htmlparser.c.svn-base

这是一个用于解析xml文件的类库。使用这个类库

/*
 * HTMLparser.c : an HTML 4.0 non-verifying parser
 *
 * See Copyright for the status of this software.
 *
 * [e-mail]
 */

#define IN_LIBXML
#include "libxml.h"
#ifdef LIBXML_HTML_ENABLED

#include <string.h>
#ifdef HAVE_CTYPE_H
#include <ctype.h>
#endif
#ifdef HAVE_STDLIB_H
#include <stdlib.h>
#endif
#ifdef HAVE_SYS_STAT_H
#include <sys/stat.h>
#endif
#ifdef HAVE_FCNTL_H
#include <fcntl.h>
#endif
#ifdef HAVE_UNISTD_H
#include <unistd.h>
#endif
#ifdef HAVE_ZLIB_H
#include <zlib.h>
#endif

#include <libxml/xmlmemory.h>
#include <libxml/tree.h>
#include <libxml/parser.h>
#include <libxml/parserInternals.h>
#include <libxml/xmlerror.h>
#include <libxml/HTMLparser.h>
#include <libxml/HTMLtree.h>
#include <libxml/entities.h>
#include <libxml/encoding.h>
#include <libxml/valid.h>
#include <libxml/xmlIO.h>
#include <libxml/globals.h>
#include <libxml/uri.h>

#define HTML_MAX_NAMELEN 1000
#define HTML_PARSER_BIG_BUFFER_SIZE 1000
#define HTML_PARSER_BUFFER_SIZE 100

/* #define DEBUG */
/* #define DEBUG_PUSH */

static int htmlOmittedDefaultValue = 1;

xmlChar * htmlDecodeEntities(htmlParserCtxtPtr ctxt, int len,
			     xmlChar end, xmlChar  end2, xmlChar end3);
static void htmlParseComment(htmlParserCtxtPtr ctxt);

/************************************************************************
 *									*
 * 		Some factorized error routines				*
 *									*
 ************************************************************************/

/**
 * htmlErrMemory:
 * @ctxt:  an HTML parser context
 * @extra:  extra informations
 *
 * Handle a redefinition of attribute error
 */
static void
htmlErrMemory(xmlParserCtxtPtr ctxt, const char *extra)
{
    if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
        (ctxt->instate == XML_PARSER_EOF))
	return;
    if (ctxt != NULL) {
        ctxt->errNo = XML_ERR_NO_MEMORY;
        ctxt->instate = XML_PARSER_EOF;
        ctxt->disableSAX = 1;
    }
    if (extra)
        __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_PARSER,
                        XML_ERR_NO_MEMORY, XML_ERR_FATAL, NULL, 0, extra,
                        NULL, NULL, 0, 0,
                        "Memory allocation failed : %s\n", extra);
    else
        __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_PARSER,
                        XML_ERR_NO_MEMORY, XML_ERR_FATAL, NULL, 0, NULL,
                        NULL, NULL, 0, 0, "Memory allocation failed\n");
}

/**
 * htmlParseErr:
 * @ctxt:  an HTML parser context
 * @error:  the error number
 * @msg:  the error message
 * @str1:  string infor
 * @str2:  string infor
 *
 * Handle a fatal parser error, i.e. violating Well-Formedness constraints
 */
static void
htmlParseErr(xmlParserCtxtPtr ctxt, xmlParserErrors error,
             const char *msg, const xmlChar *str1, const xmlChar *str2)
{
    if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
        (ctxt->instate == XML_PARSER_EOF))
	return;
    ctxt->errNo = error;
    __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_HTML, error,
                    XML_ERR_ERROR, NULL, 0,
		    (const char *) str1, (const char *) str2,
		    NULL, 0, 0,
		    msg, str1, str2);
    ctxt->wellFormed = 0;
}

/**
 * htmlParseErrInt:
 * @ctxt:  an HTML parser context
 * @error:  the error number
 * @msg:  the error message
 * @val:  integer info
 *
 * Handle a fatal parser error, i.e. violating Well-Formedness constraints
 */
static void
htmlParseErrInt(xmlParserCtxtPtr ctxt, xmlParserErrors error,
             const char *msg, int val)
{
    if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
        (ctxt->instate == XML_PARSER_EOF))
	return;
    ctxt->errNo = error;
    __xmlRaiseError(NULL, NULL, NULL, ctxt, NULL, XML_FROM_HTML, error,
                    XML_ERR_ERROR, NULL, 0, NULL, NULL,
		    NULL, val, 0, msg, val);
    ctxt->wellFormed = 0;
}

/************************************************************************
 *									*
 * 		Parser stacks related functions and macros		*
 *									*
 ************************************************************************/

/**
 * htmlnamePush:
 * @ctxt:  an HTML parser context
 * @value:  the element name
 *
 * Pushes a new element name on top of the name stack
 *
 * Returns 0 in case of error, the index in the stack otherwise
 */
static int
htmlnamePush(htmlParserCtxtPtr ctxt, const xmlChar * value)
{
    if (ctxt->nameNr >= ctxt->nameMax) {
        ctxt->nameMax *= 2;
        ctxt->nameTab = (const xmlChar * *)
                         xmlRealloc((xmlChar * *)ctxt->nameTab,
                                    ctxt->nameMax *
                                    sizeof(ctxt->nameTab[0]));
        if (ctxt->nameTab == NULL) {
            htmlErrMemory(ctxt, NULL);
            return (0);
        }
    }
    ctxt->nameTab[ctxt->nameNr] = value;
    ctxt->name = value;
    return (ctxt->nameNr++);
}
/**
 * htmlnamePop:
 * @ctxt: an HTML parser context
 *
 * Pops the top element name from the name stack
 *
 * Returns the name just removed
 */
static const xmlChar *
htmlnamePop(htmlParserCtxtPtr ctxt)
{
    const xmlChar *ret;

    if (ctxt->nameNr <= 0)
        return (0);
    ctxt->nameNr--;
    if (ctxt->nameNr < 0)
        return (0);
    if (ctxt->nameNr > 0)
        ctxt->name = ctxt->nameTab[ctxt->nameNr - 1];
    else
        ctxt->name = NULL;
    ret = ctxt->nameTab[ctxt->nameNr];
    ctxt->nameTab[ctxt->nameNr] = 0;
    return (ret);
}

/*
 * Macros for accessing the content. Those should be used only by the parser,
 * and not exported.
 *
 * Dirty macros, i.e. one need to make assumption on the context to use them
 *
 *   CUR_PTR return the current pointer to the xmlChar to be parsed.
 *   CUR     returns the current xmlChar value, i.e. a 8 bit value if compiled
 *           in ISO-Latin or UTF-8, and the current 16 bit value if compiled
 *           in UNICODE mode. This should be used internally by the parser
 *           only to compare to ASCII values otherwise it would break when
 *           running with UTF-8 encoding.
 *   NXT(n)  returns the n'th next xmlChar. Same as CUR is should be used only
 *           to compare on ASCII based substring.
 *   UPP(n)  returns the n'th next xmlChar converted to uppercase. Same as CUR
 *           it should be used only to compare on ASCII based substring.
 *   SKIP(n) Skip n xmlChar, and must also be used only to skip ASCII defined
 *           strings without newlines within the parser.
 *
 * Clean macros, not dependent of an ASCII context, expect UTF-8 encoding
 *
 *   CURRENT Returns the current char value, with the full decoding of
 *           UTF-8 if we are using this mode. It returns an int.
 *   NEXT    Skip to the next character, this does the proper decoding
 *           in UTF-8 mode. It also pop-up unfinished entities on the fly.
 *   NEXTL(l) Skip the current unicode character of l xmlChars long.
 *   COPY(to) copy one char to *to, increment CUR_PTR and to accordingly
 */

#define UPPER (toupper(*ctxt->input->cur))

#define SKIP(val) ctxt->nbChars += (val),ctxt->input->cur += (val),ctxt->input->col+=(val)

#define NXT(val) ctxt->input->cur[(val)]

#define UPP(val) (toupper(ctxt->input->cur[(val)]))

#define CUR_PTR ctxt->input->cur

#define SHRINK if ((ctxt->input->cur - ctxt->input->base > 2 * INPUT_CHUNK) && \
		   (ctxt->input->end - ctxt->input->cur < 2 * INPUT_CHUNK)) \
	xmlParserInputShrink(ctxt->input)

#define GROW if ((ctxt->progressive == 0) &&				\
		 (ctxt->input->end - ctxt->input->cur < INPUT_CHUNK))	\
	xmlParserInputGrow(ctxt->input, INPUT_CHUNK)

#define CURRENT ((int) (*ctxt->input->cur))

#define SKIP_BLANKS htmlSkipBlankChars(ctxt)

/* Inported from XML */

/* #define CUR (ctxt->token ? ctxt->token : (int) (*ctxt->input->cur)) */
#define CUR ((int) (*ctxt->input->cur))
#define NEXT xmlNextChar(ctxt)

#define RAW (ctxt->token ? -1 : (*ctxt->input->cur))
#define NXT(val) ctxt->input->cur[(val)]
#define CUR_PTR ctxt->input->cur


#define NEXTL(l) do {							\
    if (*(ctxt->input->cur) == '\n') {					\
	ctxt->input->line++; ctxt->input->col = 1;			\
    } else ctxt->input->col++;						\
    ctxt->token = 0; ctxt->input->cur += l; ctxt->nbChars++;		\
  } while (0)
    
/************
    \
    if (*ctxt->input->cur == '%') xmlParserHandlePEReference(ctxt);	\
    if (*ctxt->input->cur == '&') xmlParserHandleReference(ctxt);
 ************/

#define CUR_CHAR(l) htmlCurrentChar(ctxt, &l)
#define CUR_SCHAR(s, l) xmlStringCurrentChar(ctxt, s, &l)

#define COPY_BUF(l,b,i,v)						\
    if (l == 1) b[i++] = (xmlChar) v;					\
    else i += xmlCopyChar(l,&b[i],v)

/**
 * htmlCurrentChar:
 * @ctxt:  the HTML parser context
 * @len:  pointer to the length of the char read
 *
 * The current char value, if using UTF-8 this may actually span multiple
 * bytes in the input buffer. Implement the end of line normalization:
 * 2.11 End-of-Line Handling
 * If the encoding is unspecified, in the case we find an ISO-Latin-1
 * char, then the encoding converter is plugged in automatically.
 *
 * Returns the current char value and its length
 */

static int
htmlCurrentChar(xmlParserCtxtPtr ctxt, int *len) {
    if (ctxt->instate == XML_PARSER_EOF)
	return(0);

    if (ctxt->token != 0) {
	*len = 0;
	return(ctxt->token);
    }	
    if (ctxt->charset == XML_CHAR_ENCODING_UTF8) {
	/*
	 * We are supposed to handle UTF8, check it's valid
	 * From rfc2044: encoding of the Unicode values on UTF-8:
	 *
	 * UCS-4 range (hex.)           UTF-8 octet sequence (binary)
	 * 0000 0000-0000 007F   0xxxxxxx
	 * 0000 0080-0000 07FF   110xxxxx 10xxxxxx
	 * 0000 0800-0000 FFFF   1110xxxx 10xxxxxx 10xxxxxx 
	 *
	 * Check for the 0x110000 limit too
	 */
	const unsigned char *cur = ctxt->input->cur;
	unsigned char c;
	unsigned int val;

	c = *cur;
	if (c & 0x80) {
	    if (cur[1] == 0)
		xmlParserInputGrow(ctxt->input, INPUT_CHUNK);
	    if ((cur[1] & 0xc0) != 0x80)
		goto encoding_error;
	    if ((c & 0xe0) == 0xe0) {

		if (cur[2] == 0)
		    xmlParserInputGrow(ctxt->input, INPUT_CHUNK);
		if ((cur[2] & 0xc0) != 0x80)
		    goto encoding_error;
		if ((c & 0xf0) == 0xf0) {
		    if (cur[3] == 0)
			xmlParserInputGrow(ctxt->input, INPUT_CHUNK);
		    if (((c & 0xf8) != 0xf0) ||
			((cur[3] & 0xc0) != 0x80))
			goto encoding_error;
		    /* 4-byte code */
		    *len = 4;
		    val = (cur[0] & 0x7) << 18;
		    val |= (cur[1] & 0x3f) << 12;
		    val |= (cur[2] & 0x3f) << 6;