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 #include "XmlDom.h"
18 #include "XmlPullParser.h"
19 #include "util/Util.h"
20 
21 #include <cassert>
22 #include <expat.h>
23 #include <memory>
24 #include <stack>
25 #include <string>
26 #include <tuple>
27 
28 namespace aapt {
29 namespace xml {
30 
31 constexpr char kXmlNamespaceSep = 1;
32 
33 struct Stack {
34     std::unique_ptr<xml::Node> root;
35     std::stack<xml::Node*> nodeStack;
36     std::u16string pendingComment;
37 };
38 
39 /**
40  * Extracts the namespace and name of an expanded element or attribute name.
41  */
splitName(const char * name,std::u16string * outNs,std::u16string * outName)42 static void splitName(const char* name, std::u16string* outNs, std::u16string* outName) {
43     const char* p = name;
44     while (*p != 0 && *p != kXmlNamespaceSep) {
45         p++;
46     }
47 
48     if (*p == 0) {
49         outNs->clear();
50         *outName = util::utf8ToUtf16(name);
51     } else {
52         *outNs = util::utf8ToUtf16(StringPiece(name, (p - name)));
53         *outName = util::utf8ToUtf16(p + 1);
54     }
55 }
56 
addToStack(Stack * stack,XML_Parser parser,std::unique_ptr<Node> node)57 static void addToStack(Stack* stack, XML_Parser parser, std::unique_ptr<Node> node) {
58     node->lineNumber = XML_GetCurrentLineNumber(parser);
59     node->columnNumber = XML_GetCurrentColumnNumber(parser);
60 
61     Node* thisNode = node.get();
62     if (!stack->nodeStack.empty()) {
63         stack->nodeStack.top()->addChild(std::move(node));
64     } else {
65         stack->root = std::move(node);
66     }
67 
68     if (!nodeCast<Text>(thisNode)) {
69         stack->nodeStack.push(thisNode);
70     }
71 }
72 
startNamespaceHandler(void * userData,const char * prefix,const char * uri)73 static void XMLCALL startNamespaceHandler(void* userData, const char* prefix, const char* uri) {
74     XML_Parser parser = reinterpret_cast<XML_Parser>(userData);
75     Stack* stack = reinterpret_cast<Stack*>(XML_GetUserData(parser));
76 
77     std::unique_ptr<Namespace> ns = util::make_unique<Namespace>();
78     if (prefix) {
79         ns->namespacePrefix = util::utf8ToUtf16(prefix);
80     }
81 
82     if (uri) {
83         ns->namespaceUri = util::utf8ToUtf16(uri);
84     }
85 
86     addToStack(stack, parser, std::move(ns));
87 }
88 
endNamespaceHandler(void * userData,const char * prefix)89 static void XMLCALL endNamespaceHandler(void* userData, const char* prefix) {
90     XML_Parser parser = reinterpret_cast<XML_Parser>(userData);
91     Stack* stack = reinterpret_cast<Stack*>(XML_GetUserData(parser));
92 
93     assert(!stack->nodeStack.empty());
94     stack->nodeStack.pop();
95 }
96 
lessAttribute(const Attribute & lhs,const Attribute & rhs)97 static bool lessAttribute(const Attribute& lhs, const Attribute& rhs) {
98     return std::tie(lhs.namespaceUri, lhs.name, lhs.value) <
99             std::tie(rhs.namespaceUri, rhs.name, rhs.value);
100 }
101 
startElementHandler(void * userData,const char * name,const char ** attrs)102 static void XMLCALL startElementHandler(void* userData, const char* name, const char** attrs) {
103     XML_Parser parser = reinterpret_cast<XML_Parser>(userData);
104     Stack* stack = reinterpret_cast<Stack*>(XML_GetUserData(parser));
105 
106     std::unique_ptr<Element> el = util::make_unique<Element>();
107     splitName(name, &el->namespaceUri, &el->name);
108 
109     while (*attrs) {
110         Attribute attribute;
111         splitName(*attrs++, &attribute.namespaceUri, &attribute.name);
112         attribute.value = util::utf8ToUtf16(*attrs++);
113 
114         // Insert in sorted order.
115         auto iter = std::lower_bound(el->attributes.begin(), el->attributes.end(), attribute,
116                                      lessAttribute);
117         el->attributes.insert(iter, std::move(attribute));
118     }
119 
120     el->comment = std::move(stack->pendingComment);
121     addToStack(stack, parser, std::move(el));
122 }
123 
endElementHandler(void * userData,const char * name)124 static void XMLCALL endElementHandler(void* userData, const char* name) {
125     XML_Parser parser = reinterpret_cast<XML_Parser>(userData);
126     Stack* stack = reinterpret_cast<Stack*>(XML_GetUserData(parser));
127 
128     assert(!stack->nodeStack.empty());
129     //stack->nodeStack.top()->comment = std::move(stack->pendingComment);
130     stack->nodeStack.pop();
131 }
132 
characterDataHandler(void * userData,const char * s,int len)133 static void XMLCALL characterDataHandler(void* userData, const char* s, int len) {
134     XML_Parser parser = reinterpret_cast<XML_Parser>(userData);
135     Stack* stack = reinterpret_cast<Stack*>(XML_GetUserData(parser));
136 
137     if (!s || len <= 0) {
138         return;
139     }
140 
141     // See if we can just append the text to a previous text node.
142     if (!stack->nodeStack.empty()) {
143         Node* currentParent = stack->nodeStack.top();
144         if (!currentParent->children.empty()) {
145             Node* lastChild = currentParent->children.back().get();
146             if (Text* text = nodeCast<Text>(lastChild)) {
147                 text->text += util::utf8ToUtf16(StringPiece(s, len));
148                 return;
149             }
150         }
151     }
152 
153     std::unique_ptr<Text> text = util::make_unique<Text>();
154     text->text = util::utf8ToUtf16(StringPiece(s, len));
155     addToStack(stack, parser, std::move(text));
156 }
157 
commentDataHandler(void * userData,const char * comment)158 static void XMLCALL commentDataHandler(void* userData, const char* comment) {
159     XML_Parser parser = reinterpret_cast<XML_Parser>(userData);
160     Stack* stack = reinterpret_cast<Stack*>(XML_GetUserData(parser));
161 
162     if (!stack->pendingComment.empty()) {
163         stack->pendingComment += '\n';
164     }
165     stack->pendingComment += util::utf8ToUtf16(comment);
166 }
167 
inflate(std::istream * in,IDiagnostics * diag,const Source & source)168 std::unique_ptr<XmlResource> inflate(std::istream* in, IDiagnostics* diag, const Source& source) {
169     Stack stack;
170 
171     XML_Parser parser = XML_ParserCreateNS(nullptr, kXmlNamespaceSep);
172     XML_SetUserData(parser, &stack);
173     XML_UseParserAsHandlerArg(parser);
174     XML_SetElementHandler(parser, startElementHandler, endElementHandler);
175     XML_SetNamespaceDeclHandler(parser, startNamespaceHandler, endNamespaceHandler);
176     XML_SetCharacterDataHandler(parser, characterDataHandler);
177     XML_SetCommentHandler(parser, commentDataHandler);
178 
179     char buffer[1024];
180     while (!in->eof()) {
181         in->read(buffer, sizeof(buffer) / sizeof(buffer[0]));
182         if (in->bad() && !in->eof()) {
183             stack.root = {};
184             diag->error(DiagMessage(source) << strerror(errno));
185             break;
186         }
187 
188         if (XML_Parse(parser, buffer, in->gcount(), in->eof()) == XML_STATUS_ERROR) {
189             stack.root = {};
190             diag->error(DiagMessage(source.withLine(XML_GetCurrentLineNumber(parser)))
191                         << XML_ErrorString(XML_GetErrorCode(parser)));
192             break;
193         }
194     }
195 
196     XML_ParserFree(parser);
197     if (stack.root) {
198         return util::make_unique<XmlResource>(ResourceFile{ {}, {}, source }, std::move(stack.root));
199     }
200     return {};
201 }
202 
copyAttributes(Element * el,android::ResXMLParser * parser)203 static void copyAttributes(Element* el, android::ResXMLParser* parser) {
204     const size_t attrCount = parser->getAttributeCount();
205     if (attrCount > 0) {
206         el->attributes.reserve(attrCount);
207         for (size_t i = 0; i < attrCount; i++) {
208             Attribute attr;
209             size_t len;
210             const char16_t* str16 = parser->getAttributeNamespace(i, &len);
211             if (str16) {
212                 attr.namespaceUri.assign(str16, len);
213             }
214 
215             str16 = parser->getAttributeName(i, &len);
216             if (str16) {
217                 attr.name.assign(str16, len);
218             }
219 
220             str16 = parser->getAttributeStringValue(i, &len);
221             if (str16) {
222                 attr.value.assign(str16, len);
223             }
224             el->attributes.push_back(std::move(attr));
225         }
226     }
227 }
228 
inflate(const void * data,size_t dataLen,IDiagnostics * diag,const Source & source)229 std::unique_ptr<XmlResource> inflate(const void* data, size_t dataLen, IDiagnostics* diag,
230                                      const Source& source) {
231     // We import the android namespace because on Windows NO_ERROR is a macro, not an enum, which
232     // causes errors when qualifying it with android::
233     using namespace android;
234 
235     std::unique_ptr<Node> root;
236     std::stack<Node*> nodeStack;
237 
238     ResXMLTree tree;
239     if (tree.setTo(data, dataLen) != NO_ERROR) {
240         return {};
241     }
242 
243     ResXMLParser::event_code_t code;
244     while ((code = tree.next()) != ResXMLParser::BAD_DOCUMENT &&
245             code != ResXMLParser::END_DOCUMENT) {
246         std::unique_ptr<Node> newNode;
247         switch (code) {
248             case ResXMLParser::START_NAMESPACE: {
249                 std::unique_ptr<Namespace> node = util::make_unique<Namespace>();
250                 size_t len;
251                 const char16_t* str16 = tree.getNamespacePrefix(&len);
252                 if (str16) {
253                     node->namespacePrefix.assign(str16, len);
254                 }
255 
256                 str16 = tree.getNamespaceUri(&len);
257                 if (str16) {
258                     node->namespaceUri.assign(str16, len);
259                 }
260                 newNode = std::move(node);
261                 break;
262             }
263 
264             case ResXMLParser::START_TAG: {
265                 std::unique_ptr<Element> node = util::make_unique<Element>();
266                 size_t len;
267                 const char16_t* str16 = tree.getElementNamespace(&len);
268                 if (str16) {
269                     node->namespaceUri.assign(str16, len);
270                 }
271 
272                 str16 = tree.getElementName(&len);
273                 if (str16) {
274                     node->name.assign(str16, len);
275                 }
276 
277                 copyAttributes(node.get(), &tree);
278 
279                 newNode = std::move(node);
280                 break;
281             }
282 
283             case ResXMLParser::TEXT: {
284                 std::unique_ptr<Text> node = util::make_unique<Text>();
285                 size_t len;
286                 const char16_t* str16 = tree.getText(&len);
287                 if (str16) {
288                     node->text.assign(str16, len);
289                 }
290                 newNode = std::move(node);
291                 break;
292             }
293 
294             case ResXMLParser::END_NAMESPACE:
295             case ResXMLParser::END_TAG:
296                 assert(!nodeStack.empty());
297                 nodeStack.pop();
298                 break;
299 
300             default:
301                 assert(false);
302                 break;
303         }
304 
305         if (newNode) {
306             newNode->lineNumber = tree.getLineNumber();
307 
308             Node* thisNode = newNode.get();
309             if (!root) {
310                 assert(nodeStack.empty());
311                 root = std::move(newNode);
312             } else {
313                 assert(!nodeStack.empty());
314                 nodeStack.top()->addChild(std::move(newNode));
315             }
316 
317             if (!nodeCast<Text>(thisNode)) {
318                 nodeStack.push(thisNode);
319             }
320         }
321     }
322     return util::make_unique<XmlResource>(ResourceFile{}, std::move(root));
323 }
324 
findRootElement(XmlResource * doc)325 Element* findRootElement(XmlResource* doc) {
326     return findRootElement(doc->root.get());
327 }
328 
findRootElement(Node * node)329 Element* findRootElement(Node* node) {
330     if (!node) {
331         return nullptr;
332     }
333 
334     Element* el = nullptr;
335     while ((el = nodeCast<Element>(node)) == nullptr) {
336         if (node->children.empty()) {
337             return nullptr;
338         }
339         // We are looking for the first element, and namespaces can only have one child.
340         node = node->children.front().get();
341     }
342     return el;
343 }
344 
addChild(std::unique_ptr<Node> child)345 void Node::addChild(std::unique_ptr<Node> child) {
346     child->parent = this;
347     children.push_back(std::move(child));
348 }
349 
findAttribute(const StringPiece16 & ns,const StringPiece16 & name)350 Attribute* Element::findAttribute(const StringPiece16& ns, const StringPiece16& name) {
351     for (auto& attr : attributes) {
352         if (ns == attr.namespaceUri && name == attr.name) {
353             return &attr;
354         }
355     }
356     return nullptr;
357 }
358 
findChild(const StringPiece16 & ns,const StringPiece16 & name)359 Element* Element::findChild(const StringPiece16& ns, const StringPiece16& name) {
360     return findChildWithAttribute(ns, name, {}, {}, {});
361 }
362 
findChildWithAttribute(const StringPiece16 & ns,const StringPiece16 & name,const StringPiece16 & attrNs,const StringPiece16 & attrName,const StringPiece16 & attrValue)363 Element* Element::findChildWithAttribute(const StringPiece16& ns, const StringPiece16& name,
364                                          const StringPiece16& attrNs, const StringPiece16& attrName,
365                                          const StringPiece16& attrValue) {
366     for (auto& childNode : children) {
367         Node* child = childNode.get();
368         while (nodeCast<Namespace>(child)) {
369             if (child->children.empty()) {
370                 break;
371             }
372             child = child->children[0].get();
373         }
374 
375         if (Element* el = nodeCast<Element>(child)) {
376             if (ns == el->namespaceUri && name == el->name) {
377                 if (attrNs.empty() && attrName.empty()) {
378                     return el;
379                 }
380 
381                 Attribute* attr = el->findAttribute(attrNs, attrName);
382                 if (attr && attrValue == attr->value) {
383                     return el;
384                 }
385             }
386         }
387     }
388     return nullptr;
389 }
390 
getChildElements()391 std::vector<Element*> Element::getChildElements() {
392     std::vector<Element*> elements;
393     for (auto& childNode : children) {
394         Node* child = childNode.get();
395         while (nodeCast<Namespace>(child)) {
396             if (child->children.empty()) {
397                 break;
398             }
399             child = child->children[0].get();
400         }
401 
402         if (Element* el = nodeCast<Element>(child)) {
403             elements.push_back(el);
404         }
405     }
406     return elements;
407 }
408 
visit(Namespace * ns)409 void PackageAwareVisitor::visit(Namespace* ns) {
410    bool added = false;
411    if (Maybe<ExtractedPackage> maybePackage = extractPackageFromNamespace(ns->namespaceUri)) {
412        ExtractedPackage& package = maybePackage.value();
413        mPackageDecls.push_back(PackageDecl{ ns->namespacePrefix, std::move(package) });
414        added = true;
415    }
416 
417    Visitor::visit(ns);
418 
419    if (added) {
420        mPackageDecls.pop_back();
421    }
422 }
423 
transformPackageAlias(const StringPiece16 & alias,const StringPiece16 & localPackage) const424 Maybe<ExtractedPackage> PackageAwareVisitor::transformPackageAlias(
425        const StringPiece16& alias, const StringPiece16& localPackage) const {
426    if (alias.empty()) {
427        return ExtractedPackage{ localPackage.toString(), false /* private */ };
428    }
429 
430    const auto rend = mPackageDecls.rend();
431    for (auto iter = mPackageDecls.rbegin(); iter != rend; ++iter) {
432        if (alias == iter->prefix) {
433            if (iter->package.package.empty()) {
434                return ExtractedPackage{ localPackage.toString(),
435                                               iter->package.privateNamespace };
436            }
437            return iter->package;
438        }
439    }
440    return {};
441 }
442 
443 } // namespace xml
444 } // namespace aapt
445