2 * Licensed to the University Corporation for Advanced Internet
3 * Development, Inc. (UCAID) under one or more contributor license
4 * agreements. See the NOTICE file distributed with this work for
5 * additional information regarding copyright ownership.
7 * UCAID licenses this file to you under the Apache License,
8 * Version 2.0 (the "License"); you may not use this file except
9 * in compliance with the License. You may obtain a copy of the
12 * http://www.apache.org/licenses/LICENSE-2.0
14 * Unless required by applicable law or agreed to in writing,
15 * software distributed under the License is distributed on an
16 * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND,
17 * either express or implied. See the License for the specific
18 * language governing permissions and limitations under the License.
24 * A thread-safe pool of parsers that share characteristics.
28 #include "exceptions.h"
30 #include "util/CurlURLInputStream.h"
32 #include "util/ParserPool.h"
33 #include "util/Threads.h"
34 #include "util/XMLHelper.h"
36 #include <sys/types.h>
40 #include <boost/algorithm/string.hpp>
41 #include <boost/bind.hpp>
42 #include <boost/tokenizer.hpp>
43 #include <xercesc/util/PlatformUtils.hpp>
44 #include <xercesc/util/XMLUniDefs.hpp>
45 #include <xercesc/sax/SAXException.hpp>
46 #include <xercesc/framework/MemBufInputSource.hpp>
47 #include <xercesc/framework/LocalFileInputSource.hpp>
48 #include <xercesc/framework/Wrapper4InputSource.hpp>
50 using namespace xmltooling::logging;
51 using namespace xmltooling;
52 using namespace xercesc;
53 using namespace boost;
58 class MyErrorHandler : public DOMErrorHandler {
62 MyErrorHandler() : errors(0) {}
64 bool handleError(const DOMError& e)
67 xmltooling::NDC ndc("handleError");
69 Category& log=Category::getInstance(XMLTOOLING_LOGCAT".ParserPool");
71 DOMLocator* locator=e.getLocation();
72 auto_ptr_char temp(e.getMessage());
74 switch (e.getSeverity()) {
75 case DOMError::DOM_SEVERITY_WARNING:
76 log.warnStream() << "warning on line " << locator->getLineNumber()
77 << ", column " << locator->getColumnNumber()
78 << ", message: " << temp.get() << logging::eol;
81 case DOMError::DOM_SEVERITY_ERROR:
83 log.errorStream() << "error on line " << locator->getLineNumber()
84 << ", column " << locator->getColumnNumber()
85 << ", message: " << temp.get() << logging::eol;
88 case DOMError::DOM_SEVERITY_FATAL_ERROR:
90 log.errorStream() << "fatal error on line " << locator->getLineNumber()
91 << ", column " << locator->getColumnNumber()
92 << ", message: " << temp.get() << logging::eol;
97 log.errorStream() << "undefined error type on line " << locator->getLineNumber()
98 << ", column " << locator->getColumnNumber()
99 << ", message: " << temp.get() << logging::eol;
106 ParserPool::ParserPool(bool namespaceAware, bool schemaAware)
107 : m_namespaceAware(namespaceAware), m_schemaAware(schemaAware), m_lock(Mutex::create()), m_security(new SecurityManager()) {}
109 ParserPool::~ParserPool()
111 while(!m_pool.empty()) {
112 m_pool.top()->release();
117 DOMDocument* ParserPool::newDocument()
119 return DOMImplementationRegistry::getDOMImplementation(nullptr)->createDocument();
122 #ifdef XMLTOOLING_XERCESC_COMPLIANT_DOMLS
124 DOMDocument* ParserPool::parse(DOMLSInput& domsrc)
126 DOMLSParser* parser=checkoutBuilder();
127 XercesJanitor<DOMLSParser> janitor(parser);
130 parser->getDomConfig()->setParameter(XMLUni::fgDOMErrorHandler, dynamic_cast<DOMErrorHandler*>(&deh));
131 DOMDocument* doc=parser->parse(&domsrc);
135 throw XMLParserException("XML error(s) during parsing, check log for specifics");
137 parser->getDomConfig()->setParameter(XMLUni::fgDOMErrorHandler, (void*)nullptr);
138 parser->getDomConfig()->setParameter(XMLUni::fgXercesUserAdoptsDOMDocument, true);
139 checkinBuilder(janitor.release());
142 catch (XMLException& ex) {
143 parser->getDomConfig()->setParameter(XMLUni::fgDOMErrorHandler, (void*)nullptr);
144 parser->getDomConfig()->setParameter(XMLUni::fgXercesUserAdoptsDOMDocument, true);
145 checkinBuilder(janitor.release());
146 auto_ptr_char temp(ex.getMessage());
147 throw XMLParserException(string("Xerces error during parsing: ") + (temp.get() ? temp.get() : "no message"));
149 catch (XMLToolingException&) {
150 parser->getDomConfig()->setParameter(XMLUni::fgDOMErrorHandler, (void*)nullptr);
151 parser->getDomConfig()->setParameter(XMLUni::fgXercesUserAdoptsDOMDocument, true);
152 checkinBuilder(janitor.release());
159 DOMDocument* ParserPool::parse(DOMInputSource& domsrc)
161 DOMBuilder* parser=checkoutBuilder();
162 XercesJanitor<DOMBuilder> janitor(parser);
165 parser->setErrorHandler(&deh);
166 DOMDocument* doc=parser->parse(domsrc);
170 throw XMLParserException("XML error(s) during parsing, check log for specifics");
172 parser->setErrorHandler(nullptr);
173 parser->setFeature(XMLUni::fgXercesUserAdoptsDOMDocument, true);
174 checkinBuilder(janitor.release());
177 catch (XMLException& ex) {
178 parser->setErrorHandler(nullptr);
179 parser->setFeature(XMLUni::fgXercesUserAdoptsDOMDocument, true);
180 checkinBuilder(janitor.release());
181 auto_ptr_char temp(ex.getMessage());
182 throw XMLParserException(string("Xerces error during parsing: ") + (temp.get() ? temp.get() : "no message"));
184 catch (XMLToolingException&) {
185 parser->setErrorHandler(nullptr);
186 parser->setFeature(XMLUni::fgXercesUserAdoptsDOMDocument, true);
187 checkinBuilder(janitor.release());
194 DOMDocument* ParserPool::parse(istream& is)
196 StreamInputSource src(is);
197 Wrapper4InputSource domsrc(&src,false);
198 return parse(domsrc);
201 // Functor to double its argument separated by a character and append to a buffer
202 template <class T> class doubleit {
204 doubleit(T& t, const typename T::value_type& s) : temp(t), sep(s) {}
205 void operator() (const pair<const T,T>& s) { temp += s.first + sep + s.first + sep; }
207 const typename T::value_type& sep;
210 bool ParserPool::loadSchema(const XMLCh* nsURI, const XMLCh* pathname)
212 // Just check the pathname and then directly register the pair into the map.
214 auto_ptr_char p(pathname);
216 struct _stat stat_buf;
217 if (_stat(p.get(), &stat_buf) != 0)
219 struct stat stat_buf;
220 if (stat(p.get(), &stat_buf) != 0)
224 xmltooling::NDC ndc("loadSchema");
226 Category& log=Category::getInstance(XMLTOOLING_LOGCAT".ParserPool");
227 auto_ptr_char n(nsURI);
228 log.error("failed to load schema for (%s), file not found (%s)",n.get(),p.get());
233 m_schemaLocMap[nsURI]=pathname;
234 m_schemaLocations.erase();
235 for_each(m_schemaLocMap.begin(), m_schemaLocMap.end(), doubleit<xstring>(m_schemaLocations,chSpace));
240 bool ParserPool::loadCatalogs(const char* pathnames)
242 string temp(pathnames);
243 boost::tokenizer< char_separator<char> > catpaths(temp, char_separator<char>(PATH_SEPARATOR_STR));
245 catpaths.begin(), catpaths.end(),
246 // Call loadCatalog with an inner call to s->c_str() on each entry.
247 boost::bind(static_cast<bool (ParserPool::*)(const char*)>(&ParserPool::loadCatalog), this, boost::bind(&string::c_str, _1))
249 return catpaths.begin() != catpaths.end();
252 bool ParserPool::loadCatalog(const char* pathname)
254 auto_ptr_XMLCh temp(pathname);
255 return loadCatalog(temp.get());
258 bool ParserPool::loadCatalog(const XMLCh* pathname)
261 xmltooling::NDC ndc("loadCatalog");
263 Category& log=Category::getInstance(XMLTOOLING_LOGCAT".ParserPool");
266 static const XMLCh catalog[] = UNICODE_LITERAL_7(c,a,t,a,l,o,g);
267 static const XMLCh system[] = UNICODE_LITERAL_6(s,y,s,t,e,m);
268 static const XMLCh systemId[] = UNICODE_LITERAL_8(s,y,s,t,e,m,I,d);
269 static const XMLCh uri[] = UNICODE_LITERAL_3(u,r,i);
270 static const XMLCh CATALOG_NS[] = {
271 chLatin_u, chLatin_r, chLatin_n, chColon,
272 chLatin_o, chLatin_a, chLatin_s, chLatin_i, chLatin_s, chColon,
273 chLatin_n, chLatin_a, chLatin_m, chLatin_e, chLatin_s, chColon,
274 chLatin_t, chLatin_c, chColon,
275 chLatin_e, chLatin_n, chLatin_t, chLatin_i, chLatin_t, chLatin_y, chColon,
276 chLatin_x, chLatin_m, chLatin_l, chLatin_n, chLatin_s, chColon,
277 chLatin_x, chLatin_m, chLatin_l, chColon,
278 chLatin_c, chLatin_a, chLatin_t, chLatin_a, chLatin_l, chLatin_o, chLatin_g, chNull
281 // Parse the catalog with the internal parser pool.
283 if (log.isDebugEnabled()) {
284 auto_ptr_char temp(pathname);
285 log.debug("loading XML catalog from %s", temp.get());
288 LocalFileInputSource fsrc(nullptr,pathname);
289 Wrapper4InputSource domsrc(&fsrc,false);
291 DOMDocument* doc=XMLToolingConfig::getConfig().getParser().parse(domsrc);
292 XercesJanitor<DOMDocument> janitor(doc);
294 // Check root element.
295 const DOMElement* root=doc->getDocumentElement();
296 if (!XMLHelper::isNodeNamed(root,CATALOG_NS,catalog)) {
297 auto_ptr_char temp(pathname);
298 log.error("unknown root element, failed to load XML catalog from %s", temp.get());
302 // Fetch all the <system> elements.
303 DOMNodeList* mappings=root->getElementsByTagNameNS(CATALOG_NS,system);
305 for (XMLSize_t i=0; i<mappings->getLength(); i++) {
306 root=static_cast<DOMElement*>(mappings->item(i));
307 const XMLCh* from=root->getAttributeNS(nullptr,systemId);
308 const XMLCh* to=root->getAttributeNS(nullptr,uri);
309 m_schemaLocMap[from]=to;
311 m_schemaLocations.erase();
312 for_each(m_schemaLocMap.begin(), m_schemaLocMap.end(), doubleit<xstring>(m_schemaLocations,chSpace));
314 catch (std::exception& e) {
315 log.error("catalog loader caught exception: %s", e.what());
322 #ifdef XMLTOOLING_XERCESC_COMPLIANT_DOMLS
323 DOMLSInput* ParserPool::resolveResource(
324 const XMLCh *const resourceType,
325 const XMLCh *const namespaceUri,
326 const XMLCh *const publicId,
327 const XMLCh *const systemId,
328 const XMLCh *const baseURI
331 DOMInputSource* ParserPool::resolveEntity(
332 const XMLCh* const publicId, const XMLCh* const systemId, const XMLCh* const baseURI
337 xmltooling::NDC ndc("resolveEntity");
341 xstring sysId(systemId);
343 Category& log=Category::getInstance(XMLTOOLING_LOGCAT".ParserPool");
344 if (log.isDebugEnabled()) {
345 auto_ptr_char sysId(systemId);
346 auto_ptr_char base(baseURI);
347 log.debug("asked to resolve %s with baseURI %s",sysId.get(),base.get() ? base.get() : "(null)");
350 // Find well-known schemas in the specified location.
351 map<xstring,xstring>::const_iterator i = m_schemaLocMap.find(sysId);
352 if (i != m_schemaLocMap.end())
353 return new Wrapper4InputSource(new LocalFileInputSource(baseURI, i->second.c_str()));
355 // Check for entity as a suffix of a value in the map.
356 bool (*p_ends_with)(const xstring&, const xstring&) = ends_with;
358 m_schemaLocMap.begin(), m_schemaLocMap.end(),
359 boost::bind(p_ends_with, boost::bind(&map<xstring,xstring>::value_type::second, _1), boost::ref(sysId))
361 if (i != m_schemaLocMap.end())
362 return new Wrapper4InputSource(new LocalFileInputSource(baseURI, i->second.c_str()));
364 // We'll allow anything without embedded slashes.
365 if (XMLString::indexOf(systemId, chForwardSlash) == -1 && XMLString::indexOf(systemId, chBackSlash) == -1)
366 return new Wrapper4InputSource(new LocalFileInputSource(baseURI, systemId));
368 // Shortcircuit the request.
369 auto_ptr_char temp(systemId);
370 log.debug("unauthorized entity request (%s), blocking it", temp.get());
371 static const XMLByte nullbuf[] = {0};
372 return new Wrapper4InputSource(new MemBufInputSource(nullbuf, 0, systemId));
375 #ifdef XMLTOOLING_XERCESC_COMPLIANT_DOMLS
377 DOMLSParser* ParserPool::createBuilder()
379 static const XMLCh impltype[] = { chLatin_L, chLatin_S, chNull };
380 DOMImplementation* impl=DOMImplementationRegistry::getDOMImplementation(impltype);
381 DOMLSParser* parser=static_cast<DOMImplementationLS*>(impl)->createLSParser(DOMImplementationLS::MODE_SYNCHRONOUS,nullptr);
382 parser->getDomConfig()->setParameter(XMLUni::fgDOMNamespaces, m_namespaceAware);
384 parser->getDomConfig()->setParameter(XMLUni::fgDOMNamespaces, true);
385 parser->getDomConfig()->setParameter(XMLUni::fgXercesSchema, true);
386 parser->getDomConfig()->setParameter(XMLUni::fgDOMValidate, true);
387 parser->getDomConfig()->setParameter(XMLUni::fgXercesCacheGrammarFromParse, true);
389 // We build a "fake" schema location hint that binds each namespace to itself.
390 // This ensures the entity resolver will be given the namespace as a systemId it can check.
391 parser->getDomConfig()->setParameter(XMLUni::fgXercesSchemaExternalSchemaLocation, const_cast<XMLCh*>(m_schemaLocations.c_str()));
393 parser->getDomConfig()->setParameter(XMLUni::fgXercesUserAdoptsDOMDocument, true);
394 parser->getDomConfig()->setParameter(XMLUni::fgXercesDisableDefaultEntityResolution, true);
395 parser->getDomConfig()->setParameter(XMLUni::fgDOMResourceResolver, dynamic_cast<DOMLSResourceResolver*>(this));
396 parser->getDomConfig()->setParameter(XMLUni::fgXercesSecurityManager, m_security.get());
400 DOMLSParser* ParserPool::checkoutBuilder()
403 if (m_pool.empty()) {
404 DOMLSParser* builder=createBuilder();
407 DOMLSParser* p=m_pool.top();
410 p->getDomConfig()->setParameter(XMLUni::fgXercesSchemaExternalSchemaLocation, const_cast<XMLCh*>(m_schemaLocations.c_str()));
414 void ParserPool::checkinBuilder(DOMLSParser* builder)
418 m_pool.push(builder);
424 DOMBuilder* ParserPool::createBuilder()
426 static const XMLCh impltype[] = { chLatin_L, chLatin_S, chNull };
427 DOMImplementation* impl=DOMImplementationRegistry::getDOMImplementation(impltype);
428 DOMBuilder* parser=static_cast<DOMImplementationLS*>(impl)->createDOMBuilder(DOMImplementationLS::MODE_SYNCHRONOUS,0);
429 parser->setFeature(XMLUni::fgDOMNamespaces, m_namespaceAware);
431 parser->setFeature(XMLUni::fgDOMNamespaces, true);
432 parser->setFeature(XMLUni::fgXercesSchema, true);
433 parser->setFeature(XMLUni::fgDOMValidation, true);
434 parser->setFeature(XMLUni::fgXercesCacheGrammarFromParse, true);
436 // We build a "fake" schema location hint that binds each namespace to itself.
437 // This ensures the entity resolver will be given the namespace as a systemId it can check.
438 parser->setProperty(XMLUni::fgXercesSchemaExternalSchemaLocation,const_cast<XMLCh*>(m_schemaLocations.c_str()));
440 parser->setProperty(XMLUni::fgXercesSecurityManager, m_security.get());
441 parser->setFeature(XMLUni::fgXercesUserAdoptsDOMDocument, true);
442 parser->setFeature(XMLUni::fgXercesDisableDefaultEntityResolution, true);
443 parser->setEntityResolver(this);
447 DOMBuilder* ParserPool::checkoutBuilder()
450 if (m_pool.empty()) {
451 DOMBuilder* builder=createBuilder();
454 DOMBuilder* p=m_pool.top();
457 p->setProperty(XMLUni::fgXercesSchemaExternalSchemaLocation,const_cast<XMLCh*>(m_schemaLocations.c_str()));
461 void ParserPool::checkinBuilder(DOMBuilder* builder)
465 m_pool.push(builder);
471 StreamInputSource::StreamInputSource(istream& is, const char* systemId) : InputSource(systemId), m_is(is)
475 BinInputStream* StreamInputSource::makeStream() const
477 return new StreamBinInputStream(m_is);
480 StreamInputSource::StreamBinInputStream::StreamBinInputStream(istream& is) : m_is(is), m_pos(0)
484 #ifdef XMLTOOLING_XERCESC_64BITSAFE
489 StreamInputSource::StreamBinInputStream::curPos() const
494 #ifdef XMLTOOLING_XERCESC_64BITSAFE
495 const XMLCh* StreamInputSource::StreamBinInputStream::getContentType() const
501 xsecsize_t StreamInputSource::StreamBinInputStream::readBytes(XMLByte* const toFill, const xsecsize_t maxToRead)
503 XMLByte* target=toFill;
504 xsecsize_t bytes_read=0,request=maxToRead;
506 // Fulfill the rest by reading from the stream.
507 if (request && !m_is.eof() && !m_is.fail()) {
509 m_is.read(reinterpret_cast<char* const>(target),request);
510 m_pos+=m_is.gcount();
511 bytes_read+=m_is.gcount();
513 catch(ios_base::failure& e) {
514 Category::getInstance(XMLTOOLING_LOGCAT".StreamInputSource").critStream()
515 << "XML::StreamInputSource::StreamBinInputStream::readBytes caught an exception: " << e.what()
524 #ifdef XMLTOOLING_LITE
526 URLInputSource::URLInputSource(const XMLCh* url, const char* systemId, string* cacheTag) : InputSource(systemId), m_url(url)
530 URLInputSource::URLInputSource(const DOMElement* e, const char* systemId, string* cacheTag) : InputSource(systemId)
532 static const XMLCh uri[] = UNICODE_LITERAL_3(u,r,i);
533 static const XMLCh url[] = UNICODE_LITERAL_3(u,r,l);
535 const XMLCh* attr = e->getAttributeNS(nullptr, url);
536 if (!attr || !*attr) {
537 attr = e->getAttributeNS(nullptr, uri);
539 throw IOException("No URL supplied via DOM to URLInputSource constructor.");
545 BinInputStream* URLInputSource::makeStream() const
547 // Ask the URL to create us an appropriate input stream
548 return m_url.makeNewStream();
553 URLInputSource::URLInputSource(const XMLCh* url, const char* systemId, string* cacheTag)
554 : InputSource(systemId), m_cacheTag(cacheTag), m_url(url), m_root(nullptr)
558 URLInputSource::URLInputSource(const DOMElement* e, const char* systemId, string* cacheTag)
559 : InputSource(systemId), m_cacheTag(cacheTag), m_root(e)
563 BinInputStream* URLInputSource::makeStream() const
565 return m_root ? new CurlURLInputStream(m_root, m_cacheTag) : new CurlURLInputStream(m_url.get(), m_cacheTag);
570 const char URLInputSource::asciiStatusCodeElementName[] = "URLInputSourceStatus";
572 const XMLCh URLInputSource::utf16StatusCodeElementName[] = UNICODE_LITERAL_20(U,R,L,I,n,p,u,t,S,o,u,r,c,e,S,t,a,t,u,s);