1 | /* |
2 | * Copyright (C) 2013 Apple Inc. All rights reserved. |
3 | * Copyright (C) 2013 Google Inc. All rights reserved. |
4 | * |
5 | * Redistribution and use in source and binary forms, with or without |
6 | * modification, are permitted provided that the following conditions are |
7 | * met: |
8 | * |
9 | * * Redistributions of source code must retain the above copyright |
10 | * notice, this list of conditions and the following disclaimer. |
11 | * * Redistributions in binary form must reproduce the above |
12 | * copyright notice, this list of conditions and the following disclaimer |
13 | * in the documentation and/or other materials provided with the |
14 | * distribution. |
15 | * * Neither the name of Google Inc. nor the names of its |
16 | * contributors may be used to endorse or promote products derived from |
17 | * this software without specific prior written permission. |
18 | * |
19 | * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS |
20 | * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT |
21 | * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR |
22 | * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT |
23 | * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, |
24 | * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT |
25 | * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, |
26 | * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY |
27 | * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT |
28 | * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
29 | * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
30 | */ |
31 | |
32 | #include "config.h" |
33 | #include "HTMLSrcsetParser.h" |
34 | |
35 | #include "HTMLParserIdioms.h" |
36 | #include "ParsingUtilities.h" |
37 | |
38 | namespace WebCore { |
39 | |
40 | static inline bool compareByDensity(const ImageCandidate& first, const ImageCandidate& second) |
41 | { |
42 | return first.density < second.density; |
43 | } |
44 | |
45 | enum DescriptorTokenizerState { |
46 | Initial, |
47 | InParenthesis, |
48 | AfterToken, |
49 | }; |
50 | |
51 | template<typename CharType> |
52 | static void appendDescriptorAndReset(const CharType*& descriptorStart, const CharType* position, Vector<StringView>& descriptors) |
53 | { |
54 | if (position > descriptorStart) |
55 | descriptors.append(StringView(descriptorStart, position - descriptorStart)); |
56 | descriptorStart = nullptr; |
57 | } |
58 | |
59 | // The following is called appendCharacter to match the spec's terminology. |
60 | template<typename CharType> |
61 | static void appendCharacter(const CharType* descriptorStart, const CharType* position) |
62 | { |
63 | // Since we don't copy the tokens, this just set the point where the descriptor tokens start. |
64 | if (!descriptorStart) |
65 | descriptorStart = position; |
66 | } |
67 | |
68 | template<typename CharType> |
69 | static bool isEOF(const CharType* position, const CharType* end) |
70 | { |
71 | return position >= end; |
72 | } |
73 | |
74 | template<typename CharType> |
75 | static void tokenizeDescriptors(const CharType*& position, const CharType* attributeEnd, Vector<StringView>& descriptors) |
76 | { |
77 | DescriptorTokenizerState state = Initial; |
78 | const CharType* = position; |
79 | const CharType* currentDescriptorStart = descriptorsStart; |
80 | for (; ; ++position) { |
81 | switch (state) { |
82 | case Initial: |
83 | if (isEOF(position, attributeEnd)) { |
84 | appendDescriptorAndReset(currentDescriptorStart, attributeEnd, descriptors); |
85 | return; |
86 | } |
87 | if (isComma(*position)) { |
88 | appendDescriptorAndReset(currentDescriptorStart, position, descriptors); |
89 | ++position; |
90 | return; |
91 | } |
92 | if (isHTMLSpace(*position)) { |
93 | appendDescriptorAndReset(currentDescriptorStart, position, descriptors); |
94 | currentDescriptorStart = position + 1; |
95 | state = AfterToken; |
96 | } else if (*position == '(') { |
97 | appendCharacter(currentDescriptorStart, position); |
98 | state = InParenthesis; |
99 | } else |
100 | appendCharacter(currentDescriptorStart, position); |
101 | break; |
102 | case InParenthesis: |
103 | if (isEOF(position, attributeEnd)) { |
104 | appendDescriptorAndReset(currentDescriptorStart, attributeEnd, descriptors); |
105 | return; |
106 | } |
107 | if (*position == ')') { |
108 | appendCharacter(currentDescriptorStart, position); |
109 | state = Initial; |
110 | } else |
111 | appendCharacter(currentDescriptorStart, position); |
112 | break; |
113 | case AfterToken: |
114 | if (isEOF(position, attributeEnd)) |
115 | return; |
116 | if (!isHTMLSpace(*position)) { |
117 | state = Initial; |
118 | currentDescriptorStart = position; |
119 | --position; |
120 | } |
121 | break; |
122 | } |
123 | } |
124 | } |
125 | |
126 | static bool parseDescriptors(Vector<StringView>& descriptors, DescriptorParsingResult& result) |
127 | { |
128 | for (auto& descriptor : descriptors) { |
129 | if (descriptor.isEmpty()) |
130 | continue; |
131 | unsigned descriptorCharPosition = descriptor.length() - 1; |
132 | UChar descriptorChar = descriptor[descriptorCharPosition]; |
133 | descriptor = descriptor.substring(0, descriptorCharPosition); |
134 | if (descriptorChar == 'x') { |
135 | if (result.hasDensity() || result.hasHeight() || result.hasWidth()) |
136 | return false; |
137 | Optional<double> density = parseValidHTMLFloatingPointNumber(descriptor); |
138 | if (!density || density.value() < 0) |
139 | return false; |
140 | result.setDensity(density.value()); |
141 | } else if (descriptorChar == 'w') { |
142 | if (result.hasDensity() || result.hasWidth()) |
143 | return false; |
144 | Optional<int> resourceWidth = parseValidHTMLNonNegativeInteger(descriptor); |
145 | if (!resourceWidth || resourceWidth.value() <= 0) |
146 | return false; |
147 | result.setResourceWidth(resourceWidth.value()); |
148 | } else if (descriptorChar == 'h') { |
149 | // This is here only for future compat purposes. |
150 | // The value of the 'h' descriptor is not used. |
151 | if (result.hasDensity() || result.hasHeight()) |
152 | return false; |
153 | Optional<int> resourceHeight = parseValidHTMLNonNegativeInteger(descriptor); |
154 | if (!resourceHeight || resourceHeight.value() <= 0) |
155 | return false; |
156 | result.setResourceHeight(resourceHeight.value()); |
157 | } else |
158 | return false; |
159 | } |
160 | return !result.hasHeight() || result.hasWidth(); |
161 | } |
162 | |
163 | // http://picture.responsiveimages.org/#parse-srcset-attr |
164 | template<typename CharType> |
165 | static Vector<ImageCandidate> parseImageCandidatesFromSrcsetAttribute(const CharType* attributeStart, unsigned length) |
166 | { |
167 | Vector<ImageCandidate> imageCandidates; |
168 | |
169 | const CharType* attributeEnd = attributeStart + length; |
170 | |
171 | for (const CharType* position = attributeStart; position < attributeEnd;) { |
172 | // 4. Splitting loop: Collect a sequence of characters that are space characters or U+002C COMMA characters. |
173 | skipWhile<CharType, isHTMLSpaceOrComma<CharType> >(position, attributeEnd); |
174 | if (position == attributeEnd) { |
175 | // Contrary to spec language - descriptor parsing happens on each candidate, so when we reach the attributeEnd, we can exit. |
176 | break; |
177 | } |
178 | const CharType* imageURLStart = position; |
179 | // 6. Collect a sequence of characters that are not space characters, and let that be url. |
180 | |
181 | skipUntil<CharType, isHTMLSpace<CharType> >(position, attributeEnd); |
182 | const CharType* imageURLEnd = position; |
183 | |
184 | DescriptorParsingResult result; |
185 | |
186 | // 8. If url ends with a U+002C COMMA character (,) |
187 | if (isComma(*(position - 1))) { |
188 | // Remove all trailing U+002C COMMA characters from url. |
189 | imageURLEnd = position - 1; |
190 | reverseSkipWhile<CharType, isComma>(imageURLEnd, imageURLStart); |
191 | ++imageURLEnd; |
192 | // If url is empty, then jump to the step labeled splitting loop. |
193 | if (imageURLStart == imageURLEnd) |
194 | continue; |
195 | } else { |
196 | skipWhile<CharType, isHTMLSpace<CharType>>(position, attributeEnd); |
197 | Vector<StringView> descriptorTokens; |
198 | tokenizeDescriptors(position, attributeEnd, descriptorTokens); |
199 | // Contrary to spec language - descriptor parsing happens on each candidate. |
200 | // This is a black-box equivalent, to avoid storing descriptor lists for each candidate. |
201 | if (!parseDescriptors(descriptorTokens, result)) |
202 | continue; |
203 | } |
204 | |
205 | ASSERT(imageURLEnd > imageURLStart); |
206 | unsigned imageURLLength = imageURLEnd - imageURLStart; |
207 | imageCandidates.append(ImageCandidate(StringView(imageURLStart, imageURLLength), result, ImageCandidate::SrcsetOrigin)); |
208 | // 11. Return to the step labeled splitting loop. |
209 | } |
210 | return imageCandidates; |
211 | } |
212 | |
213 | Vector<ImageCandidate> parseImageCandidatesFromSrcsetAttribute(StringView attribute) |
214 | { |
215 | // FIXME: We should consider replacing the direct pointers in the parsing process with StringView and positions. |
216 | if (attribute.is8Bit()) |
217 | return parseImageCandidatesFromSrcsetAttribute<LChar>(attribute.characters8(), attribute.length()); |
218 | else |
219 | return parseImageCandidatesFromSrcsetAttribute<UChar>(attribute.characters16(), attribute.length()); |
220 | } |
221 | |
222 | static ImageCandidate pickBestImageCandidate(float deviceScaleFactor, Vector<ImageCandidate>& imageCandidates, float sourceSize) |
223 | { |
224 | bool ignoreSrc = false; |
225 | if (imageCandidates.isEmpty()) |
226 | return ImageCandidate(); |
227 | |
228 | // http://picture.responsiveimages.org/#normalize-source-densities |
229 | for (auto& candidate : imageCandidates) { |
230 | if (candidate.resourceWidth > 0) { |
231 | candidate.density = static_cast<float>(candidate.resourceWidth) / sourceSize; |
232 | ignoreSrc = true; |
233 | } else if (candidate.density < 0) |
234 | candidate.density = DefaultDensityValue; |
235 | } |
236 | |
237 | std::stable_sort(imageCandidates.begin(), imageCandidates.end(), compareByDensity); |
238 | |
239 | unsigned i; |
240 | for (i = 0; i < imageCandidates.size() - 1; ++i) { |
241 | if ((imageCandidates[i].density >= deviceScaleFactor) && (!ignoreSrc || !imageCandidates[i].srcOrigin())) |
242 | break; |
243 | } |
244 | |
245 | if (imageCandidates[i].srcOrigin() && ignoreSrc) { |
246 | ASSERT(i > 0); |
247 | --i; |
248 | } |
249 | float winningDensity = imageCandidates[i].density; |
250 | |
251 | unsigned winner = i; |
252 | // 16. If an entry b in candidates has the same associated ... pixel density as an earlier entry a in candidates, |
253 | // then remove entry b |
254 | while ((i > 0) && (imageCandidates[--i].density == winningDensity)) |
255 | winner = i; |
256 | |
257 | return imageCandidates[winner]; |
258 | } |
259 | |
260 | ImageCandidate bestFitSourceForImageAttributes(float deviceScaleFactor, const AtomicString& srcAttribute, const AtomicString& srcsetAttribute, float sourceSize) |
261 | { |
262 | if (srcsetAttribute.isNull()) { |
263 | if (srcAttribute.isNull()) |
264 | return ImageCandidate(); |
265 | return ImageCandidate(StringView(srcAttribute), DescriptorParsingResult(), ImageCandidate::SrcOrigin); |
266 | } |
267 | |
268 | Vector<ImageCandidate> imageCandidates = parseImageCandidatesFromSrcsetAttribute(StringView(srcsetAttribute)); |
269 | |
270 | if (!srcAttribute.isEmpty()) |
271 | imageCandidates.append(ImageCandidate(StringView(srcAttribute), DescriptorParsingResult(), ImageCandidate::SrcOrigin)); |
272 | |
273 | return pickBestImageCandidate(deviceScaleFactor, imageCandidates, sourceSize); |
274 | } |
275 | |
276 | } // namespace WebCore |
277 | |