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