|
1 /* |
|
2 * Copyright (C) 1999 Lars Knoll (knoll@kde.org) |
|
3 * (C) 1999 Antti Koivisto (koivisto@kde.org) |
|
4 * (C) 2001 Dirk Mueller (mueller@kde.org) |
|
5 * (C) 2006 Alexey Proskuryakov (ap@webkit.org) |
|
6 * Copyright (C) 2004, 2005, 2006, 2007, 2008 Apple Inc. All rights reserved. |
|
7 * |
|
8 * This library is free software; you can redistribute it and/or |
|
9 * modify it under the terms of the GNU Library General Public |
|
10 * License as published by the Free Software Foundation; either |
|
11 * version 2 of the License, or (at your option) any later version. |
|
12 * |
|
13 * This library is distributed in the hope that it will be useful, |
|
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
|
16 * Library General Public License for more details. |
|
17 * |
|
18 * You should have received a copy of the GNU Library General Public License |
|
19 * along with this library; see the file COPYING.LIB. If not, write to |
|
20 * the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, |
|
21 * Boston, MA 02110-1301, USA. |
|
22 */ |
|
23 |
|
24 #include "config.h" |
|
25 #include "AtomicString.h" |
|
26 #include "KURL.h" |
|
27 #include "LinkHash.h" |
|
28 #include "PlatformString.h" |
|
29 #include "StringHash.h" |
|
30 #include "StringImpl.h" |
|
31 |
|
32 namespace WebCore { |
|
33 |
|
34 static inline int findSlashDotDotSlash(const UChar* characters, size_t length) |
|
35 { |
|
36 if (length < 4) |
|
37 return -1; |
|
38 unsigned loopLimit = length - 3; |
|
39 for (unsigned i = 0; i < loopLimit; ++i) { |
|
40 if (characters[i] == '/' && characters[i + 1] == '.' && characters[i + 2] == '.' && characters[i + 3] == '/') |
|
41 return i; |
|
42 } |
|
43 return -1; |
|
44 } |
|
45 |
|
46 static inline int findSlashSlash(const UChar* characters, size_t length, int position) |
|
47 { |
|
48 if (length < 2) |
|
49 return -1; |
|
50 unsigned loopLimit = length - 1; |
|
51 for (unsigned i = position; i < loopLimit; ++i) { |
|
52 if (characters[i] == '/' && characters[i + 1] == '/') |
|
53 return i; |
|
54 } |
|
55 return -1; |
|
56 } |
|
57 |
|
58 static inline int findSlashDotSlash(const UChar* characters, size_t length) |
|
59 { |
|
60 if (length < 3) |
|
61 return -1; |
|
62 unsigned loopLimit = length - 2; |
|
63 for (unsigned i = 0; i < loopLimit; ++i) { |
|
64 if (characters[i] == '/' && characters[i + 1] == '.' && characters[i + 2] == '/') |
|
65 return i; |
|
66 } |
|
67 return -1; |
|
68 } |
|
69 |
|
70 static inline bool containsColonSlashSlash(const UChar* characters, unsigned length) |
|
71 { |
|
72 if (length < 3) |
|
73 return false; |
|
74 unsigned loopLimit = length - 2; |
|
75 for (unsigned i = 0; i < loopLimit; ++i) { |
|
76 if (characters[i] == ':' && characters[i + 1] == '/' && characters[i + 2] == '/') |
|
77 return true; |
|
78 } |
|
79 return false; |
|
80 } |
|
81 |
|
82 static inline void cleanPath(Vector<UChar, 512>& path) |
|
83 { |
|
84 // FIXME: Should not do this in the query or anchor part. |
|
85 int pos; |
|
86 while ((pos = findSlashDotDotSlash(path.data(), path.size())) != -1) { |
|
87 int prev = reverseFind(path.data(), path.size(), '/', pos - 1); |
|
88 // don't remove the host, i.e. http://foo.org/../foo.html |
|
89 if (prev < 0 || (prev > 3 && path[prev - 2] == ':' && path[prev - 1] == '/')) |
|
90 path.remove(pos, 3); |
|
91 else |
|
92 path.remove(prev, pos - prev + 3); |
|
93 } |
|
94 |
|
95 // FIXME: Should not do this in the query part. |
|
96 // Set refPos to -2 to mean "I haven't looked for the anchor yet". |
|
97 // We don't want to waste a function call on the search for the the anchor |
|
98 // in the vast majority of cases where there is no "//" in the path. |
|
99 pos = 0; |
|
100 int refPos = -2; |
|
101 while ((pos = findSlashSlash(path.data(), path.size(), pos)) != -1) { |
|
102 if (refPos == -2) |
|
103 refPos = find(path.data(), path.size(), '#'); |
|
104 if (refPos > 0 && pos >= refPos) |
|
105 break; |
|
106 |
|
107 if (pos == 0 || path[pos - 1] != ':') |
|
108 path.remove(pos); |
|
109 else |
|
110 pos += 2; |
|
111 } |
|
112 |
|
113 // FIXME: Should not do this in the query or anchor part. |
|
114 while ((pos = findSlashDotSlash(path.data(), path.size())) != -1) |
|
115 path.remove(pos, 2); |
|
116 } |
|
117 |
|
118 |
|
119 static inline bool matchLetter(UChar c, UChar lowercaseLetter) |
|
120 { |
|
121 return (c | 0x20) == lowercaseLetter; |
|
122 } |
|
123 |
|
124 static inline bool needsTrailingSlash(const UChar* characters, unsigned length) |
|
125 { |
|
126 if (length < 6) |
|
127 return false; |
|
128 if (!matchLetter(characters[0], 'h') |
|
129 || !matchLetter(characters[1], 't') |
|
130 || !matchLetter(characters[2], 't') |
|
131 || !matchLetter(characters[3], 'p')) |
|
132 return false; |
|
133 if (!(characters[4] == ':' |
|
134 || (matchLetter(characters[4], 's') && characters[5] == ':'))) |
|
135 return false; |
|
136 |
|
137 unsigned pos = characters[4] == ':' ? 5 : 6; |
|
138 |
|
139 // Skip initial two slashes if present. |
|
140 if (pos + 1 < length && characters[pos] == '/' && characters[pos + 1] == '/') |
|
141 pos += 2; |
|
142 |
|
143 // Find next slash. |
|
144 while (pos < length && characters[pos] != '/') |
|
145 ++pos; |
|
146 |
|
147 return pos == length; |
|
148 } |
|
149 |
|
150 static ALWAYS_INLINE LinkHash visitedLinkHashInline(const UChar* url, unsigned length) |
|
151 { |
|
152 return AlreadyHashed::avoidDeletedValue(StringImpl::computeHash(url, length)); |
|
153 } |
|
154 |
|
155 LinkHash visitedLinkHash(const UChar* url, unsigned length) |
|
156 { |
|
157 return visitedLinkHashInline(url, length); |
|
158 } |
|
159 |
|
160 static ALWAYS_INLINE void visitedURLInline(const KURL& base, const AtomicString& attributeURL, Vector<UChar, 512>& buffer) |
|
161 { |
|
162 if (attributeURL.isNull()) |
|
163 return; |
|
164 |
|
165 const UChar* characters = attributeURL.characters(); |
|
166 unsigned length = attributeURL.length(); |
|
167 |
|
168 // This is a poor man's completeURL. Faster with less memory allocation. |
|
169 // FIXME: It's missing a lot of what completeURL does and a lot of what KURL does. |
|
170 // For example, it does not handle international domain names properly. |
|
171 |
|
172 // FIXME: It is wrong that we do not do further processing on strings that have "://" in them: |
|
173 // 1) The "://" could be in the query or anchor. |
|
174 // 2) The URL's path could have a "/./" or a "/../" or a "//" sequence in it. |
|
175 |
|
176 // FIXME: needsTrailingSlash does not properly return true for a URL that has no path, but does |
|
177 // have a query or anchor. |
|
178 |
|
179 bool hasColonSlashSlash = containsColonSlashSlash(characters, length); |
|
180 |
|
181 if (hasColonSlashSlash && !needsTrailingSlash(characters, length)) { |
|
182 buffer.append(attributeURL.characters(), attributeURL.length()); |
|
183 return; |
|
184 } |
|
185 |
|
186 |
|
187 if (hasColonSlashSlash) { |
|
188 // FIXME: This is incorrect for URLs that have a query or anchor; the "/" needs to go at the |
|
189 // end of the path, *before* the query or anchor. |
|
190 buffer.append(characters, length); |
|
191 buffer.append('/'); |
|
192 return; |
|
193 } |
|
194 |
|
195 if (!length) |
|
196 buffer.append(base.string().characters(), base.string().length()); |
|
197 else { |
|
198 switch (characters[0]) { |
|
199 case '/': |
|
200 buffer.append(base.string().characters(), base.pathStart()); |
|
201 break; |
|
202 case '#': |
|
203 buffer.append(base.string().characters(), base.pathEnd()); |
|
204 break; |
|
205 default: |
|
206 buffer.append(base.string().characters(), base.pathAfterLastSlash()); |
|
207 break; |
|
208 } |
|
209 } |
|
210 buffer.append(characters, length); |
|
211 cleanPath(buffer); |
|
212 if (needsTrailingSlash(buffer.data(), buffer.size())) { |
|
213 // FIXME: This is incorrect for URLs that have a query or anchor; the "/" needs to go at the |
|
214 // end of the path, *before* the query or anchor. |
|
215 buffer.append('/'); |
|
216 } |
|
217 |
|
218 return; |
|
219 } |
|
220 |
|
221 void visitedURL(const KURL& base, const AtomicString& attributeURL, Vector<UChar, 512>& buffer) |
|
222 { |
|
223 return visitedURLInline(base, attributeURL, buffer); |
|
224 } |
|
225 |
|
226 LinkHash visitedLinkHash(const KURL& base, const AtomicString& attributeURL) |
|
227 { |
|
228 Vector<UChar, 512> url; |
|
229 visitedURLInline(base, attributeURL, url); |
|
230 if (url.isEmpty()) |
|
231 return 0; |
|
232 |
|
233 return visitedLinkHashInline(url.data(), url.size()); |
|
234 } |
|
235 |
|
236 } // namespace WebCore |