xmlsave.c

xml开源解析代码.版本为libxml2-2.6.29,可支持GB3212.网络消息发送XML时很有用.

     */
    if ((encoding != NULL) && (oldctxtenc == NULL) &&
	((ctxt->options & XML_SAVE_NO_DECL) == 0)) {
	if ((enc != XML_CHAR_ENCODING_UTF8) &&
	    (enc != XML_CHAR_ENCODING_NONE) &&
	    (enc != XML_CHAR_ENCODING_ASCII)) {
	    xmlOutputBufferFlush(buf);
	    xmlCharEncCloseFunc(buf->encoder);
	    xmlBufferFree(buf->conv);
	    buf->encoder = NULL;
	    buf->conv = NULL;
	}
	ctxt->escape = oldescape;
	ctxt->escapeAttr = oldescapeAttr;
    }
    return(0);
}

#ifdef LIBXML_HTML_ENABLED
/************************************************************************
 *									*
 *		Functions specific to XHTML serialization		*
 *									*
 ************************************************************************/

/**
 * xhtmlIsEmpty:
 * @node:  the node
 *
 * Check if a node is an empty xhtml node
 *
 * Returns 1 if the node is an empty node, 0 if not and -1 in case of error
 */
static int
xhtmlIsEmpty(xmlNodePtr node) {
    if (node == NULL)
	return(-1);
    if (node->type != XML_ELEMENT_NODE)
	return(0);
    if ((node->ns != NULL) && (!xmlStrEqual(node->ns->href, XHTML_NS_NAME)))
	return(0);
    if (node->children != NULL)
	return(0);
    switch (node->name[0]) {
	case 'a':
	    if (xmlStrEqual(node->name, BAD_CAST "area"))
		return(1);
	    return(0);
	case 'b':
	    if (xmlStrEqual(node->name, BAD_CAST "br"))
		return(1);
	    if (xmlStrEqual(node->name, BAD_CAST "base"))
		return(1);
	    if (xmlStrEqual(node->name, BAD_CAST "basefont"))
		return(1);
	    return(0);
	case 'c':
	    if (xmlStrEqual(node->name, BAD_CAST "col"))
		return(1);
	    return(0);
	case 'f':
	    if (xmlStrEqual(node->name, BAD_CAST "frame"))
		return(1);
	    return(0);
	case 'h':
	    if (xmlStrEqual(node->name, BAD_CAST "hr"))
		return(1);
	    return(0);
	case 'i':
	    if (xmlStrEqual(node->name, BAD_CAST "img"))
		return(1);
	    if (xmlStrEqual(node->name, BAD_CAST "input"))
		return(1);
	    if (xmlStrEqual(node->name, BAD_CAST "isindex"))
		return(1);
	    return(0);
	case 'l':
	    if (xmlStrEqual(node->name, BAD_CAST "link"))
		return(1);
	    return(0);
	case 'm':
	    if (xmlStrEqual(node->name, BAD_CAST "meta"))
		return(1);
	    return(0);
	case 'p':
	    if (xmlStrEqual(node->name, BAD_CAST "param"))
		return(1);
	    return(0);
    }
    return(0);
}

/**
 * xhtmlAttrListDumpOutput:
 * @cur:  the first attribute pointer
 *
 * Dump a list of XML attributes
 */
static void
xhtmlAttrListDumpOutput(xmlSaveCtxtPtr ctxt, xmlAttrPtr cur) {
    xmlAttrPtr xml_lang = NULL;
    xmlAttrPtr lang = NULL;
    xmlAttrPtr name = NULL;
    xmlAttrPtr id = NULL;
    xmlNodePtr parent;
    xmlOutputBufferPtr buf;

    if (cur == NULL) return;
    buf = ctxt->buf;
    parent = cur->parent;
    while (cur != NULL) {
	if ((cur->ns == NULL) && (xmlStrEqual(cur->name, BAD_CAST "id")))
	    id = cur;
	else
	if ((cur->ns == NULL) && (xmlStrEqual(cur->name, BAD_CAST "name")))
	    name = cur;
	else
	if ((cur->ns == NULL) && (xmlStrEqual(cur->name, BAD_CAST "lang")))
	    lang = cur;
	else
	if ((cur->ns != NULL) && (xmlStrEqual(cur->name, BAD_CAST "lang")) &&
	    (xmlStrEqual(cur->ns->prefix, BAD_CAST "xml")))
	    xml_lang = cur;
	else if ((cur->ns == NULL) && 
		 ((cur->children == NULL) ||
		  (cur->children->content == NULL) ||
		  (cur->children->content[0] == 0)) &&
		 (htmlIsBooleanAttr(cur->name))) {
	    if (cur->children != NULL)
		xmlFreeNode(cur->children);
	    cur->children = xmlNewText(cur->name);
	    if (cur->children != NULL)
		cur->children->parent = (xmlNodePtr) cur;
	}
        xmlAttrDumpOutput(ctxt, cur);
	cur = cur->next;
    }
    /*
     * C.8
     */
    if ((name != NULL) && (id == NULL)) {
	if ((parent != NULL) && (parent->name != NULL) &&
	    ((xmlStrEqual(parent->name, BAD_CAST "a")) ||
	     (xmlStrEqual(parent->name, BAD_CAST "p")) ||
	     (xmlStrEqual(parent->name, BAD_CAST "div")) ||
	     (xmlStrEqual(parent->name, BAD_CAST "img")) ||
	     (xmlStrEqual(parent->name, BAD_CAST "map")) ||
	     (xmlStrEqual(parent->name, BAD_CAST "applet")) ||
	     (xmlStrEqual(parent->name, BAD_CAST "form")) ||
	     (xmlStrEqual(parent->name, BAD_CAST "frame")) ||
	     (xmlStrEqual(parent->name, BAD_CAST "iframe")))) {
	    xmlOutputBufferWrite(buf, 5, " id=\"");
	    xmlAttrSerializeContent(buf, name);
	    xmlOutputBufferWrite(buf, 1, "\"");
	}
    }
    /*
     * C.7.
     */
    if ((lang != NULL) && (xml_lang == NULL)) {
	xmlOutputBufferWrite(buf, 11, " xml:lang=\"");
	xmlAttrSerializeContent(buf, lang);
	xmlOutputBufferWrite(buf, 1, "\"");
    } else 
    if ((xml_lang != NULL) && (lang == NULL)) {
	xmlOutputBufferWrite(buf, 7, " lang=\"");
	xmlAttrSerializeContent(buf, xml_lang);
	xmlOutputBufferWrite(buf, 1, "\"");
    }
}

/**
 * xhtmlNodeListDumpOutput:
 * @buf:  the XML buffer output
 * @doc:  the XHTML document
 * @cur:  the first node
 * @level: the imbrication level for indenting
 * @format: is formatting allowed
 * @encoding:  an optional encoding string
 *
 * Dump an XML node list, recursive behaviour, children are printed too.
 * Note that @format = 1 provide node indenting only if xmlIndentTreeOutput = 1
 * or xmlKeepBlanksDefault(0) was called
 */
static void
xhtmlNodeListDumpOutput(xmlSaveCtxtPtr ctxt, xmlNodePtr cur) {
    xmlOutputBufferPtr buf;

    if (cur == NULL) return;
    buf = ctxt->buf;
    while (cur != NULL) {
	if ((ctxt->format) && (xmlIndentTreeOutput) &&
	    (cur->type == XML_ELEMENT_NODE))
	    xmlOutputBufferWrite(buf, ctxt->indent_size *
	                         (ctxt->level > ctxt->indent_nr ? 
				  ctxt->indent_nr : ctxt->level),
				 ctxt->indent);
        xhtmlNodeDumpOutput(ctxt, cur);
	if (ctxt->format) {
	    xmlOutputBufferWrite(buf, 1, "\n");
	}
	cur = cur->next;
    }
}

/**
 * xhtmlNodeDumpOutput:
 * @buf:  the XML buffer output
 * @doc:  the XHTML document
 * @cur:  the current node
 * @level: the imbrication level for indenting
 * @format: is formatting allowed
 * @encoding:  an optional encoding string
 *
 * Dump an XHTML node, recursive behaviour, children are printed too.
 */
static void
xhtmlNodeDumpOutput(xmlSaveCtxtPtr ctxt, xmlNodePtr cur) {
    int format, addmeta = 0;
    xmlNodePtr tmp;
    xmlChar *start, *end;
    xmlOutputBufferPtr buf;

    if (cur == NULL) return;
    if ((cur->type == XML_DOCUMENT_NODE) ||
        (cur->type == XML_HTML_DOCUMENT_NODE)) {
        xmlDocContentDumpOutput(ctxt, (xmlDocPtr) cur);
	return;
    }
    if (cur->type == XML_XINCLUDE_START)
	return;
    if (cur->type == XML_XINCLUDE_END)
	return;
    if (cur->type == XML_DTD_NODE) {
        xmlDtdDumpOutput(ctxt, (xmlDtdPtr) cur);
	return;
    }
    if (cur->type == XML_DOCUMENT_FRAG_NODE) {
        xhtmlNodeListDumpOutput(ctxt, cur->children);
	return;
    }
    buf = ctxt->buf;
    if (cur->type == XML_ELEMENT_DECL) {
        xmlDumpElementDecl(buf->buffer, (xmlElementPtr) cur);
	return;
    }
    if (cur->type == XML_ATTRIBUTE_DECL) {
        xmlDumpAttributeDecl(buf->buffer, (xmlAttributePtr) cur);
	return;
    }
    if (cur->type == XML_ENTITY_DECL) {
        xmlDumpEntityDecl(buf->buffer, (xmlEntityPtr) cur);
	return;
    }
    if (cur->type == XML_TEXT_NODE) {
	if (cur->content != NULL) {
	    if ((cur->name == xmlStringText) ||
		(cur->name != xmlStringTextNoenc)) {
                xmlOutputBufferWriteEscape(buf, cur->content, ctxt->escape);
	    } else {
		/*
		 * Disable escaping, needed for XSLT
		 */
		xmlOutputBufferWriteString(buf, (const char *) cur->content);
	    }
	}

	return;
    }
    if (cur->type == XML_PI_NODE) {
	if (cur->content != NULL) {
	    xmlOutputBufferWrite(buf, 2, "<?");
	    xmlOutputBufferWriteString(buf, (const char *)cur->name);
	    if (cur->content != NULL) {
		xmlOutputBufferWrite(buf, 1, " ");
		xmlOutputBufferWriteString(buf, (const char *)cur->content);
	    }
	    xmlOutputBufferWrite(buf, 2, "?>");
	} else {
	    xmlOutputBufferWrite(buf, 2, "<?");
	    xmlOutputBufferWriteString(buf, (const char *)cur->name);
	    xmlOutputBufferWrite(buf, 2, "?>");
	}
	return;
    }
    if (cur->type == XML_COMMENT_NODE) {
	if (cur->content != NULL) {
	    xmlOutputBufferWrite(buf, 4, "<!--");
	    xmlOutputBufferWriteString(buf, (const char *)cur->content);
	    xmlOutputBufferWrite(buf, 3, "-->");
	}
	return;
    }
    if (cur->type == XML_ENTITY_REF_NODE) {
        xmlOutputBufferWrite(buf, 1, "&");
	xmlOutputBufferWriteString(buf, (const char *)cur->name);
        xmlOutputBufferWrite(buf, 1, ";");
	return;
    }
    if (cur->type == XML_CDATA_SECTION_NODE) {
	start = end = cur->content;
	while (*end != '\0') {
	    if (*end == ']' && *(end + 1) == ']' && *(end + 2) == '>') {
		end = end + 2;
		xmlOutputBufferWrite(buf, 9, "<![CDATA[");
		xmlOutputBufferWrite(buf, end - start, (const char *)start);
		xmlOutputBufferWrite(buf, 3, "]]>");
		start = end;
	    }
	    end++;
	}
	if (start != end) {
	    xmlOutputBufferWrite(buf, 9, "<![CDATA[");
	    xmlOutputBufferWriteString(buf, (const char *)start);
	    xmlOutputBufferWrite(buf, 3, "]]>");
	}
	return;
    }

    format = ctxt->format;
    if (format == 1) {
	tmp = cur->children;
	while (tmp != NULL) {
	    if ((tmp->type == XML_TEXT_NODE) || 
		(tmp->type == XML_ENTITY_REF_NODE)) {
		format = 0;
		break;
	    }
	    tmp = tmp->next;
	}
    }
    xmlOutputBufferWrite(buf, 1, "<");
    if ((cur->ns != NULL) && (cur->ns->prefix != NULL)) {
        xmlOutputBufferWriteString(buf, (const char *)cur->ns->prefix);
	xmlOutputBufferWrite(buf, 1, ":");
    }

    xmlOutputBufferWriteString(buf, (const char *)cur->name);
    if (cur->nsDef)
        xmlNsListDumpOutput(buf, cur->nsDef);
    if ((xmlStrEqual(cur->name, BAD_CAST "html") &&
	(cur->ns == NULL) && (cur->nsDef == NULL))) {
	/*
	 * 3.1.1. Strictly Conforming Documents A.3.1.1 3/
	 */
	xmlOutputBufferWriteString(buf,
		" xmlns=\"http://www.w3.org/1999/xhtml\"");
    }
    if (cur->properties != NULL)
        xhtmlAttrListDumpOutput(ctxt, cur->properties);

	if ((cur->type == XML_ELEMENT_NODE) && 
		(cur->parent != NULL) && 
		(cur->parent->parent == (xmlNodePtr) cur->doc) && 
		xmlStrEqual(cur->name, BAD_CAST"head") && 
		xmlStrEqual(cur->parent->name, BAD_CAST"html")) {

		tmp = cur->children;
		while (tmp != NULL) {
			if (xmlStrEqual(tmp->name, BAD_CAST"meta")) {
				xmlChar *httpequiv;

				httpequiv = xmlGetProp(tmp, BAD_CAST"http-equiv");
				if (httpequiv != NULL) {
					if (xmlStrcasecmp(httpequiv, BAD_CAST"Content-Type") == 0) {
						xmlFree(httpequiv);
						break;
					}
					xmlFree(httpequiv);
				}
			}
			tmp = tmp->next;
		}
		if (tmp == NULL)
			addmeta = 1;
	}

    if ((cur->type == XML_ELEMENT_NODE) && (cur->children == NULL)) {
	if (((cur->ns == NULL) || (cur->ns->prefix == NULL)) &&
	    ((xhtmlIsEmpty(cur) == 1) && (addmeta == 0))) {
	    /*
	     * C.2. Empty Elements
	     */
	    xmlOutputBufferWrite(buf, 3, " />");
	} else {
		if (addmeta == 1) {
			xmlOutputBufferWrite(buf, 1, ">");
			if (ctxt->format) {
				xmlOutputBufferWrite(buf, 1, "\n");
				if (xmlIndentTreeOutput)
					xmlOutputBufferWrite(buf, ctxt->indent_size *
					(ctxt->level + 1 > ctxt->indent_nr ? 
					ctxt->indent_nr : ctxt->level + 1), ctxt->indent);
			}
			xmlOutputBufferWriteString(buf,
				"<meta http-equiv=\"Content-Type\" content=\"text/html; charset=");
			if (ctxt->encoding) {
				xmlOutputBufferWriteString(buf, (const char *)ctxt->encoding);
			} else {
				xmlOutputBufferWrite(buf, 5, "UTF-8");
			}
			xmlOutputBufferWrite(buf, 4, "\" />");
			if (ctxt->format)
				xmlOutputBufferWrite(buf, 1, "\n");
		} else {
			xmlOutputBufferWrite(buf, 1, ">");
		}
	    /*
	     * C.3. Element Minimization and Empty Element Content
	     */
	    xmlOutputBufferWrite(buf, 2, "</");
	    if ((cur->ns != NULL) && (cur->ns->prefix != NULL)) {
		xmlOutputBufferWriteString(buf, (const char *)cur->ns->prefix);
		xmlOutputBufferWrite(buf, 1, ":");
	    }
	    xmlOutputBufferWriteString(buf, (const char *)cur->name);
	    xmlOutputBufferWrite(buf, 1, ">");
	}
	return;
    }
    xmlOutputBufferWrite(buf, 1, ">");
	if (addmeta == 1) {
		if (ctxt->format) {
			xmlOutputBufferWrite(buf, 1, "\n");
			if (xmlIndentTreeOutput)
				xmlOutputBufferWrite(buf, ctxt->indent_size *
				(ctxt->level + 1 > ctxt->indent_nr ? 
				ctxt->indent_nr : ctxt->level + 1), ctxt->indent);
		}
		xmlOutputBufferWriteString(buf,
			"<meta http-equiv=\"Content-Type\" content=\"text/html; charset=");
		if (ctxt->encoding) {
			xmlOutputBufferWriteString(buf, (const char *)ctxt->encoding);
		} else {
			xmlOutputBufferWrite(buf, 5, "UTF-8");
		}
		xmlOutputBufferWrite(buf, 4, "\" />");
	}
    if ((cur->type != XML_ELEMENT_NODE) && (cur->content != NULL)) {
	xmlOutputBufferWriteEscape(buf, cur->content, ctxt->escape);
    }

#if 0
    /*
    * This was removed due to problems with HTML processors.
    * See bug #345147.
    */
    /*
     * 4.8. Script and Style elements
     */
    if ((cur->type == XML_ELEMENT_NODE) &&
	((xmlStrEqual(cur->name, BAD_CAST "script")) ||
	 (xmlStrEqual(cur->name, BAD_CAST "style"))) &&
	((cur->ns == NULL) ||
	 (xmlStrEqual(cur->ns->href, XHTML_NS_NAME)))) {
	xmlNodePtr child = cur->children;

	while (child != NULL) {
	    if (child->type == XML_TEXT_NODE) {
		if ((xmlStrchr(child->content, '<') == NULL) &&
		    (xmlStrchr(child->content, '&') == NULL) &&
		    (xmlStrstr(child->content, BAD_CAST "]]>") == NULL)) {
		    /* Nothing to escape, so just output as is... */
		    /* FIXME: Should we do something about "--" also? */
		    int level = ctxt->level;
		    int indent = ctxt->format;

		    ctxt->level = 0;
		    ctxt->format = 0;