src/esi/Libxml2Parser.cc

   1 /*
   2  * Copyright (C) 1996-2015 The Squid Software Foundation and contributors
   3  *
   4  * Squid software is distributed under GPLv2+ license and includes
   5  * contributions from numerous individuals and organizations.
   6  * Please see the COPYING and CONTRIBUTORS files for details.
   7  */
   8
   9 /*
  10  * The ESI Libxml2 parser is Copyright (c) 2004 by Joachim Bauch
  11  * http://www.joachim-bauch.de
  12  * mail@joachim-bauch.de
  13  */
  14
  15 #include "squid.h"
  16
  17 #if USE_SQUID_ESI && HAVE_LIBXML2
  18
  19 #include "esi/Libxml2Parser.h"
  20
  21 // the global document that will store the resolved entity
  22 // definitions
  23 static htmlDocPtr entity_doc = NULL;
  24
  25 EsiParserDefinition(ESILibxml2Parser);
  26
  27 // the SAX callback functions
  28 void esi_startElementSAXFunc(void * ctx, const xmlChar * name, const xmlChar ** atts)
  29 {
  30     int count=0;
  31     xmlChar **tmp = (xmlChar **)atts;
  32
  33     while (tmp && *tmp != NULL) {
  34         ++count;
  35         ++tmp;
  36     }
  37
  38     // we increased on every key and value
  39     count /= 2;
  40
  41     ESILibxml2Parser *p = (ESILibxml2Parser *)ctx;
  42
  43     p->getClient()->start((const char *)name, (const char **)atts, count);
  44 }
  45
  46 void esi_endElementSAXFunc(void * ctx, const xmlChar * name)
  47 {
  48     ESILibxml2Parser *p = (ESILibxml2Parser *)ctx;
  49     p->getClient()->end((const char *)name);
  50 }
  51
  52 void esi_commentSAXFunc(void * ctx, const xmlChar * value)
  53 {
  54     ESILibxml2Parser *p = (ESILibxml2Parser *)ctx;
  55     p->getClient()->parserComment((const char *)value);
  56 }
  57
  58 void esi_charactersSAXFunc(void *ctx, const xmlChar *ch, int len)
  59 {
  60     ESILibxml2Parser *p = (ESILibxml2Parser *)ctx;
  61     p->getClient()->parserDefault((const char *)ch, len);
  62 }
  63
  64 xmlEntityPtr esi_getEntitySAXFunc(void * ctx,  const xmlChar * name)
  65 {
  66     xmlEntityPtr res = xmlGetDocEntity(entity_doc, name);
  67
  68     if (res == NULL) {
  69         const htmlEntityDesc *ent = htmlEntityLookup(name);
  70
  71         if (ent != NULL) {
  72             char tmp[32];
  73             snprintf(tmp, 32, "&#%d;", ent->value);
  74             res = xmlAddDocEntity(entity_doc, (const xmlChar *)name, XML_INTERNAL_GENERAL_ENTITY, NULL, NULL, (const xmlChar *)tmp);
  75         }
  76     }
  77
  78     return res;
  79 }
  80
  81 ESILibxml2Parser::ESILibxml2Parser(ESIParserClient *aClient) : theClient (aClient)
  82 {
  83     xmlSAXHandler sax;
  84     htmlDefaultSAXHandlerInit();
  85     memset(&sax, 0, sizeof(sax));
  86     sax.startElement = esi_startElementSAXFunc;
  87     sax.endElement = esi_endElementSAXFunc;
  88     sax.comment = esi_commentSAXFunc;
  89     sax.characters = esi_charactersSAXFunc;
  90     sax.getEntity = esi_getEntitySAXFunc;
  91
  92     /* TODO: grab the document encoding from the headers */
  93     parser = xmlCreatePushParserCtxt(&sax, static_cast<void *>(this), NULL, 0, NULL);
  94     xmlSetFeature(parser, "substitute entities", 0);
  95
  96     if (entity_doc == NULL)
  97         entity_doc = htmlNewDoc(NULL, NULL);
  98 }
  99
 100 ESILibxml2Parser::~ESILibxml2Parser()
 101 {
 102     xmlFreeParserCtxt(parser);
 103     parser = NULL;
 104 }
 105
 106 bool
 107 ESILibxml2Parser::parse(char const *dataToParse, size_t const lengthOfData, bool const endOfStream)
 108 {
 109     return (xmlParseChunk(parser, dataToParse, lengthOfData, endOfStream) == 0);
 110 }
 111
 112 long int
 113 ESILibxml2Parser::lineNumber() const
 114 {
 115     return (long int)xmlSAX2GetLineNumber(parser);
 116 }
 117
 118 char const *
 119 ESILibxml2Parser::errorString() const
 120 {
 121     xmlErrorPtr error = xmlGetLastError();
 122
 123     if (error == NULL)
 124         return NULL;
 125
 126     return error->message;
 127 }
 128
 129 #endif /* USE_SQUID_ESI */
 130