1*8d192041SAndrew Rist /************************************************************** 2cdf0e10cSrcweir * 3*8d192041SAndrew Rist * Licensed to the Apache Software Foundation (ASF) under one 4*8d192041SAndrew Rist * or more contributor license agreements. See the NOTICE file 5*8d192041SAndrew Rist * distributed with this work for additional information 6*8d192041SAndrew Rist * regarding copyright ownership. The ASF licenses this file 7*8d192041SAndrew Rist * to you under the Apache License, Version 2.0 (the 8*8d192041SAndrew Rist * "License"); you may not use this file except in compliance 9*8d192041SAndrew Rist * with the License. You may obtain a copy of the License at 10*8d192041SAndrew Rist * 11*8d192041SAndrew Rist * http://www.apache.org/licenses/LICENSE-2.0 12*8d192041SAndrew Rist * 13*8d192041SAndrew Rist * Unless required by applicable law or agreed to in writing, 14*8d192041SAndrew Rist * software distributed under the License is distributed on an 15*8d192041SAndrew Rist * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY 16*8d192041SAndrew Rist * KIND, either express or implied. See the License for the 17*8d192041SAndrew Rist * specific language governing permissions and limitations 18*8d192041SAndrew Rist * under the License. 19*8d192041SAndrew Rist * 20*8d192041SAndrew Rist *************************************************************/ 21*8d192041SAndrew Rist 22*8d192041SAndrew Rist 23cdf0e10cSrcweir 24cdf0e10cSrcweir #ifndef _SAX_FASTPARSER_HXX_ 25cdf0e10cSrcweir #define _SAX_FASTPARSER_HXX_ 26cdf0e10cSrcweir 27cdf0e10cSrcweir #include <vector> 28cdf0e10cSrcweir #include <stack> 29cdf0e10cSrcweir #include <hash_map> 30cdf0e10cSrcweir #include <boost/shared_ptr.hpp> 31cdf0e10cSrcweir #include <rtl/ref.hxx> 32cdf0e10cSrcweir #include <com/sun/star/xml/sax/XFastParser.hpp> 33cdf0e10cSrcweir #include <com/sun/star/xml/sax/XFastTokenHandler.hpp> 34cdf0e10cSrcweir #include <com/sun/star/xml/sax/XFastDocumentHandler.hpp> 35cdf0e10cSrcweir #include <com/sun/star/lang/XServiceInfo.hpp> 36cdf0e10cSrcweir #include <cppuhelper/implbase2.hxx> 37cdf0e10cSrcweir 38cdf0e10cSrcweir #include <expat.h> 39cdf0e10cSrcweir #include "xml2utf.hxx" 40cdf0e10cSrcweir 41cdf0e10cSrcweir #include <sax/fastattribs.hxx> 42cdf0e10cSrcweir 43cdf0e10cSrcweir #define PARSER_IMPLEMENTATION_NAME "com.sun.star.comp.extensions.xml.sax.FastParser" 44cdf0e10cSrcweir #define PARSER_SERVICE_NAME "com.sun.star.xml.sax.FastParser" 45cdf0e10cSrcweir 46cdf0e10cSrcweir namespace sax_fastparser { 47cdf0e10cSrcweir 48cdf0e10cSrcweir class FastLocatorImpl; 49cdf0e10cSrcweir struct NamespaceDefine; 50cdf0e10cSrcweir struct SaxContextImpl; 51cdf0e10cSrcweir 52cdf0e10cSrcweir typedef ::boost::shared_ptr< SaxContextImpl > SaxContextImplPtr; 53cdf0e10cSrcweir typedef ::boost::shared_ptr< NamespaceDefine > NamespaceDefineRef; 54cdf0e10cSrcweir 55cdf0e10cSrcweir typedef ::std::hash_map< ::rtl::OUString, sal_Int32, 56cdf0e10cSrcweir ::rtl::OUStringHash, ::std::equal_to< ::rtl::OUString > > NamespaceMap; 57cdf0e10cSrcweir 58cdf0e10cSrcweir // -------------------------------------------------------------------- 59cdf0e10cSrcweir 60cdf0e10cSrcweir struct ParserData 61cdf0e10cSrcweir { 62cdf0e10cSrcweir ::com::sun::star::uno::Reference< ::com::sun::star::xml::sax::XFastDocumentHandler > mxDocumentHandler; 63cdf0e10cSrcweir ::com::sun::star::uno::Reference< ::com::sun::star::xml::sax::XFastTokenHandler > mxTokenHandler; 64cdf0e10cSrcweir ::com::sun::star::uno::Reference< ::com::sun::star::xml::sax::XErrorHandler > mxErrorHandler; 65cdf0e10cSrcweir ::com::sun::star::uno::Reference< ::com::sun::star::xml::sax::XEntityResolver > mxEntityResolver; 66cdf0e10cSrcweir ::com::sun::star::lang::Locale maLocale; 67cdf0e10cSrcweir 68cdf0e10cSrcweir ParserData(); 69cdf0e10cSrcweir ~ParserData(); 70cdf0e10cSrcweir }; 71cdf0e10cSrcweir 72cdf0e10cSrcweir // -------------------------------------------------------------------- 73cdf0e10cSrcweir 74cdf0e10cSrcweir // Entity binds all information needed for a single file 75cdf0e10cSrcweir struct Entity : public ParserData 76cdf0e10cSrcweir { 77cdf0e10cSrcweir ::com::sun::star::xml::sax::InputSource maStructSource; 78cdf0e10cSrcweir XML_Parser mpParser; 79cdf0e10cSrcweir ::sax_expatwrap::XMLFile2UTFConverter maConverter; 80cdf0e10cSrcweir ::rtl::Reference< FastAttributeList > mxAttributes; 81cdf0e10cSrcweir 82cdf0e10cSrcweir // Exceptions cannot be thrown through the C-XmlParser (possible resource leaks), 83cdf0e10cSrcweir // therefore the exception must be saved somewhere. 84cdf0e10cSrcweir ::com::sun::star::uno::Any maSavedException; 85cdf0e10cSrcweir 86cdf0e10cSrcweir ::std::stack< SaxContextImplPtr > maContextStack; 87cdf0e10cSrcweir ::std::vector< NamespaceDefineRef > maNamespaceDefines; 88cdf0e10cSrcweir 89cdf0e10cSrcweir explicit Entity( const ParserData& rData ); 90cdf0e10cSrcweir ~Entity(); 91cdf0e10cSrcweir }; 92cdf0e10cSrcweir 93cdf0e10cSrcweir // -------------------------------------------------------------------- 94cdf0e10cSrcweir 95cdf0e10cSrcweir // This class implements the external Parser interface 96cdf0e10cSrcweir class FastSaxParser : public ::cppu::WeakImplHelper2< ::com::sun::star::xml::sax::XFastParser, ::com::sun::star::lang::XServiceInfo > 97cdf0e10cSrcweir { 98cdf0e10cSrcweir public: 99cdf0e10cSrcweir FastSaxParser(); 100cdf0e10cSrcweir virtual ~FastSaxParser(); 101cdf0e10cSrcweir 102cdf0e10cSrcweir // The implementation details 103cdf0e10cSrcweir static ::com::sun::star::uno::Sequence< ::rtl::OUString > getSupportedServiceNames_Static(void); 104cdf0e10cSrcweir 105cdf0e10cSrcweir // XFastParser 106cdf0e10cSrcweir virtual void SAL_CALL parseStream( const ::com::sun::star::xml::sax::InputSource& aInputSource ) throw (::com::sun::star::xml::sax::SAXException, ::com::sun::star::io::IOException, ::com::sun::star::uno::RuntimeException); 107cdf0e10cSrcweir virtual void SAL_CALL setFastDocumentHandler( const ::com::sun::star::uno::Reference< ::com::sun::star::xml::sax::XFastDocumentHandler >& Handler ) throw (::com::sun::star::uno::RuntimeException); 108cdf0e10cSrcweir virtual void SAL_CALL setTokenHandler( const ::com::sun::star::uno::Reference< ::com::sun::star::xml::sax::XFastTokenHandler >& Handler ) throw (::com::sun::star::uno::RuntimeException); 109cdf0e10cSrcweir virtual void SAL_CALL registerNamespace( const ::rtl::OUString& NamespaceURL, sal_Int32 NamespaceToken ) throw (::com::sun::star::lang::IllegalArgumentException, ::com::sun::star::uno::RuntimeException); 110cdf0e10cSrcweir virtual void SAL_CALL setErrorHandler( const ::com::sun::star::uno::Reference< ::com::sun::star::xml::sax::XErrorHandler >& Handler ) throw (::com::sun::star::uno::RuntimeException); 111cdf0e10cSrcweir virtual void SAL_CALL setEntityResolver( const ::com::sun::star::uno::Reference< ::com::sun::star::xml::sax::XEntityResolver >& Resolver ) throw (::com::sun::star::uno::RuntimeException); 112cdf0e10cSrcweir virtual void SAL_CALL setLocale( const ::com::sun::star::lang::Locale& rLocale ) throw (::com::sun::star::uno::RuntimeException); 113cdf0e10cSrcweir 114cdf0e10cSrcweir // XServiceInfo 115cdf0e10cSrcweir virtual ::rtl::OUString SAL_CALL getImplementationName( ) throw (::com::sun::star::uno::RuntimeException); 116cdf0e10cSrcweir virtual sal_Bool SAL_CALL supportsService( const ::rtl::OUString& ServiceName ) throw (::com::sun::star::uno::RuntimeException); 117cdf0e10cSrcweir virtual ::com::sun::star::uno::Sequence< ::rtl::OUString > SAL_CALL getSupportedServiceNames( ) throw (::com::sun::star::uno::RuntimeException); 118cdf0e10cSrcweir 119cdf0e10cSrcweir // called by the C callbacks of the expat parser 120cdf0e10cSrcweir void callbackStartElement( const XML_Char* name, const XML_Char** atts ); 121cdf0e10cSrcweir void callbackEndElement( const XML_Char* name ); 122cdf0e10cSrcweir void callbackCharacters( const XML_Char* s, int nLen ); 123cdf0e10cSrcweir int callbackExternalEntityRef( XML_Parser parser, const XML_Char *openEntityNames, const XML_Char *base, const XML_Char *systemId, const XML_Char *publicId); 124cdf0e10cSrcweir pushEntity(const Entity & rEntity)125cdf0e10cSrcweir inline void pushEntity( const Entity& rEntity ) { maEntities.push( rEntity ); } popEntity()126cdf0e10cSrcweir inline void popEntity() { maEntities.pop(); } getEntity()127cdf0e10cSrcweir Entity& getEntity() { return maEntities.top(); } 128cdf0e10cSrcweir 129cdf0e10cSrcweir private: 130cdf0e10cSrcweir void parse(); 131cdf0e10cSrcweir 132cdf0e10cSrcweir sal_Int32 GetToken( const ::rtl::OString& rToken ); 133cdf0e10cSrcweir sal_Int32 GetToken( const sal_Char* pToken, sal_Int32 nTokenLen = 0 ); 134cdf0e10cSrcweir sal_Int32 GetTokenWithPrefix( const ::rtl::OString& rPrefix, const ::rtl::OString& rName ) throw (::com::sun::star::xml::sax::SAXException); 135cdf0e10cSrcweir sal_Int32 GetTokenWithPrefix( const sal_Char*pPrefix, int nPrefixLen, const sal_Char* pName, int nNameLen ) throw (::com::sun::star::xml::sax::SAXException); 136cdf0e10cSrcweir ::rtl::OUString GetNamespaceURL( const ::rtl::OString& rPrefix ) throw (::com::sun::star::xml::sax::SAXException); 137cdf0e10cSrcweir ::rtl::OUString GetNamespaceURL( const sal_Char*pPrefix, int nPrefixLen ) throw (::com::sun::star::xml::sax::SAXException); 138cdf0e10cSrcweir sal_Int32 GetNamespaceToken( const ::rtl::OUString& rNamespaceURL ); 139cdf0e10cSrcweir sal_Int32 GetTokenWithNamespaceURL( const ::rtl::OUString& rNamespaceURL, const sal_Char* pName, int nNameLen ); 140cdf0e10cSrcweir void DefineNamespace( const ::rtl::OString& rPrefix, const sal_Char* pNamespaceURL ); 141cdf0e10cSrcweir sal_Int32 CreateCustomToken( const sal_Char* pToken, int len = 0 ); 142cdf0e10cSrcweir 143cdf0e10cSrcweir void pushContext(); 144cdf0e10cSrcweir void popContext(); 145cdf0e10cSrcweir 146cdf0e10cSrcweir void splitName( const XML_Char *pwName, const XML_Char *&rpPrefix, sal_Int32 &rPrefixLen, const XML_Char *&rpName, sal_Int32 &rNameLen ); 147cdf0e10cSrcweir 148cdf0e10cSrcweir private: 149cdf0e10cSrcweir ::osl::Mutex maMutex; 150cdf0e10cSrcweir 151cdf0e10cSrcweir ::rtl::Reference< FastLocatorImpl > mxDocumentLocator; 152cdf0e10cSrcweir NamespaceMap maNamespaceMap; 153cdf0e10cSrcweir 154cdf0e10cSrcweir ParserData maData; /// Cached parser configuration for next call of parseStream(). 155cdf0e10cSrcweir ::std::stack< Entity > maEntities; /// Entity stack for each call of parseStream(). 156cdf0e10cSrcweir }; 157cdf0e10cSrcweir 158cdf0e10cSrcweir } 159cdf0e10cSrcweir 160cdf0e10cSrcweir #endif // _SAX_FASTPARSER_HXX_ 161