1/*
2 * This file is part of the XSL implementation.
3 *
4 * Copyright (C) 2004, 2005, 2006, 2008, 2012 Apple Inc. All rights reserved.
5 *
6 * This library is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Library General Public
8 * License as published by the Free Software Foundation; either
9 * version 2 of the License, or (at your option) any later version.
10 *
11 * This library is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Library General Public License for more details.
15 *
16 * You should have received a copy of the GNU Library General Public License
17 * along with this library; see the file COPYING.LIB. If not, write to
18 * the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
19 * Boston, MA 02110-1301, USA.
20 */
21
22#include "config.h"
23
24#if ENABLE(XSLT)
25
26#include "CachedResourceLoader.h"
27#include "Document.h"
28#include "Frame.h"
29#include "Page.h"
30#include "PageConsoleClient.h"
31#include "TransformSource.h"
32#include "XMLDocumentParser.h"
33#include "XMLDocumentParserScope.h"
34#include "XSLImportRule.h"
35#include "XSLTProcessor.h"
36#include <libxml/uri.h>
37#include <libxslt/xsltutils.h>
38#include <wtf/CheckedArithmetic.h>
39
40#if OS(DARWIN) && !PLATFORM(GTK)
41#include "SoftLinkLibxslt.h"
42#endif
43
44namespace WebCore {
45
46XSLStyleSheet::XSLStyleSheet(XSLImportRule* parentRule, const String& originalURL, const URL& finalURL)
47 : m_ownerNode(nullptr)
48 , m_originalURL(originalURL)
49 , m_finalURL(finalURL)
50 , m_embedded(false)
51 , m_processed(false) // Child sheets get marked as processed when the libxslt engine has finally seen them.
52{
53 if (parentRule)
54 m_parentStyleSheet = parentRule->parentStyleSheet();
55}
56
57XSLStyleSheet::XSLStyleSheet(Node* parentNode, const String& originalURL, const URL& finalURL, bool embedded)
58 : m_ownerNode(parentNode)
59 , m_originalURL(originalURL)
60 , m_finalURL(finalURL)
61 , m_embedded(embedded)
62 , m_processed(true) // The root sheet starts off processed.
63{
64}
65
66XSLStyleSheet::~XSLStyleSheet()
67{
68 clearXSLStylesheetDocument();
69
70 for (auto& import : m_children) {
71 ASSERT(import->parentStyleSheet() == this);
72 import->setParentStyleSheet(nullptr);
73 }
74}
75
76bool XSLStyleSheet::isLoading() const
77{
78 for (auto& import : m_children) {
79 if (import->isLoading())
80 return true;
81 }
82 return false;
83}
84
85void XSLStyleSheet::checkLoaded()
86{
87 if (isLoading())
88 return;
89 if (XSLStyleSheet* styleSheet = parentStyleSheet())
90 styleSheet->checkLoaded();
91 if (ownerNode())
92 ownerNode()->sheetLoaded();
93}
94
95xmlDocPtr XSLStyleSheet::document()
96{
97 if (m_embedded && ownerDocument() && ownerDocument()->transformSource())
98 return ownerDocument()->transformSource()->platformSource();
99 return m_stylesheetDoc;
100}
101
102void XSLStyleSheet::clearDocuments()
103{
104 clearXSLStylesheetDocument();
105
106 for (auto& import : m_children) {
107 if (import->styleSheet())
108 import->styleSheet()->clearDocuments();
109 }
110}
111
112void XSLStyleSheet::clearXSLStylesheetDocument()
113{
114 if (!m_stylesheetDocTaken) {
115 if (m_stylesheetDoc)
116 xmlFreeDoc(m_stylesheetDoc);
117 } else
118 m_stylesheetDocTaken = false;
119 m_stylesheetDoc = nullptr;
120}
121
122CachedResourceLoader* XSLStyleSheet::cachedResourceLoader()
123{
124 Document* document = ownerDocument();
125 if (!document)
126 return nullptr;
127 return &document->cachedResourceLoader();
128}
129
130bool XSLStyleSheet::parseString(const String& string)
131{
132 // Parse in a single chunk into an xmlDocPtr
133 const UChar BOM = 0xFEFF;
134 const unsigned char BOMHighByte = *reinterpret_cast<const unsigned char*>(&BOM);
135 clearXSLStylesheetDocument();
136
137 PageConsoleClient* console = nullptr;
138 Frame* frame = ownerDocument()->frame();
139 if (frame && frame->page())
140 console = &frame->page()->console();
141
142 XMLDocumentParserScope scope(cachedResourceLoader(), XSLTProcessor::genericErrorFunc, XSLTProcessor::parseErrorFunc, console);
143
144 auto upconvertedCharacters = StringView(string).upconvertedCharacters();
145 const char* buffer = reinterpret_cast<const char*>(upconvertedCharacters.get());
146 Checked<unsigned, RecordOverflow> unsignedSize = string.length();
147 unsignedSize *= sizeof(UChar);
148 if (unsignedSize.hasOverflowed() || unsignedSize.unsafeGet() > std::numeric_limits<int>::max())
149 return false;
150
151 int size = static_cast<int>(unsignedSize.unsafeGet());
152 xmlParserCtxtPtr ctxt = xmlCreateMemoryParserCtxt(buffer, size);
153 if (!ctxt)
154 return false;
155
156 if (m_parentStyleSheet && m_parentStyleSheet->m_stylesheetDoc) {
157 // The XSL transform may leave the newly-transformed document
158 // with references to the symbol dictionaries of the style sheet
159 // and any of its children. XML document disposal can corrupt memory
160 // if a document uses more than one symbol dictionary, so we
161 // ensure that all child stylesheets use the same dictionaries as their
162 // parents.
163 xmlDictFree(ctxt->dict);
164 ctxt->dict = m_parentStyleSheet->m_stylesheetDoc->dict;
165 xmlDictReference(ctxt->dict);
166 }
167
168 m_stylesheetDoc = xmlCtxtReadMemory(ctxt, buffer, size,
169 finalURL().string().utf8().data(),
170 BOMHighByte == 0xFF ? "UTF-16LE" : "UTF-16BE",
171 XML_PARSE_NOENT | XML_PARSE_DTDATTR | XML_PARSE_NOWARNING | XML_PARSE_NOCDATA);
172 xmlFreeParserCtxt(ctxt);
173
174 loadChildSheets();
175
176 return !!m_stylesheetDoc;
177}
178
179void XSLStyleSheet::loadChildSheets()
180{
181 if (!document())
182 return;
183
184 xmlNodePtr stylesheetRoot = document()->children;
185
186 // Top level children may include other things such as DTD nodes, we ignore those.
187 while (stylesheetRoot && stylesheetRoot->type != XML_ELEMENT_NODE)
188 stylesheetRoot = stylesheetRoot->next;
189
190 if (m_embedded) {
191 // We have to locate (by ID) the appropriate embedded stylesheet element, so that we can walk the
192 // import/include list.
193 xmlAttrPtr idNode = xmlGetID(document(), (const xmlChar*)(finalURL().string().utf8().data()));
194 if (!idNode)
195 return;
196 stylesheetRoot = idNode->parent;
197 } else {
198 // FIXME: Need to handle an external URI with a # in it. This is a pretty minor edge case, so we'll deal
199 // with it later.
200 }
201
202 if (stylesheetRoot) {
203 // Walk the children of the root element and look for import/include elements.
204 // Imports must occur first.
205 xmlNodePtr curr = stylesheetRoot->children;
206 while (curr) {
207 if (curr->type != XML_ELEMENT_NODE) {
208 curr = curr->next;
209 continue;
210 }
211 if (IS_XSLT_ELEM(curr) && IS_XSLT_NAME(curr, "import")) {
212 xmlChar* uriRef = xsltGetNsProp(curr, (const xmlChar*)"href", XSLT_NAMESPACE);
213 loadChildSheet(String::fromUTF8((const char*)uriRef));
214 xmlFree(uriRef);
215 } else
216 break;
217 curr = curr->next;
218 }
219
220 // Now handle includes.
221 while (curr) {
222 if (curr->type == XML_ELEMENT_NODE && IS_XSLT_ELEM(curr) && IS_XSLT_NAME(curr, "include")) {
223 xmlChar* uriRef = xsltGetNsProp(curr, (const xmlChar*)"href", XSLT_NAMESPACE);
224 loadChildSheet(String::fromUTF8((const char*)uriRef));
225 xmlFree(uriRef);
226 }
227 curr = curr->next;
228 }
229 }
230}
231
232void XSLStyleSheet::loadChildSheet(const String& href)
233{
234 auto childRule = std::make_unique<XSLImportRule>(this, href);
235 m_children.append(childRule.release());
236 m_children.last()->loadSheet();
237}
238
239xsltStylesheetPtr XSLStyleSheet::compileStyleSheet()
240{
241 // FIXME: Hook up error reporting for the stylesheet compilation process.
242 if (m_embedded)
243 return xsltLoadStylesheetPI(document());
244
245 // Certain libxslt versions are corrupting the xmlDoc on compilation
246 // failures - hence attempting to recompile after a failure is unsafe.
247 if (m_compilationFailed)
248 return nullptr;
249
250 // xsltParseStylesheetDoc makes the document part of the stylesheet
251 // so we have to release our pointer to it.
252 ASSERT(m_stylesheetDoc);
253 ASSERT(!m_stylesheetDocTaken);
254 xsltStylesheetPtr result = xsltParseStylesheetDoc(m_stylesheetDoc);
255 if (result)
256 m_stylesheetDocTaken = true;
257 else
258 m_compilationFailed = true;
259 return result;
260}
261
262void XSLStyleSheet::setParentStyleSheet(XSLStyleSheet* parent)
263{
264 m_parentStyleSheet = parent;
265}
266
267Document* XSLStyleSheet::ownerDocument()
268{
269 for (XSLStyleSheet* styleSheet = this; styleSheet; styleSheet = styleSheet->parentStyleSheet()) {
270 Node* node = styleSheet->ownerNode();
271 if (node)
272 return &node->document();
273 }
274 return 0;
275}
276
277xmlDocPtr XSLStyleSheet::locateStylesheetSubResource(xmlDocPtr parentDoc, const xmlChar* uri)
278{
279 bool matchedParent = (parentDoc == document());
280 for (auto& import : m_children) {
281 XSLStyleSheet* child = import->styleSheet();
282 if (!child)
283 continue;
284 if (matchedParent) {
285 if (child->processed())
286 continue; // libxslt has been given this sheet already.
287
288 // Check the URI of the child stylesheet against the doc URI.
289 // In order to ensure that libxml canonicalized both URLs, we get the original href
290 // string from the import rule and canonicalize it using libxml before comparing it
291 // with the URI argument.
292 CString importHref = import->href().utf8();
293 xmlChar* base = xmlNodeGetBase(parentDoc, (xmlNodePtr)parentDoc);
294 xmlChar* childURI = xmlBuildURI((const xmlChar*)importHref.data(), base);
295 bool equalURIs = xmlStrEqual(uri, childURI);
296 xmlFree(base);
297 xmlFree(childURI);
298 if (equalURIs) {
299 child->markAsProcessed();
300 return child->document();
301 }
302 continue;
303 }
304 xmlDocPtr result = import->styleSheet()->locateStylesheetSubResource(parentDoc, uri);
305 if (result)
306 return result;
307 }
308
309 return 0;
310}
311
312void XSLStyleSheet::markAsProcessed()
313{
314 ASSERT(!m_processed);
315 ASSERT(!m_stylesheetDocTaken);
316 m_processed = true;
317 m_stylesheetDocTaken = true;
318}
319
320} // namespace WebCore
321
322#endif // ENABLE(XSLT)
323