shell/source/all/xml_parser.cxx

   1 /*************************************************************************
   2  *
   3  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   4  *
   5  * Copyright 2008 by Sun Microsystems, Inc.
   6  *
   7  * OpenOffice.org - a multi-platform office productivity suite
   8  *
   9  * $RCSfile: xml_parser.cxx,v $
  10  * $Revision: 1.10 $
  11  *
  12  * This file is part of OpenOffice.org.
  13  *
  14  * OpenOffice.org is free software: you can redistribute it and/or modify
  15  * it under the terms of the GNU Lesser General Public License version 3
  16  * only, as published by the Free Software Foundation.
  17  *
  18  * OpenOffice.org is distributed in the hope that it will be useful,
  19  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  20  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  21  * GNU Lesser General Public License version 3 for more details
  22  * (a copy is included in the LICENSE file that accompanied this code).
  23  *
  24  * You should have received a copy of the GNU Lesser General Public License
  25  * version 3 along with OpenOffice.org.  If not, see
  26  * <http://www.openoffice.org/license.html>
  27  * for a copy of the LGPLv3 License.
  28  *
  29  ************************************************************************/
  30
  31 // MARKER(update_precomp.py): autogen include statement, do not remove
  32 #include "precompiled_shell.hxx"
  33
  34 #ifndef XML_PARSER_HXX_INCLUDED
  35 #include "internal/xml_parser.hxx"
  36 #endif
  37 #include "internal/i_xml_parser_event_handler.hxx"
  38
  39 #include <assert.h>
  40
  41 namespace /* private */
  42 {
  43
  44     //######################################################
  45     /*  Extracts the local part of tag without
  46         namespace decoration e.g. meta:creator -> creator */
  47     const XML_Char COLON = (XML_Char)':';
  48
  49     const XML_Char* get_local_name(const XML_Char* rawname)
  50     {
  51         const XML_Char* p = rawname;
  52
  53         // go to the end
  54         while (*p) p++;
  55
  56         // go back until the first ':'
  57         while (*p != COLON && p > rawname)
  58             p--;
  59
  60         // if we are on a colon one step forward
  61         if (*p == COLON)
  62             p++;
  63
  64         return p;
  65     }
  66
  67     //################################################
  68     inline xml_parser* get_parser_instance(void* data)
  69     {
  70         return reinterpret_cast<xml_parser*>(XML_GetUserData(
  71             reinterpret_cast<XML_Parser>(data)));
  72     }
  73
  74     //################################################
  75     bool has_only_whitespaces(const XML_Char* s, int len)
  76     {
  77         const XML_Char* p = s;
  78         for (int i = 0; i < len; i++)
  79             if (*p++ != ' ') return false;
  80         return true;
  81     }
  82 }
  83
  84 //###################################################
  85 xml_parser::xml_parser(const XML_Char* EncodingName) :
  86     document_handler_(0),
  87     xml_parser_(XML_ParserCreate(EncodingName))
  88 {
  89     init();
  90 }
  91
  92 //###################################################
  93 xml_parser::xml_parser(const XML_Char* EncodingName, XML_Char /*NamespaceSeparator*/) :
  94     document_handler_(0),
  95     xml_parser_(XML_ParserCreate(EncodingName))
  96 {
  97     init();
  98 }
  99
 100 //###################################################
 101 xml_parser::~xml_parser()
 102 {
 103     XML_ParserFree(xml_parser_);
 104 }
 105
 106 //###################################################
 107 /* Callback functions will be called by the parser on
 108    different events */
 109
 110 //###################################################
 111 extern "C"
 112 {
 113
 114 static void xml_start_element_handler(void* UserData, const XML_Char* name, const XML_Char** atts)
 115 {
 116     assert(UserData != NULL);
 117
 118     xml_parser* pImpl  = get_parser_instance(UserData);
 119
 120     i_xml_parser_event_handler* pDocHdl = pImpl->get_document_handler();
 121     if (pDocHdl)
 122     {
 123         xml_tag_attribute_container_t attributes;
 124
 125         int i = 0;
 126
 127         while(atts[i])
 128         {
 129             attributes[reinterpret_cast<const char_t*>(get_local_name(atts[i]))] = reinterpret_cast<const char_t*>(atts[i+1]);
 130             i += 2; // skip to next pair
 131         }
 132
 133         pDocHdl->start_element(
 134             reinterpret_cast<const char_t*>(name), reinterpret_cast<const char_t*>(get_local_name(name)), attributes);
 135     }
 136 }
 137
 138 //###################################################
 139 static void xml_end_element_handler(void* UserData, const XML_Char* name)
 140 {
 141     assert(UserData);
 142
 143     xml_parser* pImpl  = get_parser_instance(UserData);
 144     i_xml_parser_event_handler* pDocHdl = pImpl->get_document_handler();
 145     if (pDocHdl)
 146         pDocHdl->end_element(reinterpret_cast<const char_t*>(name), reinterpret_cast<const char_t*>(get_local_name(name)));
 147 }
 148
 149 //###################################################
 150 static void xml_character_data_handler(void* UserData, const XML_Char* s, int len)
 151 {
 152     assert(UserData);
 153
 154     xml_parser* pImpl  = get_parser_instance(UserData);
 155     i_xml_parser_event_handler* pDocHdl = pImpl->get_document_handler();
 156     if (pDocHdl)
 157     {
 158         if (has_only_whitespaces(s,len))
 159             pDocHdl->ignore_whitespace(string_t(reinterpret_cast<const char_t*>(s), len));
 160         else
 161             pDocHdl->characters(string_t(reinterpret_cast<const char_t*>(s), len));
 162     }
 163 }
 164
 165 //###################################################
 166 static void xml_comment_handler(void* UserData, const XML_Char* Data)
 167 {
 168     assert(UserData);
 169
 170     xml_parser* pImpl  = get_parser_instance(UserData);
 171     i_xml_parser_event_handler* pDocHdl = pImpl->get_document_handler();
 172     if (pDocHdl)
 173         pDocHdl->comment(reinterpret_cast<const char_t*>(Data));
 174 }
 175
 176 } // extern "C"
 177
 178 //###################################################
 179 void xml_parser::init()
 180 {
 181     XML_SetUserData(xml_parser_, this);
 182
 183     // we use the parser as handler argument,
 184     // so we could use it if necessary, the
 185     // UserData are usable anyway using
 186     // XML_GetUserData(...)
 187     XML_UseParserAsHandlerArg(xml_parser_);
 188
 189     XML_SetElementHandler(
 190         xml_parser_,
 191         xml_start_element_handler,
 192         xml_end_element_handler);
 193
 194     XML_SetCharacterDataHandler(
 195         xml_parser_,
 196         xml_character_data_handler);
 197
 198     XML_SetCommentHandler(
 199         xml_parser_,
 200         xml_comment_handler);
 201 }
 202
 203 //###################################################
 204 void xml_parser::parse(const char* XmlData, size_t Length, bool IsFinal)
 205 {
 206     if (0 == XML_Parse(xml_parser_, XmlData, Length, IsFinal))
 207         throw xml_parser_exception(
 208             (char*)XML_ErrorString(XML_GetErrorCode(xml_parser_)),
 209             (int)XML_GetErrorCode(xml_parser_),
 210             XML_GetCurrentLineNumber(xml_parser_),
 211             XML_GetCurrentColumnNumber(xml_parser_),
 212             XML_GetCurrentByteIndex(xml_parser_));
 213 }
 214
 215 //###################################################
 216 void xml_parser::set_document_handler(
 217     i_xml_parser_event_handler* event_handler)
 218 {
 219     document_handler_ = event_handler;
 220 }
 221
 222 //###################################################
 223 i_xml_parser_event_handler* xml_parser::get_document_handler() const
 224 {
 225     return document_handler_;
 226 }
 227
 228 //###################################################
 229 void xml_parser::set_encoding(const XML_Char* Encoding)
 230 {
 231     XML_SetEncoding(xml_parser_, Encoding);
 232 }