xpath.c

libxml,在UNIX/LINUX下非常重要的一个库,为XML相关应用提供方便.目前上载的是最新版本,若要取得最新版本,请参考里面的readme.

		    fprintf(output, "'name' "); break;
	    }
	    switch (type) {
                case NODE_TYPE_NODE:
		    fprintf(output, "'node' "); break;
                case NODE_TYPE_COMMENT:
		    fprintf(output, "'comment' "); break;
                case NODE_TYPE_TEXT:
		    fprintf(output, "'text' "); break;
                case NODE_TYPE_PI:
		    fprintf(output, "'PI' "); break;
	    }
	    if (prefix != NULL)
		fprintf(output, "%s:", prefix);
	    if (name != NULL)
		fprintf(output, "%s", (const char *) name);
	    break;

        }
	case XPATH_OP_VALUE: {
	    xmlXPathObjectPtr object = (xmlXPathObjectPtr) op->value4;

	    fprintf(output, "ELEM ");
	    xmlXPathDebugDumpObject(output, object, 0);
	    goto finish;
	}
	case XPATH_OP_VARIABLE: {
	    const xmlChar *prefix = op->value5;
	    const xmlChar *name = op->value4;

	    if (prefix != NULL)
		fprintf(output, "VARIABLE %s:%s", prefix, name);
	    else
		fprintf(output, "VARIABLE %s", name);
	    break;
	}
	case XPATH_OP_FUNCTION: {
	    int nbargs = op->value;
	    const xmlChar *prefix = op->value5;
	    const xmlChar *name = op->value4;

	    if (prefix != NULL)
		fprintf(output, "FUNCTION %s:%s(%d args)",
			prefix, name, nbargs);
	    else
		fprintf(output, "FUNCTION %s(%d args)", name, nbargs);
	    break;
	}
        case XPATH_OP_ARG: fprintf(output, "ARG"); break;
        case XPATH_OP_PREDICATE: fprintf(output, "PREDICATE"); break;
        case XPATH_OP_FILTER: fprintf(output, "FILTER"); break;
#ifdef LIBXML_XPTR_ENABLED
        case XPATH_OP_RANGETO: fprintf(output, "RANGETO"); break;
#endif
	default:
        fprintf(output, "UNKNOWN %d\n", op->op); return;
    }
    fprintf(output, "\n");
finish:
    if (op->ch1 >= 0)
	xmlXPathDebugDumpStepOp(output, comp, &comp->steps[op->ch1], depth + 1);
    if (op->ch2 >= 0)
	xmlXPathDebugDumpStepOp(output, comp, &comp->steps[op->ch2], depth + 1);
}

/**
 * xmlXPathDebugDumpCompExpr:
 * @output:  the FILE * for the output
 * @comp:  the precompiled XPath expression
 * @depth:  the indentation level.
 *
 * Dumps the tree of the compiled XPath expression.
 */
void
xmlXPathDebugDumpCompExpr(FILE *output, xmlXPathCompExprPtr comp,
	                  int depth) {
    int i;
    char shift[100];

    if ((output == NULL) || (comp == NULL)) return;

    for (i = 0;((i < depth) && (i < 25));i++)
        shift[2 * i] = shift[2 * i + 1] = ' ';
    shift[2 * i] = shift[2 * i + 1] = 0;

    fprintf(output, shift);

    if (comp == NULL) {
	fprintf(output, "Compiled Expression is NULL\n");
	return;
    }
    fprintf(output, "Compiled Expression : %d elements\n",
	    comp->nbStep);
    i = comp->last;
    xmlXPathDebugDumpStepOp(output, comp, &comp->steps[i], depth + 1);
}
#endif /* LIBXML_DEBUG_ENABLED */

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

/**
 * valuePop:
 * @ctxt: an XPath evaluation context
 *
 * Pops the top XPath object from the value stack
 *
 * Returns the XPath object just removed
 */
extern xmlXPathObjectPtr
valuePop(xmlXPathParserContextPtr ctxt)
{
    xmlXPathObjectPtr ret;

    if ((ctxt == NULL) || (ctxt->valueNr <= 0))
        return (0);
    ctxt->valueNr--;
    if (ctxt->valueNr > 0)
        ctxt->value = ctxt->valueTab[ctxt->valueNr - 1];
    else
        ctxt->value = NULL;
    ret = ctxt->valueTab[ctxt->valueNr];
    ctxt->valueTab[ctxt->valueNr] = 0;
    return (ret);
}
/**
 * valuePush:
 * @ctxt:  an XPath evaluation context
 * @value:  the XPath object
 *
 * Pushes a new XPath object on top of the value stack
 *
 * returns the number of items on the value stack
 */
extern int
valuePush(xmlXPathParserContextPtr ctxt, xmlXPathObjectPtr value)
{
    if ((ctxt == NULL) || (value == NULL)) return(-1);
    if (ctxt->valueNr >= ctxt->valueMax) {
        xmlXPathObjectPtr *tmp;

        tmp = (xmlXPathObjectPtr *) xmlRealloc(ctxt->valueTab,
                                             2 * ctxt->valueMax *
                                             sizeof(ctxt->valueTab[0]));
        if (tmp == NULL) {
            xmlGenericError(xmlGenericErrorContext, "realloc failed !\n");
            return (0);
        }
        ctxt->valueMax *= 2;
	ctxt->valueTab = tmp;
    }
    ctxt->valueTab[ctxt->valueNr] = value;
    ctxt->value = value;
    return (ctxt->valueNr++);
}

/**
 * xmlXPathPopBoolean:
 * @ctxt:  an XPath parser context
 *
 * Pops a boolean from the stack, handling conversion if needed.
 * Check error with #xmlXPathCheckError.
 *
 * Returns the boolean
 */
int
xmlXPathPopBoolean (xmlXPathParserContextPtr ctxt) {
    xmlXPathObjectPtr obj;
    int ret;

    obj = valuePop(ctxt);
    if (obj == NULL) {
	xmlXPathSetError(ctxt, XPATH_INVALID_OPERAND);
	return(0);
    }
    if (obj->type != XPATH_BOOLEAN)
	ret = xmlXPathCastToBoolean(obj);
    else
        ret = obj->boolval;
    xmlXPathFreeObject(obj);
    return(ret);
}

/**
 * xmlXPathPopNumber:
 * @ctxt:  an XPath parser context
 *
 * Pops a number from the stack, handling conversion if needed.
 * Check error with #xmlXPathCheckError.
 *
 * Returns the number
 */
double
xmlXPathPopNumber (xmlXPathParserContextPtr ctxt) {
    xmlXPathObjectPtr obj;
    double ret;

    obj = valuePop(ctxt);
    if (obj == NULL) {
	xmlXPathSetError(ctxt, XPATH_INVALID_OPERAND);
	return(0);
    }
    if (obj->type != XPATH_NUMBER)
	ret = xmlXPathCastToNumber(obj);
    else
        ret = obj->floatval;
    xmlXPathFreeObject(obj);
    return(ret);
}

/**
 * xmlXPathPopString:
 * @ctxt:  an XPath parser context
 *
 * Pops a string from the stack, handling conversion if needed.
 * Check error with #xmlXPathCheckError.
 *
 * Returns the string
 */
xmlChar *
xmlXPathPopString (xmlXPathParserContextPtr ctxt) {
    xmlXPathObjectPtr obj;
    xmlChar * ret;

    obj = valuePop(ctxt);
    if (obj == NULL) {
	xmlXPathSetError(ctxt, XPATH_INVALID_OPERAND);
	return(NULL);
    }
    ret = xmlXPathCastToString(obj);	/* this does required strdup */
    /* TODO: needs refactoring somewhere else */
    if (obj->stringval == ret)
	obj->stringval = NULL;
    xmlXPathFreeObject(obj);
    return(ret);
}

/**
 * xmlXPathPopNodeSet:
 * @ctxt:  an XPath parser context
 *
 * Pops a node-set from the stack, handling conversion if needed.
 * Check error with #xmlXPathCheckError.
 *
 * Returns the node-set
 */
xmlNodeSetPtr
xmlXPathPopNodeSet (xmlXPathParserContextPtr ctxt) {
    xmlXPathObjectPtr obj;
    xmlNodeSetPtr ret;

    if (ctxt == NULL) return(NULL);
    if (ctxt->value == NULL) {
	xmlXPathSetError(ctxt, XPATH_INVALID_OPERAND);
	return(NULL);
    }
    if (!xmlXPathStackIsNodeSet(ctxt)) {
	xmlXPathSetTypeError(ctxt);
	return(NULL);
    }
    obj = valuePop(ctxt);
    ret = obj->nodesetval;
#if 0
    /* to fix memory leak of not clearing obj->user */
    if (obj->boolval && obj->user != NULL)
        xmlFreeNodeList((xmlNodePtr) obj->user);
#endif
    xmlXPathFreeNodeSetList(obj);
    return(ret);
}

/**
 * xmlXPathPopExternal:
 * @ctxt:  an XPath parser context
 *
 * Pops an external object from the stack, handling conversion if needed.
 * Check error with #xmlXPathCheckError.
 *
 * Returns the object
 */
void *
xmlXPathPopExternal (xmlXPathParserContextPtr ctxt) {
    xmlXPathObjectPtr obj;
    void * ret;

    if ((ctxt == NULL) || (ctxt->value == NULL)) {
	xmlXPathSetError(ctxt, XPATH_INVALID_OPERAND);
	return(NULL);
    }
    if (ctxt->value->type != XPATH_USERS) {
	xmlXPathSetTypeError(ctxt);
	return(NULL);
    }
    obj = valuePop(ctxt);
    ret = obj->user;
    xmlXPathFreeObject(obj);
    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
 *           in ISO-Latin or UTF-8.
 *           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.
 *   SKIP(n) Skip n xmlChar, and must also be used only to skip ASCII defined
 *           strings within the parser.
 *   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.
 *           It returns the pointer to the current xmlChar.
 */

#define CUR (*ctxt->cur)
#define SKIP(val) ctxt->cur += (val)
#define NXT(val) ctxt->cur[(val)]
#define CUR_PTR ctxt->cur
#define CUR_CHAR(l) xmlXPathCurrentChar(ctxt, &l)

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

#define NEXTL(l)  ctxt->cur += l

#define SKIP_BLANKS 							\
    while (IS_BLANK_CH(*(ctxt->cur))) NEXT

#define CURRENT (*ctxt->cur)
#define NEXT ((*ctxt->cur) ?  ctxt->cur++: ctxt->cur)


#ifndef DBL_DIG
#define DBL_DIG 16
#endif
#ifndef DBL_EPSILON
#define DBL_EPSILON 1E-9
#endif

#define UPPER_DOUBLE 1E9
#define LOWER_DOUBLE 1E-5

#define INTEGER_DIGITS DBL_DIG
#define FRACTION_DIGITS (DBL_DIG + 1)
#define EXPONENT_DIGITS (3 + 2)

/**
 * xmlXPathFormatNumber:
 * @number:     number to format
 * @buffer:     output buffer
 * @buffersize: size of output buffer
 *
 * Convert the number into a string representation.
 */
static void
xmlXPathFormatNumber(double number, char buffer[], int buffersize)
{
    switch (xmlXPathIsInf(number)) {
    case 1:
	if (buffersize > (int)sizeof("Infinity"))
	    snprintf(buffer, buffersize, "Infinity");
	break;
    case -1:
	if (buffersize > (int)sizeof("-Infinity"))
	    snprintf(buffer, buffersize, "-Infinity");
	break;
    default:
	if (xmlXPathIsNaN(number)) {
	    if (buffersize > (int)sizeof("NaN"))
		snprintf(buffer, buffersize, "NaN");
	} else if (number == 0 && xmlXPathGetSign(number) != 0) {
	    snprintf(buffer, buffersize, "0");
	} else if (number == ((int) number)) {
	    char work[30];
	    char *ptr, *cur;
	    int res, value = (int) number;

            ptr = &buffer[0];
	    if (value < 0) {
		*ptr++ = '-';
		value = -value;
	    }
	    if (value == 0) {
		*ptr++ = '0';
	    } else {
		cur = &work[0];
		while (value != 0) {
		    res = value % 10;
		    value = value / 10;
		    *cur++ = '0' + res;
		}
		cur--;
		while ((cur >= &work[0]) && (ptr - buffer < buffersize)) {
		    *ptr++ = *cur--;
		}
	    }
	    if (ptr - buffer < buffersize) {
		*ptr = 0;
	    } else if (buffersize > 0) {
		ptr--;
		*ptr = 0;
	    }
	} else {
	    /* 3 is sign, decimal point, and terminating zero */
	    char work[DBL_DIG + EXPONENT_DIGITS + 3];
	    int integer_place, fraction_place;
	    char *ptr;
	    char *after_fraction;
	    double absolute_value;
	    int size;

	    absolute_value = fabs(number);

	    /*
	     * First choose format - scientific or regular floating point.
	     * In either case, result is in work, and after_fraction points
	     * just past the fractional part.
	    */
	    if ( ((absolute_value > UPPER_DOUBLE) ||
		  (absolute_value < LOWER_DOUBLE)) &&
		 (absolute_value != 0.0) ) {
		/* Use scientific notation */
		integer_place = DBL_DIG + EXPONENT_DIGITS + 1;
		fraction_place = DBL_DIG - 1;
		snprintf(work, sizeof(work),"%*.*e",
			 integer_place, fraction_place, number);
		after_fraction = strchr(work + DBL_DIG, 'e');
	    }
	    else {
		/* Use regular notation */
		if (absolute_value > 0.0)
		    integer_place = 1 + (int)log10(absolute_value);
		else
		    integer_place = 0;
		fraction_place = (integer_place > 0)
		    ? DBL_DIG - integer_place
		    : DBL_DIG;
		size = snprintf(work, sizeof(work), "%0.*f",
				fraction_place, number);
		after_fraction = work + size;
	    }

	    /* Remove fractional trailing zeroes */
	    ptr = after_fraction;
	    while (*(--ptr) == '0')
		;
	    if (*ptr != '.')
	        ptr++;
	    while ((*ptr++ = *after_fraction++) != 0);

	    /* Finally copy result back to caller */
	    size = strlen(work) + 1;
	    if (size > buffersize) {
		work[buffersize - 1] = 0;
		size = buffersize;
	    }
	    memmove(buffer, work, size);
	}
	break;
    }
}


/************************************************************************
 *									*
 *			Routines to handle NodeSets			*
 *									*
 ************************************************************************/

/**
 * xmlXPathOrderDocElems:
 * @doc:  an input document
 *
 * Call this routine to speed up XPath computation on static documents.
 * This stamps all the element nodes with the document order
 * Like for line information, the order is kept in the element->content
 * field, the value stored is actually - the node number (starting at -1)
 * to be able to differentiate from line numbers.
 *
 * Returns the number of elements found in the document or -1 in case
 *    of error.
 */
long
xmlXPathOrderDocElems(xmlDocPtr doc) {
    long count = 0;
    xmlNodePtr cur;

    if (doc == NULL)
	return(-1);
    cur = doc->children;
    while (cur != NULL) {