xmlparse.c

一套很值得分析的短信SMS开发源代码。是我今年早些时候从taobao上买来的。但我现在也没看完（先说清楚

/*
Copyright (c) 1998, 1999, 2000 Thai Open Source Software Center Ltd
See the file COPYING for copying permission.
*/

static char RCSId[]
  = "$Header: /home/cvs/apr-util/xml/expat/lib/xmlparse.c,v 1.4 2001/08/30 05:44:18 wrowe Exp $";

#ifdef COMPILED_FROM_DSP
#  include "winconfig.h"
#  define XMLPARSEAPI __declspec(dllexport)
#  include "expat.h"
#  undef XMLPARSEAPI
#else
#include <config.h>

#ifdef HAVE_STRING_H
#  include <string.h>
#endif

#ifndef __CYGWIN__
#ifdef __declspec
#  define XMLPARSEAPI __declspec(dllexport)
#endif
#endif

#include "expat.h"

#ifdef __declspec
#  undef XMLPARSEAPI
#endif
#endif /* ndef COMPILED_FROM_DSP */

#include <stddef.h>

#ifdef XML_UNICODE
#define XML_ENCODE_MAX XML_UTF16_ENCODE_MAX
#define XmlConvert XmlUtf16Convert
#define XmlGetInternalEncoding XmlGetUtf16InternalEncoding
#define XmlGetInternalEncodingNS XmlGetUtf16InternalEncodingNS
#define XmlEncode XmlUtf16Encode
#define MUST_CONVERT(enc, s) (!(enc)->isUtf16 || (((unsigned long)s) & 1))
typedef unsigned short ICHAR;
#else
#define XML_ENCODE_MAX XML_UTF8_ENCODE_MAX
#define XmlConvert XmlUtf8Convert
#define XmlGetInternalEncoding XmlGetUtf8InternalEncoding
#define XmlGetInternalEncodingNS XmlGetUtf8InternalEncodingNS
#define XmlEncode XmlUtf8Encode
#define MUST_CONVERT(enc, s) (!(enc)->isUtf8)
typedef char ICHAR;
#endif


#ifndef XML_NS

#define XmlInitEncodingNS XmlInitEncoding
#define XmlInitUnknownEncodingNS XmlInitUnknownEncoding
#undef XmlGetInternalEncodingNS
#define XmlGetInternalEncodingNS XmlGetInternalEncoding
#define XmlParseXmlDeclNS XmlParseXmlDecl

#endif

#ifdef XML_UNICODE_WCHAR_T
#define XML_T(x) L ## x
#else
#define XML_T(x) x
#endif

/* Round up n to be a multiple of sz, where sz is a power of 2. */
#define ROUND_UP(n, sz) (((n) + ((sz) - 1)) & ~((sz) - 1))

#include "xmltok.h"
#include "xmlrole.h"

typedef const XML_Char *KEY;

typedef struct {
  KEY name;
} NAMED;

typedef struct {
  NAMED **v;
  size_t size;
  size_t used;
  size_t usedLim;
  XML_Memory_Handling_Suite *mem;
} HASH_TABLE;

typedef struct {
  NAMED **p;
  NAMED **end;
} HASH_TABLE_ITER;

#define INIT_TAG_BUF_SIZE 32  /* must be a multiple of sizeof(XML_Char) */
#define INIT_DATA_BUF_SIZE 1024
#define INIT_ATTS_SIZE 16
#define INIT_BLOCK_SIZE 1024
#define INIT_BUFFER_SIZE 1024

#define EXPAND_SPARE 24

typedef struct binding {
  struct prefix *prefix;
  struct binding *nextTagBinding;
  struct binding *prevPrefixBinding;
  const struct attribute_id *attId;
  XML_Char *uri;
  int uriLen;
  int uriAlloc;
} BINDING;

typedef struct prefix {
  const XML_Char *name;
  BINDING *binding;
} PREFIX;

typedef struct {
  const XML_Char *str;
  const XML_Char *localPart;
  int uriLen;
} TAG_NAME;

typedef struct tag {
  struct tag *parent;
  const char *rawName;
  int rawNameLength;
  TAG_NAME name;
  char *buf;
  char *bufEnd;
  BINDING *bindings;
} TAG;

typedef struct {
  const XML_Char *name;
  const XML_Char *textPtr;
  int textLen;
  const XML_Char *systemId;
  const XML_Char *base;
  const XML_Char *publicId;
  const XML_Char *notation;
  char open;
  char is_param;
} ENTITY;

typedef struct {
  enum XML_Content_Type		type;
  enum XML_Content_Quant	quant;
  const XML_Char *		name;
  int				firstchild;
  int				lastchild;
  int				childcnt;
  int				nextsib;
} CONTENT_SCAFFOLD;

typedef struct block {
  struct block *next;
  int size;
  XML_Char s[1];
} BLOCK;

typedef struct {
  BLOCK *blocks;
  BLOCK *freeBlocks;
  const XML_Char *end;
  XML_Char *ptr;
  XML_Char *start;
  XML_Memory_Handling_Suite *mem;
} STRING_POOL;

/* The XML_Char before the name is used to determine whether
an attribute has been specified. */
typedef struct attribute_id {
  XML_Char *name;
  PREFIX *prefix;
  char maybeTokenized;
  char xmlns;
} ATTRIBUTE_ID;

typedef struct {
  const ATTRIBUTE_ID *id;
  char isCdata;
  const XML_Char *value;
} DEFAULT_ATTRIBUTE;

typedef struct {
  const XML_Char *name;
  PREFIX *prefix;
  const ATTRIBUTE_ID *idAtt;
  int nDefaultAtts;
  int allocDefaultAtts;
  DEFAULT_ATTRIBUTE *defaultAtts;
} ELEMENT_TYPE;

typedef struct {
  HASH_TABLE generalEntities;
  HASH_TABLE elementTypes;
  HASH_TABLE attributeIds;
  HASH_TABLE prefixes;
  STRING_POOL pool;
  int complete;
  int standalone;
#ifdef XML_DTD
  HASH_TABLE paramEntities;
#endif /* XML_DTD */
  PREFIX defaultPrefix;
  /* === scaffolding for building content model === */
  int in_eldecl;
  CONTENT_SCAFFOLD *scaffold;
  unsigned contentStringLen;
  unsigned scaffSize;
  unsigned scaffCount;
  int scaffLevel;
  int *scaffIndex;
} DTD;

typedef struct open_internal_entity {
  const char *internalEventPtr;
  const char *internalEventEndPtr;
  struct open_internal_entity *next;
  ENTITY *entity;
} OPEN_INTERNAL_ENTITY;

typedef enum XML_Error Processor(XML_Parser parser,
				 const char *start,
				 const char *end,
				 const char **endPtr);

static Processor prologProcessor;
static Processor prologInitProcessor;
static Processor contentProcessor;
static Processor cdataSectionProcessor;
#ifdef XML_DTD
static Processor ignoreSectionProcessor;
#endif /* XML_DTD */
static Processor epilogProcessor;
static Processor errorProcessor;
static Processor externalEntityInitProcessor;
static Processor externalEntityInitProcessor2;
static Processor externalEntityInitProcessor3;
static Processor externalEntityContentProcessor;

static enum XML_Error
handleUnknownEncoding(XML_Parser parser, const XML_Char *encodingName);
static enum XML_Error
processXmlDecl(XML_Parser parser, int isGeneralTextEntity, const char *, const char *);
static enum XML_Error
initializeEncoding(XML_Parser parser);
static enum XML_Error
doProlog(XML_Parser parser, const ENCODING *enc, const char *s,
	 const char *end, int tok, const char *next, const char **nextPtr);
static enum XML_Error
processInternalParamEntity(XML_Parser parser, ENTITY *entity);
static enum XML_Error
doContent(XML_Parser parser, int startTagLevel, const ENCODING *enc,
	  const char *start, const char *end, const char **endPtr);
static enum XML_Error
doCdataSection(XML_Parser parser, const ENCODING *, const char **startPtr, const char *end, const char **nextPtr);
#ifdef XML_DTD
static enum XML_Error
doIgnoreSection(XML_Parser parser, const ENCODING *, const char **startPtr, const char *end, const char **nextPtr);
#endif /* XML_DTD */
static enum XML_Error storeAtts(XML_Parser parser, const ENCODING *, const char *s,
				TAG_NAME *tagNamePtr, BINDING **bindingsPtr);
static
int addBinding(XML_Parser parser, PREFIX *prefix, const ATTRIBUTE_ID *attId, const XML_Char *uri, BINDING **bindingsPtr);

static int
defineAttribute(ELEMENT_TYPE *type, ATTRIBUTE_ID *,
		int isCdata, int isId, const XML_Char *dfltValue,
		XML_Parser parser);

static enum XML_Error
storeAttributeValue(XML_Parser parser, const ENCODING *, int isCdata, const char *, const char *,
		    STRING_POOL *);
static enum XML_Error
appendAttributeValue(XML_Parser parser, const ENCODING *, int isCdata, const char *, const char *,
		    STRING_POOL *);
static ATTRIBUTE_ID *
getAttributeId(XML_Parser parser, const ENCODING *enc, const char *start, const char *end);
static int setElementTypePrefix(XML_Parser parser, ELEMENT_TYPE *);
static enum XML_Error
storeEntityValue(XML_Parser parser, const ENCODING *enc, const char *start, const char *end);
static int
reportProcessingInstruction(XML_Parser parser, const ENCODING *enc, const char *start, const char *end);
static int
reportComment(XML_Parser parser, const ENCODING *enc, const char *start, const char *end);
static void
reportDefault(XML_Parser parser, const ENCODING *enc, const char *start, const char *end);

static const XML_Char *getContext(XML_Parser parser);
static int setContext(XML_Parser parser, const XML_Char *context);
static void normalizePublicId(XML_Char *s);
static int dtdInit(DTD *, XML_Parser parser);

static void dtdDestroy(DTD *, XML_Parser parser);

static int dtdCopy(DTD *newDtd, const DTD *oldDtd, XML_Parser parser);

static int copyEntityTable(HASH_TABLE *, STRING_POOL *, const HASH_TABLE *,
			   XML_Parser parser);

#ifdef XML_DTD
static void dtdSwap(DTD *, DTD *);
#endif /* XML_DTD */

static NAMED *lookup(HASH_TABLE *table, KEY name, size_t createSize);

static void hashTableInit(HASH_TABLE *, XML_Memory_Handling_Suite *ms);

static void hashTableDestroy(HASH_TABLE *);
static void hashTableIterInit(HASH_TABLE_ITER *, const HASH_TABLE *);
static NAMED *hashTableIterNext(HASH_TABLE_ITER *);
static void poolInit(STRING_POOL *, XML_Memory_Handling_Suite *ms);
static void poolClear(STRING_POOL *);
static void poolDestroy(STRING_POOL *);
static XML_Char *poolAppend(STRING_POOL *pool, const ENCODING *enc,
			    const char *ptr, const char *end);
static XML_Char *poolStoreString(STRING_POOL *pool, const ENCODING *enc,
				  const char *ptr, const char *end);

static int poolGrow(STRING_POOL *pool);

static int nextScaffoldPart(XML_Parser parser);
static XML_Content *build_model(XML_Parser parser);

static const XML_Char *poolCopyString(STRING_POOL *pool, const XML_Char *s);
static const XML_Char *poolCopyStringN(STRING_POOL *pool, const XML_Char *s, int n);
static const XML_Char *poolAppendString(STRING_POOL *pool, const XML_Char *s);
static ELEMENT_TYPE * getElementType(XML_Parser Paraser,
				     const ENCODING *enc,
				     const char *ptr,
				     const char *end);

#define poolStart(pool) ((pool)->start)
#define poolEnd(pool) ((pool)->ptr)
#define poolLength(pool) ((pool)->ptr - (pool)->start)
#define poolChop(pool) ((void)--(pool->ptr))
#define poolLastChar(pool) (((pool)->ptr)[-1])
#define poolDiscard(pool) ((pool)->ptr = (pool)->start)
#define poolFinish(pool) ((pool)->start = (pool)->ptr)
#define poolAppendChar(pool, c) \
  (((pool)->ptr == (pool)->end && !poolGrow(pool)) \
   ? 0 \
   : ((*((pool)->ptr)++ = c), 1))

typedef struct {
  /* The first member must be userData so that the XML_GetUserData macro works. */
  void *m_userData;
  void *m_handlerArg;
  char *m_buffer;
  XML_Memory_Handling_Suite m_mem;
  /* first character to be parsed */
  const char *m_bufferPtr;
  /* past last character to be parsed */
  char *m_bufferEnd;
  /* allocated end of buffer */
  const char *m_bufferLim;
  long m_parseEndByteIndex;
  const char *m_parseEndPtr;
  XML_Char *m_dataBuf;
  XML_Char *m_dataBufEnd;
  XML_StartElementHandler m_startElementHandler;
  XML_EndElementHandler m_endElementHandler;
  XML_CharacterDataHandler m_characterDataHandler;
  XML_ProcessingInstructionHandler m_processingInstructionHandler;
  XML_CommentHandler m_commentHandler;
  XML_StartCdataSectionHandler m_startCdataSectionHandler;
  XML_EndCdataSectionHandler m_endCdataSectionHandler;
  XML_DefaultHandler m_defaultHandler;
  XML_StartDoctypeDeclHandler m_startDoctypeDeclHandler;
  XML_EndDoctypeDeclHandler m_endDoctypeDeclHandler;
  XML_UnparsedEntityDeclHandler m_unparsedEntityDeclHandler;
  XML_NotationDeclHandler m_notationDeclHandler;
  XML_StartNamespaceDeclHandler m_startNamespaceDeclHandler;
  XML_EndNamespaceDeclHandler m_endNamespaceDeclHandler;
  XML_NotStandaloneHandler m_notStandaloneHandler;
  XML_ExternalEntityRefHandler m_externalEntityRefHandler;
  void *m_externalEntityRefHandlerArg;
  XML_UnknownEncodingHandler m_unknownEncodingHandler;
  XML_ElementDeclHandler m_elementDeclHandler;
  XML_AttlistDeclHandler m_attlistDeclHandler;
  XML_EntityDeclHandler m_entityDeclHandler;
  XML_XmlDeclHandler m_xmlDeclHandler;
  const ENCODING *m_encoding;
  INIT_ENCODING m_initEncoding;
  const ENCODING *m_internalEncoding;
  const XML_Char *m_protocolEncodingName;
  int m_ns;
  int m_ns_triplets;
  void *m_unknownEncodingMem;
  void *m_unknownEncodingData;
  void *m_unknownEncodingHandlerData;
  void (*m_unknownEncodingRelease)(void *);
  PROLOG_STATE m_prologState;
  Processor *m_processor;
  enum XML_Error m_errorCode;
  const char *m_eventPtr;
  const char *m_eventEndPtr;
  const char *m_positionPtr;
  OPEN_INTERNAL_ENTITY *m_openInternalEntities;
  int m_defaultExpandInternalEntities;
  int m_tagLevel;
  ENTITY *m_declEntity;
  const XML_Char *m_doctypeName;
  const XML_Char *m_doctypeSysid;
  const XML_Char *m_doctypePubid;
  const XML_Char *m_declAttributeType;
  const XML_Char *m_declNotationName;
  const XML_Char *m_declNotationPublicId;
  ELEMENT_TYPE *m_declElementType;
  ATTRIBUTE_ID *m_declAttributeId;
  char m_declAttributeIsCdata;
  char m_declAttributeIsId;
  DTD m_dtd;
  const XML_Char *m_curBase;
  TAG *m_tagStack;
  TAG *m_freeTagList;
  BINDING *m_inheritedBindings;
  BINDING *m_freeBindingList;
  int m_attsSize;
  int m_nSpecifiedAtts;
  int m_idAttIndex;
  ATTRIBUTE *m_atts;
  POSITION m_position;
  STRING_POOL m_tempPool;
  STRING_POOL m_temp2Pool;
  char *m_groupConnector;
  unsigned m_groupSize;
  int m_hadExternalDoctype;
  XML_Char m_namespaceSeparator;
#ifdef XML_DTD
  enum XML_ParamEntityParsing m_paramEntityParsing;
  XML_Parser m_parentParser;