xmlwf.c

一个很好用的解析

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

#include <stdio.h>
#include <stdlib.h>
#include <stddef.h>
#include <string.h>

#include "xmlparse.h"
#include "codepage.h"
#include "xmlfile.h"
#include "xmltchar.h"

#ifdef _MSC_VER
#include <crtdbg.h>
#endif

/* This ensures proper sorting. */

#define NSSEP T('\001')

static void characterData(void *userData, const XML_Char *s, int len)
{
  FILE *fp = userData;
  for (; len > 0; --len, ++s) {
    switch (*s) {
    case T('&'):
      fputts(T("&amp;"), fp);
      break;
    case T('<'):
      fputts(T("&lt;"), fp);
      break;
    case T('>'):
      fputts(T("&gt;"), fp);
      break;
#ifdef W3C14N
    case 13:
      fputts(T("&#xD;"), fp);
      break;
#else
    case T('"'):
      fputts(T("&quot;"), fp);
      break;
    case 9:
    case 10:
    case 13:
      ftprintf(fp, T("&#%d;"), *s);
      break;
#endif
    default:
      puttc(*s, fp);
      break;
    }
  }
}

static void attributeValue(FILE *fp, const XML_Char *s)
{
  puttc(T('='), fp);
  puttc(T('"'), fp);
  for (;;) {
    switch (*s) {
    case 0:
    case NSSEP:
      puttc(T('"'), fp);
      return;
    case T('&'):
      fputts(T("&amp;"), fp);
      break;
    case T('<'):
      fputts(T("&lt;"), fp);
      break;
    case T('"'):
      fputts(T("&quot;"), fp);
      break;
#ifdef W3C14N
    case 9:
      fputts(T("&#x9;"), fp);
      break;
    case 10:
      fputts(T("&#xA;"), fp);
      break;
    case 13:
      fputts(T("&#xD;"), fp);
      break;
#else
    case T('>'):
      fputts(T("&gt;"), fp);
      break;
    case 9:
    case 10:
    case 13:
      ftprintf(fp, T("&#%d;"), *s);
      break;
#endif
    default:
      puttc(*s, fp);
      break;
    }
    s++;
  }
}

/* Lexicographically comparing UTF-8 encoded attribute values,
is equivalent to lexicographically comparing based on the character number. */

static int attcmp(const void *att1, const void *att2)
{
  return tcscmp(*(const XML_Char **)att1, *(const XML_Char **)att2);
}

static void startElement(void *userData, const XML_Char *name, const XML_Char **atts)
{
  int nAtts;
  const XML_Char **p;
  FILE *fp = userData;
  puttc(T('<'), fp);
  fputts(name, fp);

  p = atts;
  while (*p)
    ++p;
  nAtts = (p - atts) >> 1;
  if (nAtts > 1)
    qsort((void *)atts, nAtts, sizeof(XML_Char *) * 2, attcmp);
  while (*atts) {
    puttc(T(' '), fp);
    fputts(*atts++, fp);
    attributeValue(fp, *atts);
    atts++;
  }
  puttc(T('>'), fp);
}

static void endElement(void *userData, const XML_Char *name)
{
  FILE *fp = userData;
  puttc(T('<'), fp);
  puttc(T('/'), fp);
  fputts(name, fp);
  puttc(T('>'), fp);
}

static int nsattcmp(const void *p1, const void *p2)
{
  const XML_Char *att1 = *(const XML_Char **)p1;
  const XML_Char *att2 = *(const XML_Char **)p2;
  int sep1 = (tcsrchr(att1, NSSEP) != 0);
  int sep2 = (tcsrchr(att1, NSSEP) != 0);
  if (sep1 != sep2)
    return sep1 - sep2;
  return tcscmp(att1, att2);
}

static void startElementNS(void *userData, const XML_Char *name, const XML_Char **atts)
{
  int nAtts;
  int nsi;
  const XML_Char **p;
  FILE *fp = userData;
  const XML_Char *sep;
  puttc(T('<'), fp);

  sep = tcsrchr(name, NSSEP);
  if (sep) {
    fputts(T("n1:"), fp);
    fputts(sep + 1, fp);
    fputts(T(" xmlns:n1"), fp);
    attributeValue(fp, name);
    nsi = 2;
  }
  else {
    fputts(name, fp);
    nsi = 1;
  }

  p = atts;
  while (*p)
    ++p;
  nAtts = (p - atts) >> 1;
  if (nAtts > 1)
    qsort((void *)atts, nAtts, sizeof(XML_Char *) * 2, nsattcmp);
  while (*atts) {
    name = *atts++;
    sep = tcsrchr(name, NSSEP);
    puttc(T(' '), fp);
    if (sep) {
      ftprintf(fp, T("n%d:"), nsi);
      fputts(sep + 1, fp);
    }
    else
      fputts(name, fp);
    attributeValue(fp, *atts);
    if (sep) {
      ftprintf(fp, T(" xmlns:n%d"), nsi++);
      attributeValue(fp, name);
    }
    atts++;
  }
  puttc(T('>'), fp);
}

static void endElementNS(void *userData, const XML_Char *name)
{
  FILE *fp = userData;
  const XML_Char *sep;
  puttc(T('<'), fp);
  puttc(T('/'), fp);
  sep = tcsrchr(name, NSSEP);
  if (sep) {
    fputts(T("n1:"), fp);
    fputts(sep + 1, fp);
  }
  else
    fputts(name, fp);
  puttc(T('>'), fp);
}

#ifndef W3C14N

static void processingInstruction(void *userData, const XML_Char *target, const XML_Char *data)
{
  FILE *fp = userData;
  puttc(T('<'), fp);
  puttc(T('?'), fp);
  fputts(target, fp);
  puttc(T(' '), fp);
  fputts(data, fp);
  puttc(T('?'), fp);
  puttc(T('>'), fp);
}

#endif /* not W3C14N */

static void defaultCharacterData(XML_Parser parser, const XML_Char *s, int len)
{
  XML_DefaultCurrent(parser);
}

static void defaultStartElement(XML_Parser parser, const XML_Char *name, const XML_Char **atts)
{
  XML_DefaultCurrent(parser);
}

static void defaultEndElement(XML_Parser parser, const XML_Char *name)
{
  XML_DefaultCurrent(parser);
}

static void defaultProcessingInstruction(XML_Parser parser, const XML_Char *target, const XML_Char *data)
{
  XML_DefaultCurrent(parser);
}

static void nopCharacterData(XML_Parser parser, const XML_Char *s, int len)
{
}

static void nopStartElement(XML_Parser parser, const XML_Char *name, const XML_Char **atts)
{
}

static void nopEndElement(XML_Parser parser, const XML_Char *name)
{
}

static void nopProcessingInstruction(XML_Parser parser, const XML_Char *target, const XML_Char *data)
{
}

static void markup(XML_Parser parser, const XML_Char *s, int len)
{
  FILE *fp = XML_GetUserData(parser);
  for (; len > 0; --len, ++s)
    puttc(*s, fp);
}

static
void metaLocation(XML_Parser parser)
{
  const XML_Char *uri = XML_GetBase(parser);
  if (uri)
    ftprintf(XML_GetUserData(parser), T(" uri=\"%s\""), uri);
  ftprintf(XML_GetUserData(parser),
           T(" byte=\"%ld\" nbytes=\"%d\" line=\"%d\" col=\"%d\""),
	   XML_GetCurrentByteIndex(parser),
	   XML_GetCurrentByteCount(parser),
	   XML_GetCurrentLineNumber(parser),
	   XML_GetCurrentColumnNumber(parser));
}

static
void metaStartDocument(XML_Parser parser)
{
  fputts(T("<document>\n"), XML_GetUserData(parser));
}

static
void metaEndDocument(XML_Parser parser)
{
  fputts(T("</document>\n"), XML_GetUserData(parser));
}

static
void metaStartElement(XML_Parser parser, const XML_Char *name, const XML_Char **atts)
{
  FILE *fp = XML_GetUserData(parser);
  const XML_Char **specifiedAttsEnd
    = atts + XML_GetSpecifiedAttributeCount(parser);
  const XML_Char **idAttPtr;
  int idAttIndex = XML_GetIdAttributeIndex(parser);
  if (idAttIndex < 0)
    idAttPtr = 0;
  else
    idAttPtr = atts + idAttIndex;
    
  ftprintf(fp, T("<starttag name=\"%s\""), name);
  metaLocation(parser);
  if (*atts) {
    fputts(T(">\n"), fp);
    do {
      ftprintf(fp, T("<attribute name=\"%s\" value=\""), atts[0]);
      characterData(fp, atts[1], tcslen(atts[1]));
      if (atts >= specifiedAttsEnd)
	fputts(T("\" defaulted=\"yes\"/>\n"), fp);
      else if (atts == idAttPtr)
	fputts(T("\" id=\"yes\"/>\n"), fp);
      else
	fputts(T("\"/>\n"), fp);
    } while (*(atts += 2));
    fputts(T("</starttag>\n"), fp);
  }
  else
    fputts(T("/>\n"), fp);
}

static
void metaEndElement(XML_Parser parser, const XML_Char *name)
{
  FILE *fp = XML_GetUserData(parser);
  ftprintf(fp, T("<endtag name=\"%s\""), name);
  metaLocation(parser);
  fputts(T("/>\n"), fp);
}

static
void metaProcessingInstruction(XML_Parser parser, const XML_Char *target, const XML_Char *data)
{
  FILE *fp = XML_GetUserData(parser);
  ftprintf(fp, T("<pi target=\"%s\" data=\""), target);
  characterData(fp, data, tcslen(data));
  puttc(T('"'), fp);
  metaLocation(parser);
  fputts(T("/>\n"), fp);
}

static
void metaComment(XML_Parser parser, const XML_Char *data)
{
  FILE *fp = XML_GetUserData(parser);
  fputts(T("<comment data=\""), fp);
  characterData(fp, data, tcslen(data));
  puttc(T('"'), fp);
  metaLocation(parser);
  fputts(T("/>\n"), fp);
}

static
void metaStartCdataSection(XML_Parser parser)
{
  FILE *fp = XML_GetUserData(parser);
  fputts(T("<startcdata"), fp);
  metaLocation(parser);
  fputts(T("/>\n"), fp);
}

static
void metaEndCdataSection(XML_Parser parser)
{
  FILE *fp = XML_GetUserData(parser);
  fputts(T("<endcdata"), fp);
  metaLocation(parser);