1 /*
2 * Copyright (C) 2015 The Android Open Source Project
3 *
4 * Licensed under the Apache License, Version 2.0 (the "License");
5 * you may not use this file except in compliance with the License.
6 * You may obtain a copy of the License at
7 *
8 * http://www.apache.org/licenses/LICENSE-2.0
9 *
10 * Unless required by applicable law or agreed to in writing, software
11 * distributed under the License is distributed on an "AS IS" BASIS,
12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 * See the License for the specific language governing permissions and
14 * limitations under the License.
15 */
16
17 #ifndef AAPT_XML_PULL_PARSER_H
18 #define AAPT_XML_PULL_PARSER_H
19
20 #include <expat.h>
21
22 #include <algorithm>
23 #include <istream>
24 #include <ostream>
25 #include <queue>
26 #include <stack>
27 #include <string>
28 #include <vector>
29
30 #include "android-base/macros.h"
31 #include "androidfw/StringPiece.h"
32
33 #include "Resource.h"
34 #include "io/Io.h"
35 #include "process/IResourceTableConsumer.h"
36 #include "xml/XmlUtil.h"
37
38 namespace aapt {
39 namespace xml {
40
41 class XmlPullParser : public IPackageDeclStack {
42 public:
43 enum class Event {
44 kBadDocument,
45 kStartDocument,
46 kEndDocument,
47
48 kStartNamespace,
49 kEndNamespace,
50 kStartElement,
51 kEndElement,
52 kText,
53 kComment,
54 kCdataStart,
55 kCdataEnd,
56 };
57
58 /**
59 * Skips to the next direct descendant node of the given start_depth,
60 * skipping namespace nodes.
61 *
62 * When NextChildNode() returns true, you can expect Comments, Text, and
63 * StartElement events.
64 */
65 static bool NextChildNode(XmlPullParser* parser, size_t start_depth);
66 static bool SkipCurrentElement(XmlPullParser* parser);
67 static bool IsGoodEvent(Event event);
68
69 explicit XmlPullParser(io::InputStream* in);
70 ~XmlPullParser();
71
72 /**
73 * Returns the current event that is being processed.
74 */
75 Event event() const;
76
77 const std::string& error() const;
78
79 /**
80 * Note, unlike XmlPullParser, the first call to next() will return
81 * StartElement of the first element.
82 */
83 Event Next();
84
85 //
86 // These are available for all nodes.
87 //
88
89 const std::string& comment() const;
90 size_t line_number() const;
91 size_t depth() const;
92
93 /**
94 * Returns the character data for a Text event.
95 */
96 const std::string& text() const;
97
98 //
99 // Namespace prefix and URI are available for StartNamespace and EndNamespace.
100 //
101
102 const std::string& namespace_prefix() const;
103 const std::string& namespace_uri() const;
104
105 //
106 // These are available for StartElement and EndElement.
107 //
108
109 const std::string& element_namespace() const;
110 const std::string& element_name() const;
111
112 /*
113 * Uses the current stack of namespaces to resolve the package. Eg:
114 * xmlns:app = "http://schemas.android.com/apk/res/com.android.app"
115 * ...
116 * android:text="@app:string/message"
117 *
118 * In this case, 'app' will be converted to 'com.android.app'.
119 *
120 * If xmlns:app="http://schemas.android.com/apk/res-auto", then
121 * 'package' will be set to 'defaultPackage'.
122 */
123 std::optional<ExtractedPackage> TransformPackageAlias(
124 const android::StringPiece& alias) const override;
125
126 struct PackageDecl {
127 std::string prefix;
128 ExtractedPackage package;
129 };
130
131 const std::vector<PackageDecl>& package_decls() const;
132
133 //
134 // Remaining methods are for retrieving information about attributes
135 // associated with a StartElement.
136 //
137 // Attributes must be in sorted order (according to the less than operator
138 // of struct Attribute).
139 //
140
141 struct Attribute {
142 std::string namespace_uri;
143 std::string name;
144 std::string value;
145
146 int compare(const Attribute& rhs) const;
147 bool operator<(const Attribute& rhs) const;
148 bool operator==(const Attribute& rhs) const;
149 bool operator!=(const Attribute& rhs) const;
150 };
151
152 using const_iterator = std::vector<Attribute>::const_iterator;
153
154 const_iterator begin_attributes() const;
155 const_iterator end_attributes() const;
156 size_t attribute_count() const;
157 const_iterator FindAttribute(android::StringPiece namespace_uri, android::StringPiece name) const;
158
159 private:
160 DISALLOW_COPY_AND_ASSIGN(XmlPullParser);
161
162 static void XMLCALL StartNamespaceHandler(void* user_data, const char* prefix,
163 const char* uri);
164 static void XMLCALL StartElementHandler(void* user_data, const char* name,
165 const char** attrs);
166 static void XMLCALL CharacterDataHandler(void* user_data, const char* s,
167 int len);
168 static void XMLCALL EndElementHandler(void* user_data, const char* name);
169 static void XMLCALL EndNamespaceHandler(void* user_data, const char* prefix);
170 static void XMLCALL CommentDataHandler(void* user_data, const char* comment);
171 static void XMLCALL StartCdataSectionHandler(void* user_data);
172 static void XMLCALL EndCdataSectionHandler(void* user_data);
173
174 struct EventData {
175 Event event;
176 size_t line_number;
177 size_t depth;
178 std::string data1;
179 std::string data2;
180 std::vector<Attribute> attributes;
181 };
182
183 io::InputStream* in_;
184 XML_Parser parser_;
185 std::queue<EventData> event_queue_;
186 std::string error_;
187 const std::string empty_;
188 size_t depth_;
189 std::stack<std::string> namespace_uris_;
190 std::vector<PackageDecl> package_aliases_;
191 };
192
193 /**
194 * Finds the attribute in the current element within the global namespace.
195 */
196 std::optional<android::StringPiece> FindAttribute(const XmlPullParser* parser,
197 const android::StringPiece& name);
198
199 /**
200 * Finds the attribute in the current element within the global namespace. The
201 * attribute's value
202 * must not be the empty string.
203 */
204 std::optional<android::StringPiece> FindNonEmptyAttribute(const XmlPullParser* parser,
205 const android::StringPiece& name);
206
207 //
208 // Implementation
209 //
210
211 inline ::std::ostream& operator<<(::std::ostream& out,
212 XmlPullParser::Event event) {
213 switch (event) {
214 case XmlPullParser::Event::kBadDocument:
215 return out << "BadDocument";
216 case XmlPullParser::Event::kStartDocument:
217 return out << "StartDocument";
218 case XmlPullParser::Event::kEndDocument:
219 return out << "EndDocument";
220 case XmlPullParser::Event::kStartNamespace:
221 return out << "StartNamespace";
222 case XmlPullParser::Event::kEndNamespace:
223 return out << "EndNamespace";
224 case XmlPullParser::Event::kStartElement:
225 return out << "StartElement";
226 case XmlPullParser::Event::kEndElement:
227 return out << "EndElement";
228 case XmlPullParser::Event::kText:
229 return out << "Text";
230 case XmlPullParser::Event::kComment:
231 return out << "Comment";
232 case XmlPullParser::Event::kCdataStart:
233 return out << "CdataStart";
234 case XmlPullParser::Event::kCdataEnd:
235 return out << "CdataEnd";
236 }
237 return out;
238 }
239
NextChildNode(XmlPullParser * parser,size_t start_depth)240 inline bool XmlPullParser::NextChildNode(XmlPullParser* parser, size_t start_depth) {
241 Event event;
242
243 // First get back to the start depth.
244 while (IsGoodEvent(event = parser->Next()) && parser->depth() > start_depth + 1) {
245 }
246
247 // Now look for the first good node.
248 while ((event != Event::kEndElement || parser->depth() > start_depth) && IsGoodEvent(event)) {
249 switch (event) {
250 case Event::kText:
251 case Event::kComment:
252 case Event::kStartElement:
253 case Event::kCdataStart:
254 case Event::kCdataEnd:
255 return true;
256 default:
257 break;
258 }
259 event = parser->Next();
260 }
261 return false;
262 }
263
SkipCurrentElement(XmlPullParser * parser)264 inline bool XmlPullParser::SkipCurrentElement(XmlPullParser* parser) {
265 int depth = 1;
266 while (depth > 0) {
267 switch (parser->Next()) {
268 case Event::kEndDocument:
269 return true;
270 case Event::kBadDocument:
271 return false;
272 case Event::kStartElement:
273 depth++;
274 break;
275 case Event::kEndElement:
276 depth--;
277 break;
278 default:
279 break;
280 }
281 }
282 return true;
283 }
284
IsGoodEvent(XmlPullParser::Event event)285 inline bool XmlPullParser::IsGoodEvent(XmlPullParser::Event event) {
286 return event != Event::kBadDocument && event != Event::kEndDocument;
287 }
288
compare(const Attribute & rhs)289 inline int XmlPullParser::Attribute::compare(const Attribute& rhs) const {
290 int cmp = namespace_uri.compare(rhs.namespace_uri);
291 if (cmp != 0) return cmp;
292 return name.compare(rhs.name);
293 }
294
295 inline bool XmlPullParser::Attribute::operator<(const Attribute& rhs) const {
296 return compare(rhs) < 0;
297 }
298
299 inline bool XmlPullParser::Attribute::operator==(const Attribute& rhs) const {
300 return compare(rhs) == 0;
301 }
302
303 inline bool XmlPullParser::Attribute::operator!=(const Attribute& rhs) const {
304 return compare(rhs) != 0;
305 }
306
FindAttribute(android::StringPiece namespace_uri,android::StringPiece name)307 inline XmlPullParser::const_iterator XmlPullParser::FindAttribute(
308 android::StringPiece namespace_uri, android::StringPiece name) const {
309 const auto end_iter = end_attributes();
310 const auto iter = std::lower_bound(
311 begin_attributes(), end_iter,
312 std::pair<android::StringPiece, android::StringPiece>(namespace_uri, name),
313 [](const Attribute& attr,
314 const std::pair<android::StringPiece, android::StringPiece>& rhs) -> bool {
315 int cmp = attr.namespace_uri.compare(
316 0, attr.namespace_uri.size(), rhs.first.data(), rhs.first.size());
317 if (cmp < 0) return true;
318 if (cmp > 0) return false;
319 cmp = attr.name.compare(0, attr.name.size(), rhs.second.data(),
320 rhs.second.size());
321 if (cmp < 0) return true;
322 return false;
323 });
324
325 if (iter != end_iter && namespace_uri == iter->namespace_uri &&
326 name == iter->name) {
327 return iter;
328 }
329 return end_iter;
330 }
331
332 } // namespace xml
333 } // namespace aapt
334
335 #endif // AAPT_XML_PULL_PARSER_H
336