status); | |
63 DCHECK(U_SUCCESS(status)) << "spoof checker failed to open with error: " | |
Ryan Sleevi
2015/08/28 00:35:15
Is DCHECK really appropriate? Does this indicate p
jungshik at Google
2015/09/01 19:47:07
This should never happen unless for some reason, I
| |
64 << u_errorName(status); | |
65 | |
66 // Use 'hightly restrictive' restritiction level to limit the script mixing | |
67 // to Latin + Han + {Hiragana + Katakana, Bopomofo, Hangul}. | |
68 // See http://www.unicode.org/reports/tr39/#Restriction_Level_Detection | |
69 // The default is highly restrictive so that it's not set explicitly. | |
70 // TODO(jshin): Firefox uses 'moderately restrictive' by default. Review | |
71 // using that, instead. | |
72 uspoof_setRestrictionLevel(checker_, USPOOF_HIGHLY_RESTRICTIVE); | |
73 | |
74 // The recommended set and inclusion set come from | |
75 // http://unicode.org/reports/tr39/ and | |
76 // http://www.unicode.org/Public/security/latest/xidmodifications.txt | |
77 // The list can undergo some changes as a new version of Unicode is | |
78 // released and we update our copy of ICU. | |
79 const icu::UnicodeSet* recommended_set = | |
80 uspoof_getRecommendedUnicodeSet(&status); | |
81 icu::UnicodeSet allowed_set; | |
82 allowed_set.addAll(*recommended_set); | |
83 const icu::UnicodeSet* inclusion_set = uspoof_getInclusionUnicodeSet(&status); | |
84 allowed_set.addAll(*inclusion_set); | |
85 | |
86 // From UAX 31 Table 6: | |
87 // http://www.unicode.org/reports/tr31/#Aspirational_Use_Scripts | |
88 const icu::UnicodeSet aspirational_scripts( | |
89 UNICODE_STRING_SIMPLE( | |
90 "[[:sc=Cans:][:sc=Plrd:][:sc=Mong:][:sc=Tfng:][:sc=Yiii:]]"), | |
91 status); | |
92 allowed_set.addAll(aspirational_scripts); | |
93 | |
94 // Add 'Black Heart Suit' and 'Circled White Star'. | |
95 // TODO(jshin): How about other heart-like characters and Emoji (e.g. | |
96 // U+1F600) ? | |
97 allowed_set.add(0x2665u); | |
98 allowed_set.add(0x272au); | |
99 | |
100 // Remove the following three characters listed in Mozilla's blacklist ( | |
101 // http://kb.mozillazine.org/Network.IDN.blacklist_chars ) but | |
102 // not yet excluded from |allowed_set| up to this point: | |
103 // Combining Long Solidus Overlay, Hebrew Punctuation Gershayim, and | |
104 // Hyphenation Point | |
105 allowed_set.remove(0x338u); // Combining Long Solidus Overlay | |
106 allowed_set.remove(0x5f4u); // Hebrew Punctuation Gershayim | |
107 allowed_set.remove(0x2027u); // Hyphenation Point | |
108 | |
109 // TODO(jshin): Decide what to do with '+' and U+0020. For now, leave | |
110 // them out as Mozilla does. | |
111 uspoof_setAllowedUnicodeSet(checker_, &allowed_set, &status); | |
112 | |
113 int32_t checks = uspoof_getChecks(checker_, &status); | |
114 // Do not allow mixed numbering systems (e.g. ASCII digits and | |
115 // Devanagari digits) or invisible characters or multiple occurrences | |
116 // there is a script mixing. | |
117 checks |= USPOOF_MIXED_NUMBERS | USPOOF_AUX_INFO; | |
118 | |
119 // USPOOF_INVISBLE should be on by this point without being | |
120 // explicitly turned on. | |
121 DCHECK(checks & USPOOF_INVISIBLE); | |
122 | |
123 // Disable whole-script-confusable check because even 'pax' (Latin) | |
124 // and b<u-umlaut>cher cannot pass the test because Cyrillic/Greek have | |
125 // confusable characters for all letters in them. | |
126 // TODO(jshin): Disabling this check has a downside. One way to alleviate | |
127 // is to check against a list of well known good domain names. | |
128 checks ^= USPOOF_WHOLE_SCRIPT_CONFUSABLE; | |
129 | |
130 uspoof_setChecks(checker_, checks, &status); | |
131 DCHECK(U_SUCCESS(status)); | |
53 } | 132 } |
54 | 133 |
55 bool IsIDNComponentInSingleScript(const base::char16* str, int str_len) { | 134 inline bool IDNSpoofChecker::check(const base::char16* label, int label_len) { |
56 UScriptCode first_script = USCRIPT_INVALID_CODE; | 135 UErrorCode status = U_ZERO_ERROR; |
57 bool is_first = true; | 136 int32_t results = uspoof_check(checker_, label, label_len, NULL, &status); |
137 DCHECK(U_SUCCESS(status)); | |
Ryan Sleevi
2015/08/28 00:35:15
Shouldn't you actually handle if this can fail?
jungshik at Google
2015/09/01 19:47:07
I agree. Done.
| |
138 if (results & USPOOF_ALL_CHECKS) | |
139 return false; | |
58 | 140 |
59 int i = 0; | 141 // If there's no script mixing, the input passes without any extra check. |
60 while (i < str_len) { | 142 if (results == USPOOF_ASCII || results == USPOOF_SINGLE_SCRIPT_RESTRICTIVE) |
61 unsigned code_point; | 143 return true; |
62 U16_NEXT(str, i, str_len, code_point); | |
63 | 144 |
64 UErrorCode err = U_ZERO_ERROR; | 145 return g_idn_spoof_checker_extra.Get().check(label, label_len); |
65 UScriptCode cur_script = uscript_getScript(code_point, &err); | |
66 if (err != U_ZERO_ERROR) | |
67 return false; // Report mixed on error. | |
68 cur_script = NormalizeScript(cur_script); | |
69 | |
70 // TODO(brettw) We may have to check for USCRIPT_INHERENT as well. | |
71 if (is_first && cur_script != USCRIPT_COMMON) { | |
72 first_script = cur_script; | |
73 is_first = false; | |
74 } else { | |
75 if (cur_script != USCRIPT_COMMON && cur_script != first_script) | |
76 return false; | |
77 } | |
78 } | |
79 return true; | |
80 } | 146 } |
81 | 147 |
82 // Check if the script of a language can be 'safely' mixed with | 148 IDNSpoofCheckerExtra::IDNSpoofCheckerExtra() { |
83 // Latin letters in the ASCII range. | 149 UErrorCode status = U_ZERO_ERROR; |
84 bool IsCompatibleWithASCIILetters(const std::string& lang) { | 150 non_ascii_latin_ = icu::UnicodeSet( |
85 // For now, just list Chinese, Japanese and Korean (positive list). | 151 UNICODE_STRING_SIMPLE("[[:sc=Latn:] - [a-zA-Z]]"), status); |
86 // An alternative is negative-listing (languages using Greek and | 152 |
87 // Cyrillic letters), but it can be more dangerous. | 153 dangerous_pattern_ = icu::RegexPattern::compile( |
88 return !lang.substr(0, 2).compare("zh") || | 154 UNICODE_STRING_SIMPLE( |
89 !lang.substr(0, 2).compare("ja") || | 155 // Lone (out-of-context) katakana no, so, zo, or n |
90 !lang.substr(0, 2).compare("ko"); | 156 // They can be mistaken for a slash. |
157 "[^\\p{Katakana}][\\u30ce\\u30f3\\u30bd\\u30be][^\\p{Katakana}]" | |
158 // Repeating Japanese accent characters. USPOOF_INVISIBLE | |
159 // only checks for a repeated occurence of the same combining | |
160 // mark, but we block a sequence of similary looking | |
161 // Japanese combining marks as well. | |
162 "|[\\u3099-\\u309c][\\u3099-\\u309c]"), | |
163 0, status); | |
164 DCHECK(U_SUCCESS(status)); | |
91 } | 165 } |
92 | 166 |
93 typedef std::map<std::string, icu::UnicodeSet*> LangToExemplarSetMap; | 167 inline bool IDNSpoofCheckerExtra::check(const base::char16* label, |
168 int label_len) { | |
169 // This is called only if script mixing is detected. | |
170 // Limit Latin letters that can be mixed with other scripts to | |
171 // ASCII-Latin instead of any Latin. | |
172 icu::UnicodeString label_string(FALSE, label, label_len); | |
173 if (non_ascii_latin_.containsSome(label_string)) | |
174 return false; | |
94 | 175 |
95 class LangToExemplarSet { | 176 UErrorCode status = U_ZERO_ERROR; |
96 public: | 177 scoped_ptr<icu::RegexMatcher> dangerous_pattern_matcher( |
97 static LangToExemplarSet* GetInstance() { | 178 dangerous_pattern_->matcher(label_string, status)); |
98 return Singleton<LangToExemplarSet>::get(); | 179 DCHECK(U_SUCCESS(status)); |
99 } | 180 return !dangerous_pattern_matcher->find(); |
100 | 181 |
101 private: | 182 // TODO(jshin): Check spoofing attempt against a list of 'good' domains |
102 LangToExemplarSetMap map; | |
103 LangToExemplarSet() { } | |
104 ~LangToExemplarSet() { | |
105 STLDeleteContainerPairSecondPointers(map.begin(), map.end()); | |
106 } | |
107 | |
108 friend class Singleton<LangToExemplarSet>; | |
109 friend struct DefaultSingletonTraits<LangToExemplarSet>; | |
110 friend bool GetExemplarSetForLang(const std::string&, icu::UnicodeSet**); | |
111 friend void SetExemplarSetForLang(const std::string&, icu::UnicodeSet*); | |
112 | |
113 DISALLOW_COPY_AND_ASSIGN(LangToExemplarSet); | |
114 }; | |
115 | |
116 bool GetExemplarSetForLang(const std::string& lang, | |
117 icu::UnicodeSet** lang_set) { | |
118 const LangToExemplarSetMap& map = LangToExemplarSet::GetInstance()->map; | |
119 LangToExemplarSetMap::const_iterator pos = map.find(lang); | |
120 if (pos != map.end()) { | |
121 *lang_set = pos->second; | |
122 return true; | |
123 } | |
124 return false; | |
125 } | |
126 | |
127 void SetExemplarSetForLang(const std::string& lang, | |
128 icu::UnicodeSet* lang_set) { | |
129 LangToExemplarSetMap& map = LangToExemplarSet::GetInstance()->map; | |
130 map.insert(std::make_pair(lang, lang_set)); | |
131 } | |
132 | |
133 static base::LazyInstance<base::Lock>::Leaky | |
134 g_lang_set_lock = LAZY_INSTANCE_INITIALIZER; | |
135 | |
136 // Returns true if all the characters in component_characters are used by | |
137 // the language |lang|. | |
138 bool IsComponentCoveredByLang(const icu::UnicodeSet& component_characters, | |
139 const std::string& lang) { | |
140 CR_DEFINE_STATIC_LOCAL( | |
141 const icu::UnicodeSet, kASCIILetters, ('a', 'z')); | |
142 icu::UnicodeSet* lang_set = nullptr; | |
143 // We're called from both the UI thread and the history thread. | |
144 { | |
145 base::AutoLock lock(g_lang_set_lock.Get()); | |
146 if (!GetExemplarSetForLang(lang, &lang_set)) { | |
147 UErrorCode status = U_ZERO_ERROR; | |
148 ULocaleData* uld = ulocdata_open(lang.c_str(), &status); | |
149 // TODO(jungshik) Turn this check on when the ICU data file is | |
150 // rebuilt with the minimal subset of locale data for languages | |
151 // to which Chrome is not localized but which we offer in the list | |
152 // of languages selectable for Accept-Languages. With the rebuilt ICU | |
153 // data, ulocdata_open never should fall back to the default locale. | |
154 // (issue 2078) | |
155 // DCHECK(U_SUCCESS(status) && status != U_USING_DEFAULT_WARNING); | |
156 if (U_SUCCESS(status) && status != U_USING_DEFAULT_WARNING) { | |
157 lang_set = reinterpret_cast<icu::UnicodeSet*>(ulocdata_getExemplarSet( | |
158 uld, nullptr, 0, ULOCDATA_ES_STANDARD, &status)); | |
159 // On success, if |lang| is compatible with ASCII Latin letters, add | |
160 // them. | |
161 if (lang_set && IsCompatibleWithASCIILetters(lang)) | |
162 lang_set->addAll(kASCIILetters); | |
163 } | |
164 | |
165 if (!lang_set) | |
166 lang_set = new icu::UnicodeSet(1, 0); | |
167 | |
168 lang_set->freeze(); | |
169 SetExemplarSetForLang(lang, lang_set); | |
170 ulocdata_close(uld); | |
171 } | |
172 } | |
173 return !lang_set->isEmpty() && lang_set->containsAll(component_characters); | |
174 } | 183 } |
175 | 184 |
176 // Returns true if the given Unicode host component is safe to display to the | 185 // Returns true if the given Unicode host component is safe to display to the |
177 // user. | 186 // user. |
178 bool IsIDNComponentSafe(const base::char16* str, | 187 bool IsIDNComponentSafe(const base::char16* label, int label_len) { |
179 int str_len, | 188 return g_idn_spoof_checker.Get().check(label, label_len); |
180 const std::string& languages) { | |
181 // Most common cases (non-IDN) do not reach here so that we don't | |
182 // need a fast return path. | |
183 // TODO(jungshik) : Check if there's any character inappropriate | |
184 // (although allowed) for domain names. | |
185 // See http://www.unicode.org/reports/tr39/#IDN_Security_Profiles and | |
186 // http://www.unicode.org/reports/tr39/data/xidmodifications.txt | |
187 // For now, we borrow the list from Mozilla and tweaked it slightly. | |
188 // (e.g. Characters like U+00A0, U+3000, U+3002 are omitted because | |
189 // they're gonna be canonicalized to U+0020 and full stop before | |
190 // reaching here.) | |
191 // The original list is available at | |
192 // http://kb.mozillazine.org/Network.IDN.blacklist_chars and | |
193 // at http://mxr.mozilla.org/seamonkey/source/modules/libpref/src/init/all.js# 703 | |
194 | |
195 UErrorCode status = U_ZERO_ERROR; | |
196 #ifdef U_WCHAR_IS_UTF16 | |
197 icu::UnicodeSet dangerous_characters( | |
198 icu::UnicodeString( | |
199 L"[[\\ \u00ad\u00bc\u00bd\u01c3\u0337\u0338" | |
200 L"\u05c3\u05f4\u06d4\u0702\u115f\u1160][\u2000-\u200b]" | |
201 L"[\u2024\u2027\u2028\u2029\u2039\u203a\u2044\u205f]" | |
202 L"[\u2154-\u2156][\u2159-\u215b][\u215f\u2215\u23ae" | |
203 L"\u29f6\u29f8\u2afb\u2afd][\u2ff0-\u2ffb][\u3014" | |
204 L"\u3015\u3033\u3164\u321d\u321e\u33ae\u33af\u33c6\u33df\ufe14" | |
205 L"\ufe15\ufe3f\ufe5d\ufe5e\ufeff\uff0e\uff06\uff61\uffa0\ufff9]" | |
206 L"[\ufffa-\ufffd]\U0001f50f\U0001f510\U0001f512\U0001f513]"), | |
207 status); | |
208 DCHECK(U_SUCCESS(status)); | |
209 icu::RegexMatcher dangerous_patterns(icu::UnicodeString( | |
210 // Lone katakana no, so, or n | |
211 L"[^\\p{Katakana}][\u30ce\u30f3\u30bd][^\\p{Katakana}]" | |
212 // Repeating Japanese accent characters | |
213 L"|[\u3099\u309a\u309b\u309c][\u3099\u309a\u309b\u309c]"), | |
214 0, status); | |
215 #else | |
216 icu::UnicodeSet dangerous_characters(icu::UnicodeString( | |
217 "[[\\u0020\\u00ad\\u00bc\\u00bd\\u01c3\\u0337\\u0338" | |
218 "\\u05c3\\u05f4\\u06d4\\u0702\\u115f\\u1160][\\u2000-\\u200b]" | |
219 "[\\u2024\\u2027\\u2028\\u2029\\u2039\\u203a\\u2044\\u205f]" | |
220 "[\\u2154-\\u2156][\\u2159-\\u215b][\\u215f\\u2215\\u23ae" | |
221 "\\u29f6\\u29f8\\u2afb\\u2afd][\\u2ff0-\\u2ffb][\\u3014" | |
222 "\\u3015\\u3033\\u3164\\u321d\\u321e\\u33ae\\u33af\\u33c6\\u33df\\ufe14" | |
223 "\\ufe15\\ufe3f\\ufe5d\\ufe5e\\ufeff\\uff0e\\uff06\\uff61\\uffa0\\ufff9]" | |
224 "[\\ufffa-\\ufffd]\\U0001f50f\\U0001f510\\U0001f512\\U0001f513]", -1, | |
225 US_INV), status); | |
226 DCHECK(U_SUCCESS(status)); | |
227 icu::RegexMatcher dangerous_patterns(icu::UnicodeString( | |
228 // Lone katakana no, so, or n | |
229 "[^\\p{Katakana}][\\u30ce\\u30f3\\u30bd][^\\p{Katakana}]" | |
230 // Repeating Japanese accent characters | |
231 "|[\\u3099\\u309a\\u309b\\u309c][\\u3099\\u309a\\u309b\\u309c]"), | |
232 0, status); | |
233 #endif | |
234 DCHECK(U_SUCCESS(status)); | |
235 icu::UnicodeSet component_characters; | |
236 icu::UnicodeString component_string(str, str_len); | |
237 component_characters.addAll(component_string); | |
238 if (dangerous_characters.containsSome(component_characters)) | |
239 return false; | |
240 | |
241 DCHECK(U_SUCCESS(status)); | |
242 dangerous_patterns.reset(component_string); | |
243 if (dangerous_patterns.find()) | |
244 return false; | |
245 | |
246 // If the language list is empty, the result is completely determined | |
247 // by whether a component is a single script or not. This will block | |
248 // even "safe" script mixing cases like <Chinese, Latin-ASCII> that are | |
249 // allowed with |languages| (while it blocks Chinese + Latin letters with | |
250 // an accent as should be the case), but we want to err on the safe side | |
251 // when |languages| is empty. | |
252 if (languages.empty()) | |
253 return IsIDNComponentInSingleScript(str, str_len); | |
254 | |
255 // |common_characters| is made up of ASCII numbers, hyphen, plus and | |
256 // underscore that are used across scripts and allowed in domain names. | |
257 // (sync'd with characters allowed in url_canon_host with square | |
258 // brackets excluded.) See kHostCharLookup[] array in url_canon_host.cc. | |
259 icu::UnicodeSet common_characters(UNICODE_STRING_SIMPLE("[[0-9]\\-_+\\ ]"), | |
260 status); | |
261 DCHECK(U_SUCCESS(status)); | |
262 // Subtract common characters because they're always allowed so that | |
263 // we just have to check if a language-specific set contains | |
264 // the remainder. | |
265 component_characters.removeAll(common_characters); | |
266 | |
267 base::StringTokenizer t(languages, ","); | |
268 while (t.GetNext()) { | |
269 if (IsComponentCoveredByLang(component_characters, t.token())) | |
270 return true; | |
271 } | |
272 return false; | |
273 } | 189 } |
274 | 190 |
275 // A wrapper to use LazyInstance<>::Leaky with ICU's UIDNA, a C pointer to | 191 // A wrapper to use LazyInstance<>::Leaky with ICU's UIDNA, a C pointer to |
276 // a UTS46/IDNA 2008 handling object opened with uidna_openUTS46(). | 192 // a UTS46/IDNA 2008 handling object opened with uidna_openUTS46(). |
277 // | 193 // |
278 // We use UTS46 with BiDiCheck to migrate from IDNA 2003 to IDNA 2008 with | 194 // We use UTS46 with BiDiCheck to migrate from IDNA 2003 to IDNA 2008 with |
279 // the backward compatibility in mind. What it does: | 195 // the backward compatibility in mind. What it does: |
280 // | 196 // |
281 // 1. Use the up-to-date Unicode data. | 197 // 1. Use the up-to-date Unicode data. |
282 // 2. Define a case folding/mapping with the up-to-date Unicode data as | 198 // 2. Define a case folding/mapping with the up-to-date Unicode data as |
(...skipping 15 matching lines...) Expand all Loading... | |
298 // TODO(jungshik): Change options as different parties (browsers, | 214 // TODO(jungshik): Change options as different parties (browsers, |
299 // registrars, search engines) converge toward a consensus. | 215 // registrars, search engines) converge toward a consensus. |
300 value = uidna_openUTS46(UIDNA_CHECK_BIDI, &err); | 216 value = uidna_openUTS46(UIDNA_CHECK_BIDI, &err); |
301 if (U_FAILURE(err)) | 217 if (U_FAILURE(err)) |
302 value = NULL; | 218 value = NULL; |
303 } | 219 } |
304 | 220 |
305 UIDNA* value; | 221 UIDNA* value; |
306 }; | 222 }; |
307 | 223 |
308 static base::LazyInstance<UIDNAWrapper>::Leaky | 224 base::LazyInstance<UIDNAWrapper>::Leaky g_uidna = LAZY_INSTANCE_INITIALIZER; |
309 g_uidna = LAZY_INSTANCE_INITIALIZER; | |
310 | 225 |
311 // Converts one component of a host (between dots) to IDN if safe. The result | 226 // Converts one component (label) of a host (between dots) to Unicode if safe. |
312 // will be APPENDED to the given output string and will be the same as the input | 227 // The result will be APPENDED to the given output string and will be the |
313 // if it is not IDN or the IDN is unsafe to display. Returns whether any | 228 // same as the input if it is not Punycode or the IDN is unsafe to display. |
314 // conversion was performed. | 229 // Returns whether any conversion was performed. |
315 bool IDNToUnicodeOneComponent(const base::char16* comp, | 230 bool IDNToUnicodeOneComponent(const base::char16* comp, |
316 size_t comp_len, | 231 size_t comp_len, |
317 const std::string& languages, | |
318 base::string16* out) { | 232 base::string16* out) { |
319 DCHECK(out); | 233 DCHECK(out); |
320 if (comp_len == 0) | 234 if (comp_len == 0) |
321 return false; | 235 return false; |
322 | 236 |
323 // Only transform if the input can be an IDN component. | 237 // Only transform if the input can be an IDN component. |
324 static const base::char16 kIdnPrefix[] = {'x', 'n', '-', '-'}; | 238 static const base::char16 kIdnPrefix[] = {'x', 'n', '-', '-'}; |
325 if ((comp_len > arraysize(kIdnPrefix)) && | 239 if ((comp_len > arraysize(kIdnPrefix)) && |
326 !memcmp(comp, kIdnPrefix, arraysize(kIdnPrefix) * sizeof(base::char16))) { | 240 !memcmp(comp, kIdnPrefix, arraysize(kIdnPrefix) * sizeof(base::char16))) { |
327 UIDNA* uidna = g_uidna.Get().value; | 241 UIDNA* uidna = g_uidna.Get().value; |
(...skipping 10 matching lines...) Expand all Loading... | |
338 // the conversion again, but with a sufficiently large buffer. | 252 // the conversion again, but with a sufficiently large buffer. |
339 output_length = uidna_labelToUnicode( | 253 output_length = uidna_labelToUnicode( |
340 uidna, comp, static_cast<int32_t>(comp_len), &(*out)[original_length], | 254 uidna, comp, static_cast<int32_t>(comp_len), &(*out)[original_length], |
341 output_length, &info, &status); | 255 output_length, &info, &status); |
342 } while ((status == U_BUFFER_OVERFLOW_ERROR && info.errors == 0)); | 256 } while ((status == U_BUFFER_OVERFLOW_ERROR && info.errors == 0)); |
343 | 257 |
344 if (U_SUCCESS(status) && info.errors == 0) { | 258 if (U_SUCCESS(status) && info.errors == 0) { |
345 // Converted successfully. Ensure that the converted component | 259 // Converted successfully. Ensure that the converted component |
346 // can be safely displayed to the user. | 260 // can be safely displayed to the user. |
347 out->resize(original_length + output_length); | 261 out->resize(original_length + output_length); |
348 if (IsIDNComponentSafe(out->data() + original_length, output_length, | 262 if (IsIDNComponentSafe(out->data() + original_length, output_length)) |
349 languages)) | |
350 return true; | 263 return true; |
351 } | 264 } |
352 | 265 |
353 // Something went wrong. Revert to original string. | 266 // Something went wrong. Revert to original string. |
354 out->resize(original_length); | 267 out->resize(original_length); |
355 } | 268 } |
356 | 269 |
357 // We get here with no IDN or on error, in which case we just append the | 270 // We get here with no IDN or on error, in which case we just append the |
358 // literal input. | 271 // literal input. |
359 out->append(comp, comp_len); | 272 out->append(comp, comp_len); |
360 return false; | 273 return false; |
361 } | 274 } |
362 | 275 |
363 // TODO(brettw) bug 734373: check the scripts for each host component and | 276 // TODO(brettw) We may want to skip this step in the case of file URLs to |
364 // don't un-IDN-ize if there is more than one. Alternatively, only IDN for | 277 // allow unicode UNC hostnames regardless of encodings. |
365 // scripts that the user has installed. For now, just put the entire | |
366 // path through IDN. Maybe this feature can be implemented in ICU itself? | |
367 // | |
368 // We may want to skip this step in the case of file URLs to allow unicode | |
369 // UNC hostnames regardless of encodings. | |
370 base::string16 IDNToUnicodeWithAdjustments( | 278 base::string16 IDNToUnicodeWithAdjustments( |
371 const std::string& host, | 279 const std::string& host, |
372 const std::string& languages, | |
373 base::OffsetAdjuster::Adjustments* adjustments) { | 280 base::OffsetAdjuster::Adjustments* adjustments) { |
374 if (adjustments) | 281 if (adjustments) |
375 adjustments->clear(); | 282 adjustments->clear(); |
376 // Convert the ASCII input to a base::string16 for ICU. | 283 // Convert the ASCII input to a base::string16 for ICU. |
377 base::string16 input16; | 284 base::string16 input16; |
378 input16.reserve(host.length()); | 285 input16.reserve(host.length()); |
379 input16.insert(input16.end(), host.begin(), host.end()); | 286 input16.insert(input16.end(), host.begin(), host.end()); |
380 | 287 |
381 // Do each component of the host separately, since we enforce script matching | 288 // Do each component of the host separately, since we enforce script matching |
382 // on a per-component basis. | 289 // on a per-component basis. |
383 base::string16 out16; | 290 base::string16 out16; |
384 { | 291 { |
385 for (size_t component_start = 0, component_end; | 292 for (size_t component_start = 0, component_end; |
386 component_start < input16.length(); | 293 component_start < input16.length(); |
387 component_start = component_end + 1) { | 294 component_start = component_end + 1) { |
388 // Find the end of the component. | 295 // Find the end of the component. |
389 component_end = input16.find('.', component_start); | 296 component_end = input16.find('.', component_start); |
390 if (component_end == base::string16::npos) | 297 if (component_end == base::string16::npos) |
391 component_end = input16.length(); // For getting the last component. | 298 component_end = input16.length(); // For getting the last component. |
392 size_t component_length = component_end - component_start; | 299 size_t component_length = component_end - component_start; |
393 size_t new_component_start = out16.length(); | 300 size_t new_component_start = out16.length(); |
394 bool converted_idn = false; | 301 bool converted_idn = false; |
395 if (component_end > component_start) { | 302 if (component_end > component_start) { |
396 // Add the substring that we just found. | 303 // Add the substring that we just found. |
397 converted_idn = IDNToUnicodeOneComponent( | 304 converted_idn = IDNToUnicodeOneComponent( |
398 input16.data() + component_start, component_length, languages, | 305 input16.data() + component_start, component_length, &out16); |
399 &out16); | |
400 } | 306 } |
401 size_t new_component_length = out16.length() - new_component_start; | 307 size_t new_component_length = out16.length() - new_component_start; |
402 | 308 |
403 if (converted_idn && adjustments) { | 309 if (converted_idn && adjustments) { |
404 adjustments->push_back(base::OffsetAdjuster::Adjustment( | 310 adjustments->push_back(base::OffsetAdjuster::Adjustment( |
405 component_start, component_length, new_component_length)); | 311 component_start, component_length, new_component_length)); |
406 } | 312 } |
407 | 313 |
408 // Need to add the dot we just found (if we found one). | 314 // Need to add the dot we just found (if we found one). |
409 if (component_end < input16.length()) | 315 if (component_end < input16.length()) |
(...skipping 19 matching lines...) Expand all Loading... | |
429 AdjustComponent(delta, &(parsed->host)); | 335 AdjustComponent(delta, &(parsed->host)); |
430 AdjustComponent(delta, &(parsed->port)); | 336 AdjustComponent(delta, &(parsed->port)); |
431 AdjustComponent(delta, &(parsed->path)); | 337 AdjustComponent(delta, &(parsed->path)); |
432 AdjustComponent(delta, &(parsed->query)); | 338 AdjustComponent(delta, &(parsed->query)); |
433 AdjustComponent(delta, &(parsed->ref)); | 339 AdjustComponent(delta, &(parsed->ref)); |
434 } | 340 } |
435 | 341 |
436 // Helper for FormatUrlWithOffsets(). | 342 // Helper for FormatUrlWithOffsets(). |
437 base::string16 FormatViewSourceUrl( | 343 base::string16 FormatViewSourceUrl( |
438 const GURL& url, | 344 const GURL& url, |
439 const std::string& languages, | |
440 FormatUrlTypes format_types, | 345 FormatUrlTypes format_types, |
441 UnescapeRule::Type unescape_rules, | 346 UnescapeRule::Type unescape_rules, |
442 url::Parsed* new_parsed, | 347 url::Parsed* new_parsed, |
443 size_t* prefix_end, | 348 size_t* prefix_end, |
444 base::OffsetAdjuster::Adjustments* adjustments) { | 349 base::OffsetAdjuster::Adjustments* adjustments) { |
445 DCHECK(new_parsed); | 350 DCHECK(new_parsed); |
446 const char kViewSource[] = "view-source:"; | 351 const char kViewSource[] = "view-source:"; |
447 const size_t kViewSourceLength = arraysize(kViewSource) - 1; | 352 const size_t kViewSourceLength = arraysize(kViewSource) - 1; |
448 | 353 |
449 // Format the underlying URL and record adjustments. | 354 // Format the underlying URL and record adjustments. |
450 const std::string& url_str(url.possibly_invalid_spec()); | 355 const std::string& url_str(url.possibly_invalid_spec()); |
451 adjustments->clear(); | 356 adjustments->clear(); |
452 base::string16 result(base::ASCIIToUTF16(kViewSource) + | 357 base::string16 result( |
358 base::ASCIIToUTF16(kViewSource) + | |
453 FormatUrlWithAdjustments(GURL(url_str.substr(kViewSourceLength)), | 359 FormatUrlWithAdjustments(GURL(url_str.substr(kViewSourceLength)), |
454 languages, format_types, unescape_rules, | 360 std::string(), format_types, unescape_rules, |
455 new_parsed, prefix_end, adjustments)); | 361 new_parsed, prefix_end, adjustments)); |
456 // Revise |adjustments| by shifting to the offsets to prefix that the above | 362 // Revise |adjustments| by shifting to the offsets to prefix that the above |
457 // call to FormatUrl didn't get to see. | 363 // call to FormatUrl didn't get to see. |
458 for (base::OffsetAdjuster::Adjustments::iterator it = adjustments->begin(); | 364 for (base::OffsetAdjuster::Adjustments::iterator it = adjustments->begin(); |
459 it != adjustments->end(); ++it) | 365 it != adjustments->end(); ++it) |
460 it->original_offset += kViewSourceLength; | 366 it->original_offset += kViewSourceLength; |
461 | 367 |
462 // Adjust positions of the parsed components. | 368 // Adjust positions of the parsed components. |
463 if (new_parsed->scheme.is_nonempty()) { | 369 if (new_parsed->scheme.is_nonempty()) { |
464 // Assume "view-source:real-scheme" as a scheme. | 370 // Assume "view-source:real-scheme" as a scheme. |
(...skipping 19 matching lines...) Expand all Loading... | |
484 const std::string& component_text, | 390 const std::string& component_text, |
485 base::OffsetAdjuster::Adjustments* adjustments) const = 0; | 391 base::OffsetAdjuster::Adjustments* adjustments) const = 0; |
486 | 392 |
487 // NOTE: No DISALLOW_COPY_AND_ASSIGN here, since gcc < 4.3.0 requires an | 393 // NOTE: No DISALLOW_COPY_AND_ASSIGN here, since gcc < 4.3.0 requires an |
488 // accessible copy constructor in order to call AppendFormattedComponent() | 394 // accessible copy constructor in order to call AppendFormattedComponent() |
489 // with an inline temporary (see http://gcc.gnu.org/bugs/#cxx%5Frvalbind ). | 395 // with an inline temporary (see http://gcc.gnu.org/bugs/#cxx%5Frvalbind ). |
490 }; | 396 }; |
491 | 397 |
492 class HostComponentTransform : public AppendComponentTransform { | 398 class HostComponentTransform : public AppendComponentTransform { |
493 public: | 399 public: |
494 explicit HostComponentTransform(const std::string& languages) | 400 explicit HostComponentTransform() {} |
495 : languages_(languages) { | |
496 } | |
497 | 401 |
498 private: | 402 private: |
499 base::string16 Execute( | 403 base::string16 Execute( |
500 const std::string& component_text, | 404 const std::string& component_text, |
501 base::OffsetAdjuster::Adjustments* adjustments) const override { | 405 base::OffsetAdjuster::Adjustments* adjustments) const override { |
502 return IDNToUnicodeWithAdjustments(component_text, languages_, | 406 return IDNToUnicodeWithAdjustments(component_text, adjustments); |
503 adjustments); | |
504 } | 407 } |
505 | |
506 const std::string& languages_; | |
507 }; | 408 }; |
508 | 409 |
509 class NonHostComponentTransform : public AppendComponentTransform { | 410 class NonHostComponentTransform : public AppendComponentTransform { |
510 public: | 411 public: |
511 explicit NonHostComponentTransform(UnescapeRule::Type unescape_rules) | 412 explicit NonHostComponentTransform(UnescapeRule::Type unescape_rules) |
512 : unescape_rules_(unescape_rules) { | 413 : unescape_rules_(unescape_rules) { |
513 } | 414 } |
514 | 415 |
515 private: | 416 private: |
516 base::string16 Execute( | 417 base::string16 Execute( |
(...skipping 49 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
566 if (output_component) { | 467 if (output_component) { |
567 output_component->begin = static_cast<int>(output_component_begin); | 468 output_component->begin = static_cast<int>(output_component_begin); |
568 output_component->len = | 469 output_component->len = |
569 static_cast<int>(output->length() - output_component_begin); | 470 static_cast<int>(output->length() - output_component_begin); |
570 } | 471 } |
571 } else if (output_component) { | 472 } else if (output_component) { |
572 output_component->reset(); | 473 output_component->reset(); |
573 } | 474 } |
574 } | 475 } |
575 | 476 |
576 } // namespace | 477 } // anonymous namespace |
577 | 478 |
578 const FormatUrlType kFormatUrlOmitNothing = 0; | 479 const FormatUrlType kFormatUrlOmitNothing = 0; |
579 const FormatUrlType kFormatUrlOmitUsernamePassword = 1 << 0; | 480 const FormatUrlType kFormatUrlOmitUsernamePassword = 1 << 0; |
580 const FormatUrlType kFormatUrlOmitHTTP = 1 << 1; | 481 const FormatUrlType kFormatUrlOmitHTTP = 1 << 1; |
581 const FormatUrlType kFormatUrlOmitTrailingSlashOnBareHostname = 1 << 2; | 482 const FormatUrlType kFormatUrlOmitTrailingSlashOnBareHostname = 1 << 2; |
582 const FormatUrlType kFormatUrlOmitAll = kFormatUrlOmitUsernamePassword | | 483 const FormatUrlType kFormatUrlOmitAll = kFormatUrlOmitUsernamePassword | |
583 kFormatUrlOmitHTTP | kFormatUrlOmitTrailingSlashOnBareHostname; | 484 kFormatUrlOmitHTTP | kFormatUrlOmitTrailingSlashOnBareHostname; |
584 | 485 |
585 base::string16 IDNToUnicode(const std::string& host, | 486 base::string16 IDNToUnicode(const std::string& host, |
586 const std::string& languages) { | 487 const std::string& languages) { |
587 return IDNToUnicodeWithAdjustments(host, languages, NULL); | 488 return IDNToUnicodeWithAdjustments(host, NULL); |
588 } | 489 } |
589 | 490 |
590 std::string GetDirectoryListingEntry(const base::string16& name, | 491 std::string GetDirectoryListingEntry(const base::string16& name, |
591 const std::string& raw_bytes, | 492 const std::string& raw_bytes, |
592 bool is_dir, | 493 bool is_dir, |
593 int64_t size, | 494 int64_t size, |
594 Time modified) { | 495 Time modified) { |
595 std::string result; | 496 std::string result; |
596 result.append("<script>addRow("); | 497 result.append("<script>addRow("); |
597 base::EscapeJSONString(name, true, &result); | 498 base::EscapeJSONString(name, true, &result); |
(...skipping 26 matching lines...) Expand all Loading... | |
624 | 525 |
625 result.append(");</script>\n"); | 526 result.append(");</script>\n"); |
626 | 527 |
627 return result; | 528 return result; |
628 } | 529 } |
629 | 530 |
630 void AppendFormattedHost(const GURL& url, | 531 void AppendFormattedHost(const GURL& url, |
631 const std::string& languages, | 532 const std::string& languages, |
632 base::string16* output) { | 533 base::string16* output) { |
633 AppendFormattedComponent(url.possibly_invalid_spec(), | 534 AppendFormattedComponent(url.possibly_invalid_spec(), |
634 url.parsed_for_possibly_invalid_spec().host, | 535 url.parsed_for_possibly_invalid_spec().host, |
635 HostComponentTransform(languages), output, NULL, NULL); | 536 HostComponentTransform(), output, NULL, NULL); |
636 } | 537 } |
637 | 538 |
638 base::string16 FormatUrlWithOffsets( | 539 base::string16 FormatUrlWithOffsets( |
639 const GURL& url, | 540 const GURL& url, |
640 const std::string& languages, | 541 const std::string& languages, |
641 FormatUrlTypes format_types, | 542 FormatUrlTypes format_types, |
642 UnescapeRule::Type unescape_rules, | 543 UnescapeRule::Type unescape_rules, |
643 url::Parsed* new_parsed, | 544 url::Parsed* new_parsed, |
644 size_t* prefix_end, | 545 size_t* prefix_end, |
645 std::vector<size_t>* offsets_for_adjustment) { | 546 std::vector<size_t>* offsets_for_adjustment) { |
646 base::OffsetAdjuster::Adjustments adjustments; | 547 base::OffsetAdjuster::Adjustments adjustments; |
647 const base::string16& format_url_return_value = | 548 const base::string16& format_url_return_value = |
648 FormatUrlWithAdjustments(url, languages, format_types, unescape_rules, | 549 FormatUrlWithAdjustments(url, std::string(), format_types, unescape_rules, |
649 new_parsed, prefix_end, &adjustments); | 550 new_parsed, prefix_end, &adjustments); |
650 base::OffsetAdjuster::AdjustOffsets(adjustments, offsets_for_adjustment); | 551 base::OffsetAdjuster::AdjustOffsets(adjustments, offsets_for_adjustment); |
651 if (offsets_for_adjustment) { | 552 if (offsets_for_adjustment) { |
652 std::for_each( | 553 std::for_each( |
653 offsets_for_adjustment->begin(), | 554 offsets_for_adjustment->begin(), |
654 offsets_for_adjustment->end(), | 555 offsets_for_adjustment->end(), |
655 base::LimitOffset<std::string>(format_url_return_value.length())); | 556 base::LimitOffset<std::string>(format_url_return_value.length())); |
656 } | 557 } |
657 return format_url_return_value; | 558 return format_url_return_value; |
658 } | 559 } |
(...skipping 15 matching lines...) Expand all Loading... | |
674 *new_parsed = url::Parsed(); | 575 *new_parsed = url::Parsed(); |
675 | 576 |
676 // Special handling for view-source:. Don't use content::kViewSourceScheme | 577 // Special handling for view-source:. Don't use content::kViewSourceScheme |
677 // because this library shouldn't depend on chrome. | 578 // because this library shouldn't depend on chrome. |
678 const char kViewSource[] = "view-source"; | 579 const char kViewSource[] = "view-source"; |
679 // Reject "view-source:view-source:..." to avoid deep recursion. | 580 // Reject "view-source:view-source:..." to avoid deep recursion. |
680 const char kViewSourceTwice[] = "view-source:view-source:"; | 581 const char kViewSourceTwice[] = "view-source:view-source:"; |
681 if (url.SchemeIs(kViewSource) && | 582 if (url.SchemeIs(kViewSource) && |
682 !base::StartsWith(url.possibly_invalid_spec(), kViewSourceTwice, | 583 !base::StartsWith(url.possibly_invalid_spec(), kViewSourceTwice, |
683 base::CompareCase::INSENSITIVE_ASCII)) { | 584 base::CompareCase::INSENSITIVE_ASCII)) { |
684 return FormatViewSourceUrl(url, languages, format_types, | 585 return FormatViewSourceUrl(url, format_types, unescape_rules, new_parsed, |
685 unescape_rules, new_parsed, prefix_end, | 586 prefix_end, adjustments); |
686 adjustments); | |
687 } | 587 } |
688 | 588 |
689 // We handle both valid and invalid URLs (this will give us the spec | 589 // We handle both valid and invalid URLs (this will give us the spec |
690 // regardless of validity). | 590 // regardless of validity). |
691 const std::string& spec = url.possibly_invalid_spec(); | 591 const std::string& spec = url.possibly_invalid_spec(); |
692 const url::Parsed& parsed = url.parsed_for_possibly_invalid_spec(); | 592 const url::Parsed& parsed = url.parsed_for_possibly_invalid_spec(); |
693 | 593 |
694 // Scheme & separators. These are ASCII. | 594 // Scheme & separators. These are ASCII. |
695 base::string16 url_string; | 595 base::string16 url_string; |
696 url_string.insert( | 596 url_string.insert( |
(...skipping 49 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
746 AppendFormattedComponent(spec, parsed.password, | 646 AppendFormattedComponent(spec, parsed.password, |
747 NonHostComponentTransform(unescape_rules), | 647 NonHostComponentTransform(unescape_rules), |
748 &url_string, &new_parsed->password, adjustments); | 648 &url_string, &new_parsed->password, adjustments); |
749 if (parsed.username.is_valid() || parsed.password.is_valid()) | 649 if (parsed.username.is_valid() || parsed.password.is_valid()) |
750 url_string.push_back('@'); | 650 url_string.push_back('@'); |
751 } | 651 } |
752 if (prefix_end) | 652 if (prefix_end) |
753 *prefix_end = static_cast<size_t>(url_string.length()); | 653 *prefix_end = static_cast<size_t>(url_string.length()); |
754 | 654 |
755 // Host. | 655 // Host. |
756 AppendFormattedComponent(spec, parsed.host, HostComponentTransform(languages), | 656 AppendFormattedComponent(spec, parsed.host, HostComponentTransform(), |
757 &url_string, &new_parsed->host, adjustments); | 657 &url_string, &new_parsed->host, adjustments); |
758 | 658 |
759 // Port. | 659 // Port. |
760 if (parsed.port.is_nonempty()) { | 660 if (parsed.port.is_nonempty()) { |
761 url_string.push_back(':'); | 661 url_string.push_back(':'); |
762 new_parsed->port.begin = url_string.length(); | 662 new_parsed->port.begin = url_string.length(); |
763 url_string.insert(url_string.end(), | 663 url_string.insert(url_string.end(), |
764 spec.begin() + parsed.port.begin, | 664 spec.begin() + parsed.port.begin, |
765 spec.begin() + parsed.port.end()); | 665 spec.begin() + parsed.port.end()); |
766 new_parsed->port.len = url_string.length() - new_parsed->port.begin; | 666 new_parsed->port.len = url_string.length() - new_parsed->port.begin; |
(...skipping 55 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... | |
822 base::string16 FormatUrl(const GURL& url, | 722 base::string16 FormatUrl(const GURL& url, |
823 const std::string& languages, | 723 const std::string& languages, |
824 FormatUrlTypes format_types, | 724 FormatUrlTypes format_types, |
825 UnescapeRule::Type unescape_rules, | 725 UnescapeRule::Type unescape_rules, |
826 url::Parsed* new_parsed, | 726 url::Parsed* new_parsed, |
827 size_t* prefix_end, | 727 size_t* prefix_end, |
828 size_t* offset_for_adjustment) { | 728 size_t* offset_for_adjustment) { |
829 Offsets offsets; | 729 Offsets offsets; |
830 if (offset_for_adjustment) | 730 if (offset_for_adjustment) |
831 offsets.push_back(*offset_for_adjustment); | 731 offsets.push_back(*offset_for_adjustment); |
832 base::string16 result = FormatUrlWithOffsets(url, languages, format_types, | 732 base::string16 result = |
833 unescape_rules, new_parsed, prefix_end, &offsets); | 733 FormatUrlWithOffsets(url, std::string(), format_types, unescape_rules, |
734 new_parsed, prefix_end, &offsets); | |
834 if (offset_for_adjustment) | 735 if (offset_for_adjustment) |
835 *offset_for_adjustment = offsets[0]; | 736 *offset_for_adjustment = offsets[0]; |
836 return result; | 737 return result; |
837 } | 738 } |
838 | 739 |
839 } // namespace net | 740 } // namespace net |
OLD | NEW |