xmlreader.c.svn-base

这是一个用于解析xml文件的类库。使用这个类库

 *          nodes to read, or -1 in case of error
 */
int
xmlTextReaderNextSibling(xmlTextReaderPtr reader) {
    if (reader == NULL)
        return(-1);
    if (reader->doc == NULL) {
        TODO
	return(-1);
    }

    if (reader->state == XML_TEXTREADER_END)
        return(0);

    if (reader->node == NULL)
        return(xmlTextReaderNextTree(reader));

    if (reader->node->next != NULL) {
        reader->node = reader->node->next;
        reader->state = XML_TEXTREADER_START;
        return(1);
    }

    return(0);
}

/************************************************************************
 *									*
 *			Constructor and destructors			*
 *									*
 ************************************************************************/
/**
 * xmlNewTextReader:
 * @input: the xmlParserInputBufferPtr used to read data
 * @URI: the URI information for the source if available
 *
 * Create an xmlTextReader structure fed with @input
 *
 * Returns the new xmlTextReaderPtr or NULL in case of error
 */
xmlTextReaderPtr
xmlNewTextReader(xmlParserInputBufferPtr input, const char *URI) {
    xmlTextReaderPtr ret;

    if (input == NULL)
	return(NULL);
    ret = xmlMalloc(sizeof(xmlTextReader));
    if (ret == NULL) {
        xmlGenericError(xmlGenericErrorContext,
		"xmlNewTextReader : malloc failed\n");
	return(NULL);
    }
    memset(ret, 0, sizeof(xmlTextReader));
    ret->doc = NULL;
    ret->entTab = NULL;
    ret->entMax = 0;
    ret->entNr = 0;
    ret->input = input;
    ret->buffer = xmlBufferCreateSize(100);
    ret->sax = (xmlSAXHandler *) xmlMalloc(sizeof(xmlSAXHandler));
    if (ret->sax == NULL) {
	xmlFree(ret);
        xmlGenericError(xmlGenericErrorContext,
		"xmlNewTextReader : malloc failed\n");
	return(NULL);
    }
    xmlSAXVersion(ret->sax, 2);
    ret->startElement = ret->sax->startElement;
    ret->sax->startElement = xmlTextReaderStartElement;
    ret->endElement = ret->sax->endElement;
    ret->sax->endElement = xmlTextReaderEndElement;
#ifdef LIBXML_SAX1_ENABLED
    if (ret->sax->initialized == XML_SAX2_MAGIC) {
#endif /* LIBXML_SAX1_ENABLED */
	ret->startElementNs = ret->sax->startElementNs;
	ret->sax->startElementNs = xmlTextReaderStartElementNs;
	ret->endElementNs = ret->sax->endElementNs;
	ret->sax->endElementNs = xmlTextReaderEndElementNs;
#ifdef LIBXML_SAX1_ENABLED
    } else {
	ret->startElementNs = NULL;
	ret->endElementNs = NULL;
    }
#endif /* LIBXML_SAX1_ENABLED */
    ret->characters = ret->sax->characters;
    ret->sax->characters = xmlTextReaderCharacters;
    ret->sax->ignorableWhitespace = xmlTextReaderCharacters;
    ret->cdataBlock = ret->sax->cdataBlock;
    ret->sax->cdataBlock = xmlTextReaderCDataBlock;

    ret->mode = XML_TEXTREADER_MODE_INITIAL;
    ret->node = NULL;
    ret->curnode = NULL;
    if (ret->input->buffer->use < 4) {
	xmlParserInputBufferRead(input, 4);
    }
    if (ret->input->buffer->use >= 4) {
	ret->ctxt = xmlCreatePushParserCtxt(ret->sax, NULL,
			(const char *) ret->input->buffer->content, 4, URI);
	ret->base = 0;
	ret->cur = 4;
    } else {
	ret->ctxt = xmlCreatePushParserCtxt(ret->sax, NULL, NULL, 0, URI);
	ret->base = 0;
	ret->cur = 0;
    }
    if (ret->ctxt == NULL) {
        xmlGenericError(xmlGenericErrorContext,
		"xmlNewTextReader : malloc failed\n");
	xmlFree(ret->sax);
	xmlFree(ret);
	return(NULL);
    }
    ret->ctxt->_private = ret;
    ret->ctxt->linenumbers = 1;
    ret->ctxt->dictNames = 1;
    ret->allocs = XML_TEXTREADER_CTXT;
    /*
     * use the parser dictionnary to allocate all elements and attributes names
     */
    ret->ctxt->docdict = 1;
    ret->dict = ret->ctxt->dict;
#ifdef LIBXML_XINCLUDE_ENABLED
    ret->xinclude = 0;
#endif
#ifdef LIBXML_PATTERN_ENABLED
    ret->patternMax = 0;
    ret->patternTab = NULL;
#endif
    return(ret);
}

/**
 * xmlNewTextReaderFilename:
 * @URI: the URI of the resource to process
 *
 * Create an xmlTextReader structure fed with the resource at @URI
 *
 * Returns the new xmlTextReaderPtr or NULL in case of error
 */
xmlTextReaderPtr
xmlNewTextReaderFilename(const char *URI) {
    xmlParserInputBufferPtr input;
    xmlTextReaderPtr ret;
    char *directory = NULL;

    input = xmlParserInputBufferCreateFilename(URI, XML_CHAR_ENCODING_NONE);
    if (input == NULL)
	return(NULL);
    ret = xmlNewTextReader(input, URI);
    if (ret == NULL) {
	xmlFreeParserInputBuffer(input);
	return(NULL);
    }
    ret->allocs |= XML_TEXTREADER_INPUT;
    if (ret->ctxt->directory == NULL)
        directory = xmlParserGetDirectory(URI);
    if ((ret->ctxt->directory == NULL) && (directory != NULL))
        ret->ctxt->directory = (char *) xmlStrdup((xmlChar *) directory);
    if (directory != NULL)
	xmlFree(directory);
    return(ret);
}

/**
 * xmlFreeTextReader:
 * @reader:  the xmlTextReaderPtr
 *
 * Deallocate all the resources associated to the reader
 */
void
xmlFreeTextReader(xmlTextReaderPtr reader) {
    if (reader == NULL)
	return;
#ifdef LIBXML_SCHEMAS_ENABLED
    if (reader->rngSchemas != NULL) {
	xmlRelaxNGFree(reader->rngSchemas);
	reader->rngSchemas = NULL;
    }
    if (reader->rngValidCtxt != NULL) {
	xmlRelaxNGFreeValidCtxt(reader->rngValidCtxt);
	reader->rngValidCtxt = NULL;
    }
#endif
#ifdef LIBXML_XINCLUDE_ENABLED
    if (reader->xincctxt != NULL)
	xmlXIncludeFreeContext(reader->xincctxt);
#endif
#ifdef LIBXML_PATTERN_ENABLED
    if (reader->patternTab != NULL) {
        int i;
	for (i = 0;i < reader->patternNr;i++) {
	    if (reader->patternTab[i] != NULL)
	        xmlFreePattern(reader->patternTab[i]);
	}
	xmlFree(reader->patternTab);
    }
#endif
    if (reader->ctxt != NULL) {
        if (reader->dict == reader->ctxt->dict)
	    reader->dict = NULL;
	if (reader->ctxt->myDoc != NULL) {
	    if (reader->preserve == 0)
		xmlTextReaderFreeDoc(reader, reader->ctxt->myDoc);
	    reader->ctxt->myDoc = NULL;
	}
	if ((reader->ctxt->vctxt.vstateTab != NULL) &&
	    (reader->ctxt->vctxt.vstateMax > 0)){
	    xmlFree(reader->ctxt->vctxt.vstateTab);
	    reader->ctxt->vctxt.vstateTab = 0;
	    reader->ctxt->vctxt.vstateMax = 0;
	}
	if (reader->allocs & XML_TEXTREADER_CTXT)
	    xmlFreeParserCtxt(reader->ctxt);
    }
    if (reader->sax != NULL)
	xmlFree(reader->sax);
    if ((reader->input != NULL)  && (reader->allocs & XML_TEXTREADER_INPUT))
	xmlFreeParserInputBuffer(reader->input);
    if (reader->faketext != NULL) {
	xmlFreeNode(reader->faketext);
    }
    if (reader->buffer != NULL)
        xmlBufferFree(reader->buffer);
    if (reader->entTab != NULL)
	xmlFree(reader->entTab);
    if (reader->dict != NULL)
        xmlDictFree(reader->dict);
    xmlFree(reader);
}

/************************************************************************
 *									*
 *			Methods for XmlTextReader			*
 *									*
 ************************************************************************/
/**
 * xmlTextReaderClose:
 * @reader:  the xmlTextReaderPtr used
 *
 * This method releases any resources allocated by the current instance
 * changes the state to Closed and close any underlying input.
 *
 * Returns 0 or -1 in case of error
 */
int
xmlTextReaderClose(xmlTextReaderPtr reader) {
    if (reader == NULL)
	return(-1);
    reader->node = NULL;
    reader->curnode = NULL;
    reader->mode = XML_TEXTREADER_MODE_CLOSED;
    if (reader->ctxt != NULL) {
	xmlStopParser(reader->ctxt);
	if (reader->ctxt->myDoc != NULL) {
	    if (reader->preserve == 0)
		xmlTextReaderFreeDoc(reader, reader->ctxt->myDoc);
	    reader->ctxt->myDoc = NULL;
	}
    }
    if ((reader->input != NULL)  && (reader->allocs & XML_TEXTREADER_INPUT)) {
	xmlFreeParserInputBuffer(reader->input);
	reader->allocs -= XML_TEXTREADER_INPUT;
    }
    return(0);
}

/**
 * xmlTextReaderGetAttributeNo:
 * @reader:  the xmlTextReaderPtr used
 * @no: the zero-based index of the attribute relative to the containing element
 *
 * Provides the value of the attribute with the specified index relative
 * to the containing element.
 *
 * Returns a string containing the value of the specified attribute, or NULL
 *    in case of error. The string must be deallocated by the caller.
 */
xmlChar *
xmlTextReaderGetAttributeNo(xmlTextReaderPtr reader, int no) {
    xmlChar *ret;
    int i;
    xmlAttrPtr cur;
    xmlNsPtr ns;

    if (reader == NULL)
	return(NULL);
    if (reader->node == NULL)
	return(NULL);
    if (reader->curnode != NULL)
	return(NULL);
    /* TODO: handle the xmlDecl */
    if (reader->node->type != XML_ELEMENT_NODE) 
	return(NULL);

    ns = reader->node->nsDef;
    for (i = 0;(i < no) && (ns != NULL);i++) {
	ns = ns->next;
    }
    if (ns != NULL)
	return(xmlStrdup(ns->href));

    cur = reader->node->properties;
    if (cur == NULL)
	return(NULL);
    for (;i < no;i++) {
	cur = cur->next;
	if (cur == NULL)
	    return(NULL);
    }
    /* TODO walk the DTD if present */

    ret = xmlNodeListGetString(reader->node->doc, cur->children, 1);
    if (ret == NULL) return(xmlStrdup((xmlChar *)""));
    return(ret);
}

/**
 * xmlTextReaderGetAttribute:
 * @reader:  the xmlTextReaderPtr used
 * @name: the qualified name of the attribute.
 *
 * Provides the value of the attribute with the specified qualified name.
 *
 * Returns a string containing the value of the specified attribute, or NULL
 *    in case of error. The string must be deallocated by the caller.
 */
xmlChar *
xmlTextReaderGetAttribute(xmlTextReaderPtr reader, const xmlChar *name) {
    xmlChar *prefix = NULL;
    xmlChar *localname;
    xmlNsPtr ns;
    xmlChar *ret = NULL;

    if ((reader == NULL) || (name == NULL))
	return(NULL);
    if (reader->node == NULL)
	return(NULL);
    if (reader->curnode != NULL)
	return(NULL);

    /* TODO: handle the xmlDecl */
    if (reader->node->type != XML_ELEMENT_NODE)
	return(NULL);

    localname = xmlSplitQName2(name, &prefix);
    if (localname == NULL)
	return(xmlGetProp(reader->node, name));
    
    ns = xmlSearchNs(reader->node->doc, reader->node, prefix);
    if (ns != NULL)
        ret = xmlGetNsProp(reader->node, localname, ns->href);

    if (localname != NULL)
        xmlFree(localname);
    if (prefix != NULL)
        xmlFree(prefix);
    return(ret);
}


/**
 * xmlTextReaderGetAttributeNs:
 * @reader:  the xmlTextReaderPtr used
 * @localName: the local name of the attribute.
 * @namespaceURI: the namespace URI of the attribute.
 *
 * Provides the value of the specified attribute
 *
 * Returns a string containing the value of the specified attribute, or NULL
 *    in case of error. The string must be deallocated by the caller.
 */
xmlChar *
xmlTextReaderGetAttributeNs(xmlTextReaderPtr reader, const xmlChar *localName,
			    const xmlChar *namespaceURI) {
    if ((reader == NULL) || (localName == NULL))
	return(NULL);
    if (reader->node == NULL)
	return(NULL);
    if (reader->curnode != NULL)
	return(NULL);

    /* TODO: handle the xmlDecl */
    if (reader->node->type != XML_ELEMENT_NODE)
	return(NULL);

    return(xmlGetNsProp(reader->node, localName, namespaceURI));
}

/**
 * xmlTextReaderGetRemainder:
 * @reader:  the xmlTextReaderPtr used
 *
 * Method to get the remainder of the buffered XML. this method stops the
 * parser, set its state to End Of File and return the input stream with
 * what is left that the parser did not use.
 *
 * The implementation is not good, the parser certainly procgressed past
 * what's left in reader->input, and there is an allocation problem. Best
 * would be to rewrite it differently.
 *
 * Returns the xmlParserInputBufferPtr attached to the XML or NULL
 *    in case of error.
 */
xmlParserInputBufferPtr
xmlTextReaderGetRemainder(xmlTextReaderPtr reader) {
    xmlParserInputBufferPtr ret = NULL;

    if (reader == NULL)
	return(NULL);
    if (reader->node == NULL)
	return(NULL);

    reader->node = NULL;
    reader->curnode = NULL;
    reader->mode = XML_TEXTREADER_MODE_EOF;
    if (reader->ctxt != NULL) {
	xmlStopParser(reader->ctxt);
	if (reader->ctxt->myDoc != NULL) {
	    if (reader->preserve == 0)
		xmlTextReaderFreeDoc(reader, reader->ctxt->myDoc);
	    reader->ctxt->myDoc = NULL;
	}
    }
    if (reader->allocs & XML_TEXTREADER_INPUT) {
	ret = reader->input;
	reader->input = NULL;
	reader->allocs -= XML_TEXTREADER_INPUT;
    } else {
	/*
	 * Hum, one may need to duplicate the data structure because
	 * without reference counting the input may be freed twice:
	 *   - by the layer which allocated it.
	 *   - by the layer to which would have been returned to.
	 */
	TODO
	return(NULL);
    }
    return(ret);
}

/**
 * xmlTextReaderLookupNamespace:
 * @reader:  the xmlTextReaderPtr used
 * @prefix: the prefix whose namespace URI is to be resolved. To return
 *          the default namespace, specify NULL
 *
 * Resolves a namespace prefix in the scope of the current element.
 *
 * Returns a string containing the namespace URI to which the prefix maps
 *    or NULL in case of error. The string must be deallocated by the caller.
 */
xmlChar *
xmlTextRea