xpointer.c.svn-base

这是一个用于解析xml文件的类库。使用这个类库

xmlXPtrNewCollapsedRange(xmlNodePtr start) {
    xmlXPathObjectPtr ret;

    if (start == NULL)
	return(NULL);

    ret = (xmlXPathObjectPtr) xmlMalloc(sizeof(xmlXPathObject));
    if (ret == NULL) {
        xmlXPtrErrMemory("allocating range");
	return(NULL);
    }
    memset(ret, 0 , (size_t) sizeof(xmlXPathObject));
    ret->type = XPATH_RANGE;
    ret->user = start;
    ret->index = -1;
    ret->user2 = NULL;
    ret->index2 = -1;
    return(ret);
}

/**
 * xmlXPtrNewRangeNodeObject:
 * @start:  the starting node
 * @end:  the ending object
 *
 * Create a new xmlXPathObjectPtr of type range from a not to an object
 *
 * Returns the newly created object.
 */
xmlXPathObjectPtr
xmlXPtrNewRangeNodeObject(xmlNodePtr start, xmlXPathObjectPtr end) {
    xmlXPathObjectPtr ret;

    if (start == NULL)
	return(NULL);
    if (end == NULL)
	return(NULL);
    switch (end->type) {
	case XPATH_POINT:
	case XPATH_RANGE:
	    break;
	case XPATH_NODESET:
	    /*
	     * Empty set ... 
	     */
	    if (end->nodesetval->nodeNr <= 0)
		return(NULL);
	    break;
	default:
	    TODO
	    return(NULL);
    }

    ret = (xmlXPathObjectPtr) xmlMalloc(sizeof(xmlXPathObject));
    if (ret == NULL) {
        xmlXPtrErrMemory("allocating range");
	return(NULL);
    }
    memset(ret, 0 , (size_t) sizeof(xmlXPathObject));
    ret->type = XPATH_RANGE;
    ret->user = start;
    ret->index = -1;
    switch (end->type) {
	case XPATH_POINT:
	    ret->user2 = end->user;
	    ret->index2 = end->index;
	    break;
	case XPATH_RANGE:
	    ret->user2 = end->user2;
	    ret->index2 = end->index2;
	    break;
	case XPATH_NODESET: {
	    ret->user2 = end->nodesetval->nodeTab[end->nodesetval->nodeNr - 1];
	    ret->index2 = -1;
	    break;
	}
	default:
	    STRANGE
	    return(NULL);
    }
    xmlXPtrRangeCheckOrder(ret);
    return(ret);
}

#define XML_RANGESET_DEFAULT	10

/**
 * xmlXPtrLocationSetCreate:
 * @val:  an initial xmlXPathObjectPtr, or NULL
 *
 * Create a new xmlLocationSetPtr of type double and of value @val
 *
 * Returns the newly created object.
 */
xmlLocationSetPtr
xmlXPtrLocationSetCreate(xmlXPathObjectPtr val) {
    xmlLocationSetPtr ret;

    ret = (xmlLocationSetPtr) xmlMalloc(sizeof(xmlLocationSet));
    if (ret == NULL) {
        xmlXPtrErrMemory("allocating locationset");
	return(NULL);
    }
    memset(ret, 0 , (size_t) sizeof(xmlLocationSet));
    if (val != NULL) {
        ret->locTab = (xmlXPathObjectPtr *) xmlMalloc(XML_RANGESET_DEFAULT *
					     sizeof(xmlXPathObjectPtr));
	if (ret->locTab == NULL) {
	    xmlXPtrErrMemory("allocating locationset");
	    xmlFree(ret);
	    return(NULL);
	}
	memset(ret->locTab, 0 ,
	       XML_RANGESET_DEFAULT * (size_t) sizeof(xmlXPathObjectPtr));
        ret->locMax = XML_RANGESET_DEFAULT;
	ret->locTab[ret->locNr++] = val;
    }
    return(ret);
}

/**
 * xmlXPtrLocationSetAdd:
 * @cur:  the initial range set
 * @val:  a new xmlXPathObjectPtr
 *
 * add a new xmlXPathObjectPtr to an existing LocationSet
 * If the location already exist in the set @val is freed.
 */
void
xmlXPtrLocationSetAdd(xmlLocationSetPtr cur, xmlXPathObjectPtr val) {
    int i;

    if (val == NULL) return;

    /*
     * check against doublons
     */
    for (i = 0;i < cur->locNr;i++) {
	if (xmlXPtrRangesEqual(cur->locTab[i], val)) {
	    xmlXPathFreeObject(val);
	    return;
	}
    }

    /*
     * grow the locTab if needed
     */
    if (cur->locMax == 0) {
        cur->locTab = (xmlXPathObjectPtr *) xmlMalloc(XML_RANGESET_DEFAULT *
					     sizeof(xmlXPathObjectPtr));
	if (cur->locTab == NULL) {
	    xmlXPtrErrMemory("adding location to set");
	    return;
	}
	memset(cur->locTab, 0 ,
	       XML_RANGESET_DEFAULT * (size_t) sizeof(xmlXPathObjectPtr));
        cur->locMax = XML_RANGESET_DEFAULT;
    } else if (cur->locNr == cur->locMax) {
        xmlXPathObjectPtr *temp;

        cur->locMax *= 2;
	temp = (xmlXPathObjectPtr *) xmlRealloc(cur->locTab, cur->locMax *
				      sizeof(xmlXPathObjectPtr));
	if (temp == NULL) {
	    xmlXPtrErrMemory("adding location to set");
	    return;
	}
	cur->locTab = temp;
    }
    cur->locTab[cur->locNr++] = val;
}

/**
 * xmlXPtrLocationSetMerge:
 * @val1:  the first LocationSet
 * @val2:  the second LocationSet
 *
 * Merges two rangesets, all ranges from @val2 are added to @val1
 *
 * Returns val1 once extended or NULL in case of error.
 */
xmlLocationSetPtr
xmlXPtrLocationSetMerge(xmlLocationSetPtr val1, xmlLocationSetPtr val2) {
    int i;

    if (val1 == NULL) return(NULL);
    if (val2 == NULL) return(val1);

    /*
     * !!!!! this can be optimized a lot, knowing that both
     *       val1 and val2 already have unicity of their values.
     */

    for (i = 0;i < val2->locNr;i++)
        xmlXPtrLocationSetAdd(val1, val2->locTab[i]);

    return(val1);
}

/**
 * xmlXPtrLocationSetDel:
 * @cur:  the initial range set
 * @val:  an xmlXPathObjectPtr
 *
 * Removes an xmlXPathObjectPtr from an existing LocationSet
 */
void
xmlXPtrLocationSetDel(xmlLocationSetPtr cur, xmlXPathObjectPtr val) {
    int i;

    if (cur == NULL) return;
    if (val == NULL) return;

    /*
     * check against doublons
     */
    for (i = 0;i < cur->locNr;i++)
        if (cur->locTab[i] == val) break;

    if (i >= cur->locNr) {
#ifdef DEBUG
        xmlGenericError(xmlGenericErrorContext, 
	        "xmlXPtrLocationSetDel: Range wasn't found in RangeList\n");
#endif
        return;
    }
    cur->locNr--;
    for (;i < cur->locNr;i++)
        cur->locTab[i] = cur->locTab[i + 1];
    cur->locTab[cur->locNr] = NULL;
}

/**
 * xmlXPtrLocationSetRemove:
 * @cur:  the initial range set
 * @val:  the index to remove
 *
 * Removes an entry from an existing LocationSet list.
 */
void
xmlXPtrLocationSetRemove(xmlLocationSetPtr cur, int val) {
    if (cur == NULL) return;
    if (val >= cur->locNr) return;
    cur->locNr--;
    for (;val < cur->locNr;val++)
        cur->locTab[val] = cur->locTab[val + 1];
    cur->locTab[cur->locNr] = NULL;
}

/**
 * xmlXPtrFreeLocationSet:
 * @obj:  the xmlLocationSetPtr to free
 *
 * Free the LocationSet compound (not the actual ranges !).
 */
void
xmlXPtrFreeLocationSet(xmlLocationSetPtr obj) {
    int i;

    if (obj == NULL) return;
    if (obj->locTab != NULL) {
	for (i = 0;i < obj->locNr; i++) {
            xmlXPathFreeObject(obj->locTab[i]);
	}
	xmlFree(obj->locTab);
    }
    xmlFree(obj);
}

/**
 * xmlXPtrNewLocationSetNodes:
 * @start:  the start NodePtr value
 * @end:  the end NodePtr value or NULL
 *
 * Create a new xmlXPathObjectPtr of type LocationSet and initialize
 * it with the single range made of the two nodes @start and @end
 *
 * Returns the newly created object.
 */
xmlXPathObjectPtr
xmlXPtrNewLocationSetNodes(xmlNodePtr start, xmlNodePtr end) {
    xmlXPathObjectPtr ret;

    ret = (xmlXPathObjectPtr) xmlMalloc(sizeof(xmlXPathObject));
    if (ret == NULL) {
        xmlXPtrErrMemory("allocating locationset");
	return(NULL);
    }
    memset(ret, 0 , (size_t) sizeof(xmlXPathObject));
    ret->type = XPATH_LOCATIONSET;
    if (end == NULL)
	ret->user = xmlXPtrLocationSetCreate(xmlXPtrNewCollapsedRange(start));
    else
	ret->user = xmlXPtrLocationSetCreate(xmlXPtrNewRangeNodes(start,end));
    return(ret);
}

/**
 * xmlXPtrNewLocationSetNodeSet:
 * @set:  a node set
 *
 * Create a new xmlXPathObjectPtr of type LocationSet and initialize
 * it with all the nodes from @set
 *
 * Returns the newly created object.
 */
xmlXPathObjectPtr
xmlXPtrNewLocationSetNodeSet(xmlNodeSetPtr set) {
    xmlXPathObjectPtr ret;

    ret = (xmlXPathObjectPtr) xmlMalloc(sizeof(xmlXPathObject));
    if (ret == NULL) {
        xmlXPtrErrMemory("allocating locationset");
	return(NULL);
    }
    memset(ret, 0 , (size_t) sizeof(xmlXPathObject));
    ret->type = XPATH_LOCATIONSET;
    if (set != NULL) {
	int i;
	xmlLocationSetPtr newset;

	newset = xmlXPtrLocationSetCreate(NULL);
	if (newset == NULL)
	    return(ret);

	for (i = 0;i < set->nodeNr;i++)
	    xmlXPtrLocationSetAdd(newset,
		        xmlXPtrNewCollapsedRange(set->nodeTab[i]));

	ret->user = (void *) newset;
    }
    return(ret);
}

/**
 * xmlXPtrWrapLocationSet:
 * @val:  the LocationSet value
 *
 * Wrap the LocationSet @val in a new xmlXPathObjectPtr
 *
 * Returns the newly created object.
 */
xmlXPathObjectPtr
xmlXPtrWrapLocationSet(xmlLocationSetPtr val) {
    xmlXPathObjectPtr ret;

    ret = (xmlXPathObjectPtr) xmlMalloc(sizeof(xmlXPathObject));
    if (ret == NULL) {
        xmlXPtrErrMemory("allocating locationset");
	return(NULL);
    }
    memset(ret, 0 , (size_t) sizeof(xmlXPathObject));
    ret->type = XPATH_LOCATIONSET;
    ret->user = (void *) val;
    return(ret);
}

/************************************************************************
 *									*
 *			The parser					*
 *									*
 ************************************************************************/

static void xmlXPtrEvalChildSeq(xmlXPathParserContextPtr ctxt, xmlChar *name);

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

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

/*
 * xmlXPtrGetChildNo:
 * @ctxt:  the XPointer Parser context
 * @index:  the child number
 *
 * Move the current node of the nodeset on the stack to the
 * given child if found
 */
static void
xmlXPtrGetChildNo(xmlXPathParserContextPtr ctxt, int indx) {
    xmlNodePtr cur = NULL;
    xmlXPathObjectPtr obj;
    xmlNodeSetPtr oldset;

    CHECK_TYPE(XPATH_NODESET);
    obj = valuePop(ctxt);
    oldset = obj->nodesetval;
    if ((indx <= 0) || (oldset == NULL) || (oldset->nodeNr != 1)) {
	xmlXPathFreeObject(obj);
	valuePush(ctxt, xmlXPathNewNodeSet(NULL));
	return;
    }
    cur = xmlXPtrGetNthChild(oldset->nodeTab[0], indx);
    if (cur == NULL) {
	xmlXPathFreeObject(obj);
	valuePush(ctxt, xmlXPathNewNodeSet(NULL));
	return;
    }
    oldset->nodeTab[0] = cur;
    valuePush(ctxt, obj);
}

/**
 * xmlXPtrEvalXPtrPart:
 * @ctxt:  the XPointer Parser context
 * @name:  the preparsed Scheme for the XPtrPart
 * 
 * XPtrPart ::= 'xpointer' '(' XPtrExpr ')'
 *            | Scheme '(' SchemeSpecificExpr ')'
 *
 * Scheme   ::=  NCName - 'xpointer' [VC: Non-XPointer schemes]
 *
 * SchemeSpecificExpr ::= StringWithBalancedParens
 *
 * StringWithBalancedParens ::=  
 *              [^()]* ('(' StringWithBalancedParens ')' [^()]*)*
 *              [VC: Parenthesis escaping]
 *
 * XPtrExpr ::= Expr [VC: Parenthesis escaping]
 *
 * VC: Parenthesis escaping:
 *   The end of an XPointer part is signaled by the right parenthesis ")"
 *   character that is balanced with the left parenthesis "(" character
 *   that began the part. Any unbalanced parenthesis character inside the
 *   expression, even within literals, must be escaped with a circumflex (^)
 *   character preceding it. If the expression contains any literal
 *   occurrences of the circumflex, each must be escaped with an additional
 *   circumflex (that is, ^^). If the unescaped parentheses in the expression
 *   are not balanced, a syntax error results.
 *
 * Parse and evaluate an XPtrPart. Basically it generates the unescaped
 * string and if the scheme is 'xpointer' it will call the XPath interpreter.
 * 
 * TODO: there is no new scheme registration mechanism
 */

static void
xmlXPtrEvalXPtrPart(xmlXPathParserContextPtr ctxt, xmlChar *name) {
    xmlChar *buffer, *cur;
    int len;
    int level;

    if (name == NULL)
    name = xmlXPathParseName(ctxt);
    if (name == NULL)
	XP_ERROR(XPATH_EXPR_ERROR);

    if (CUR != '(')
	XP_ERROR(XPATH_EXPR_ERROR);
    NEXT;
    level = 1;

    len = xmlStrlen(ctxt->cur);
    len++;
    buffer = (xmlChar *) xmlMallocAtomic(len * sizeof (xmlChar));
    if (buffer == NULL) {
        xmlXPtrErrMemory("allocating buffer");
	return;
    }

    cur = buffer;
    while (CUR != 0) {
	if (CUR == ')') {
	    level--;
	    if (level == 0) {
		NEXT;
		break;
	    }
	    *cur++ = CUR;
	} else if (CUR == '(') {
	    level++;
	    *cur++ = CUR;
	} else if (CUR == '^') {
	    NEXT;
	    if ((CUR == ')') || (CUR == '(') || (CUR == '^')) {
		*cur++ = CUR;
	    } else {
		*cur++ = '^';
		*cur++ = CUR;
	    }
	} else {
	    *cur++ = CUR;
	}