parserinternals.c.svn-base

这是一个用于解析xml文件的类库。使用这个类库

/*
 * parserInternals.c : Internal routines (and obsolete ones) needed for the
 *                     XML and HTML parsers.
 *
 * See Copyright for the status of this software.
 *
 * [e-mail]
 */

#define IN_LIBXML
#include "libxml.h"

#if defined(WIN32) && !defined (__CYGWIN__)
#define XML_DIR_SEP '\\'
#else
#define XML_DIR_SEP '/'
#endif

#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/valid.h>
#include <libxml/entities.h>
#include <libxml/xmlerror.h>
#include <libxml/encoding.h>
#include <libxml/valid.h>
#include <libxml/xmlIO.h>
#include <libxml/uri.h>
#include <libxml/dict.h>
#include <libxml/SAX.h>
#ifdef LIBXML_CATALOG_ENABLED
#include <libxml/catalog.h>
#endif
#include <libxml/globals.h>
#include <libxml/chvalid.h>

/*
 * Various global defaults for parsing
 */

/**
 * xmlCheckVersion:
 * @version: the include version number
 *
 * check the compiled lib version against the include one.
 * This can warn or immediately kill the application
 */
void
xmlCheckVersion(int version) {
    int myversion = (int) LIBXML_VERSION;

    xmlInitParser();

    if ((myversion / 10000) != (version / 10000)) {
	xmlGenericError(xmlGenericErrorContext, 
		"Fatal: program compiled against libxml %d using libxml %d\n",
		(version / 10000), (myversion / 10000));
	fprintf(stderr, 
		"Fatal: program compiled against libxml %d using libxml %d\n",
		(version / 10000), (myversion / 10000));
    }
    if ((myversion / 100) < (version / 100)) {
	xmlGenericError(xmlGenericErrorContext, 
		"Warning: program compiled against libxml %d using older %d\n",
		(version / 100), (myversion / 100));
    }
}


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


/**
 * xmlErrMemory:
 * @ctxt:  an XML parser context
 * @extra:  extra informations
 *
 * Handle a redefinition of attribute error
 */
void
xmlErrMemory(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");
}

/**
 * __xmlErrEncoding:
 * @ctxt:  an XML parser context
 * @error:  the error number
 * @msg:  the error message
 * @str1:  an string info
 * @str2:  an string info
 *
 * Handle an encoding error
 */
void
__xmlErrEncoding(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;
    if (ctxt != NULL)
        ctxt->errNo = error;
    __xmlRaiseError(NULL, NULL, NULL,
                    ctxt, NULL, XML_FROM_PARSER, error, XML_ERR_FATAL,
                    NULL, 0, (const char *) str1, (const char *) str2,
                    NULL, 0, 0, msg, str1, str2);
    if (ctxt != NULL) {
        ctxt->wellFormed = 0;
        if (ctxt->recovery == 0)
            ctxt->disableSAX = 1;
    }
}

/**
 * xmlErrInternal:
 * @ctxt:  an XML parser context
 * @msg:  the error message
 * @str:  error informations
 *
 * Handle an internal error
 */
static void
xmlErrInternal(xmlParserCtxtPtr ctxt, const char *msg, const xmlChar * str)
{
    if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
        (ctxt->instate == XML_PARSER_EOF))
	return;
    if (ctxt != NULL)
        ctxt->errNo = XML_ERR_INTERNAL_ERROR;
    __xmlRaiseError(NULL, NULL, NULL,
                    ctxt, NULL, XML_FROM_PARSER, XML_ERR_INTERNAL_ERROR,
                    XML_ERR_FATAL, NULL, 0, (const char *) str, NULL, NULL,
                    0, 0, msg, str);
    if (ctxt != NULL) {
        ctxt->wellFormed = 0;
        if (ctxt->recovery == 0)
            ctxt->disableSAX = 1;
    }
}

/**
 * xmlErrEncodingInt:
 * @ctxt:  an XML parser context
 * @error:  the error number
 * @msg:  the error message
 * @val:  an integer value
 *
 * n encoding error
 */
static void
xmlErrEncodingInt(xmlParserCtxtPtr ctxt, xmlParserErrors error,
                  const char *msg, int val)
{
    if ((ctxt != NULL) && (ctxt->disableSAX != 0) &&
        (ctxt->instate == XML_PARSER_EOF))
	return;
    if (ctxt != NULL)
        ctxt->errNo = error;
    __xmlRaiseError(NULL, NULL, NULL,
                    ctxt, NULL, XML_FROM_PARSER, error, XML_ERR_FATAL,
                    NULL, 0, NULL, NULL, NULL, val, 0, msg, val);
    if (ctxt != NULL) {
        ctxt->wellFormed = 0;
        if (ctxt->recovery == 0)
            ctxt->disableSAX = 1;
    }
}

/**
 * xmlIsLetter:
 * @c:  an unicode character (int)
 *
 * Check whether the character is allowed by the production
 * [84] Letter ::= BaseChar | Ideographic
 *
 * Returns 0 if not, non-zero otherwise
 */
int
xmlIsLetter(int c) {
    return(IS_BASECHAR(c) || IS_IDEOGRAPHIC(c));
}

/************************************************************************
 *									*
 * 		Input handling functions for progressive parsing	*
 *									*
 ************************************************************************/

/* #define DEBUG_INPUT */
/* #define DEBUG_STACK */
/* #define DEBUG_PUSH */


/* we need to keep enough input to show errors in context */
#define LINE_LEN        80

#ifdef DEBUG_INPUT
#define CHECK_BUFFER(in) check_buffer(in)

static
void check_buffer(xmlParserInputPtr in) {
    if (in->base != in->buf->buffer->content) {
        xmlGenericError(xmlGenericErrorContext,
		"xmlParserInput: base mismatch problem\n");
    }
    if (in->cur < in->base) {
        xmlGenericError(xmlGenericErrorContext,
		"xmlParserInput: cur < base problem\n");
    }
    if (in->cur > in->base + in->buf->buffer->use) {
        xmlGenericError(xmlGenericErrorContext,
		"xmlParserInput: cur > base + use problem\n");
    }
    xmlGenericError(xmlGenericErrorContext,"buffer %x : content %x, cur %d, use %d, size %d\n",
            (int) in, (int) in->buf->buffer->content, in->cur - in->base,
	    in->buf->buffer->use, in->buf->buffer->size);
}

#else
#define CHECK_BUFFER(in) 
#endif


/**
 * xmlParserInputRead:
 * @in:  an XML parser input
 * @len:  an indicative size for the lookahead
 *
 * This function refresh the input for the parser. It doesn't try to
 * preserve pointers to the input buffer, and discard already read data
 *
 * Returns the number of xmlChars read, or -1 in case of error, 0 indicate the
 * end of this entity
 */
int
xmlParserInputRead(xmlParserInputPtr in, int len) {
    int ret;
    int used;
    int indx;

#ifdef DEBUG_INPUT
    xmlGenericError(xmlGenericErrorContext, "Read\n");
#endif
    if (in->buf == NULL) return(-1);
    if (in->base == NULL) return(-1);
    if (in->cur == NULL) return(-1);
    if (in->buf->buffer == NULL) return(-1);
    if (in->buf->readcallback == NULL) return(-1);

    CHECK_BUFFER(in);

    used = in->cur - in->buf->buffer->content;
    ret = xmlBufferShrink(in->buf->buffer, used);
    if (ret > 0) {
	in->cur -= ret;
	in->consumed += ret;
    }
    ret = xmlParserInputBufferRead(in->buf, len);
    if (in->base != in->buf->buffer->content) {
        /*
	 * the buffer has been reallocated
	 */
	indx = in->cur - in->base;
	in->base = in->buf->buffer->content;
	in->cur = &in->buf->buffer->content[indx];
    }
    in->end = &in->buf->buffer->content[in->buf->buffer->use];

    CHECK_BUFFER(in);

    return(ret);
}

/**
 * xmlParserInputGrow:
 * @in:  an XML parser input
 * @len:  an indicative size for the lookahead
 *
 * This function increase the input for the parser. It tries to
 * preserve pointers to the input buffer, and keep already read data
 *
 * Returns the number of xmlChars read, or -1 in case of error, 0 indicate the
 * end of this entity
 */
int
xmlParserInputGrow(xmlParserInputPtr in, int len) {
    int ret;
    int indx;

#ifdef DEBUG_INPUT
    xmlGenericError(xmlGenericErrorContext, "Grow\n");
#endif
    if (in->buf == NULL) return(-1);
    if (in->base == NULL) return(-1);
    if (in->cur == NULL) return(-1);
    if (in->buf->buffer == NULL) return(-1);

    CHECK_BUFFER(in);

    indx = in->cur - in->base;
    if (in->buf->buffer->use > (unsigned int) indx + INPUT_CHUNK) {

	CHECK_BUFFER(in);

        return(0);
    }
    if (in->buf->readcallback != NULL)
	ret = xmlParserInputBufferGrow(in->buf, len);
    else	
        return(0);

    /*
     * NOTE : in->base may be a "dangling" i.e. freed pointer in this
     *        block, but we use it really as an integer to do some
     *        pointer arithmetic. Insure will raise it as a bug but in
     *        that specific case, that's not !
     */
    if (in->base != in->buf->buffer->content) {
        /*
	 * the buffer has been reallocated
	 */
	indx = in->cur - in->base;
	in->base = in->buf->buffer->content;
	in->cur = &in->buf->buffer->content[indx];
    }
    in->end = &in->buf->buffer->content[in->buf->buffer->use];

    CHECK_BUFFER(in);

    return(ret);
}

/**
 * xmlParserInputShrink:
 * @in:  an XML parser input
 *
 * This function removes used input for the parser.
 */
void
xmlParserInputShrink(xmlParserInputPtr in) {
    int used;
    int ret;
    int indx;

#ifdef DEBUG_INPUT
    xmlGenericError(xmlGenericErrorContext, "Shrink\n");
#endif
    if (in->buf == NULL) return;
    if (in->base == NULL) return;
    if (in->cur == NULL) return;
    if (in->buf->buffer == NULL) return;

    CHECK_BUFFER(in);

    used = in->cur - in->buf->buffer->content;
    /*
     * Do not shrink on large buffers whose only a tiny fraction
     * was consumed
     */
    if (used > INPUT_CHUNK) {
	ret = xmlBufferShrink(in->buf->buffer, used - LINE_LEN);
	if (ret > 0) {
	    in->cur -= ret;
	    in->consumed += ret;
	}
	in->end = &in->buf->buffer->content[in->buf->buffer->use];
    }

    CHECK_BUFFER(in);

    if (in->buf->buffer->use > INPUT_CHUNK) {
        return;
    }
    xmlParserInputBufferRead(in->buf, 2 * INPUT_CHUNK);
    if (in->base != in->buf->buffer->content) {
        /*
	 * the buffer has been reallocated
	 */
	indx = in->cur - in->base;
	in->base = in->buf->buffer->content;
	in->cur = &in->buf->buffer->content[indx];
    }
    in->end = &in->buf->buffer->content[in->buf->buffer->use];

    CHECK_BUFFER(in);
}

/************************************************************************
 *									*
 * 		UTF8 character input and related functions		*
 *									*
 ************************************************************************/

/**
 * xmlNextChar:
 * @ctxt:  the XML parser context
 *
 * Skip to the next char input char.
 */

void
xmlNextChar(xmlParserCtxtPtr ctxt)
{
    if (ctxt->instate == XML_PARSER_EOF)
        return;

    if (ctxt->charset == XML_CHAR_ENCODING_UTF8) {
        if ((*ctxt->input->cur == 0) &&
            (xmlParserInputGrow(ctxt->input, INPUT_CHUNK) <= 0) &&
            (ctxt->instate != XML_PARSER_COMMENT)) {
            /*
             * If we are at the end of the current entity and
             * the context allows it, we pop consumed entities
             * automatically.
             * the auto closing should be blocked in other cases
             */
            xmlPopInput(ctxt);
        } else {
            const unsigned char *cur;
            unsigned char c;

            /*
             *   2.11 End-of-Line Handling
             *   the literal two-character sequence "#xD#xA" or a standalone
             *   literal #xD, an XML processor must pass to the application
             *   the single character #xA.
             */
            if (*(ctxt->input->cur) == '\n') {
                ctxt->input->line++;
                ctxt->input->col = 1;
            } else
                ctxt->input->col++;

            /*
             * 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
             */
            cur = ctxt->input->cur;

            c = *cur;
            if (c & 0x80) {
	        if (c == 0xC0)
		    goto encoding_error;
                if (cur[1] == 0)
                    xmlParserInputGrow(ctxt->input, INPUT_CHUNK);
                if ((cur[1] & 0xc0) != 0x80)
                    goto encoding_error;
                if ((c & 0xe0) == 0xe0) {
                    unsigned int val;

                    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;