Implement PHP-specific extensions to Dom (#14754)

See RFC: https://wiki.php.net/rfc/dom_additions_84

Author: nielsdos

ext/dom/dom_ce.h

@@ -69,5 +69,6 @@ extern PHP_DOM_EXPORT zend_class_entry *dom_modern_xpath_class_entry;
 #endif
 extern PHP_DOM_EXPORT zend_class_entry *dom_namespace_node_class_entry;
 extern PHP_DOM_EXPORT zend_class_entry *dom_adjacent_position_class_entry;
+extern PHP_DOM_EXPORT zend_class_entry *dom_namespace_info_class_entry;
 
 #endif /* DOM_CE_H */

ext/dom/dom_properties.h

@@ -86,6 +86,8 @@ zend_result dom_element_schema_type_info_read(dom_object *obj, zval *retval);
 zend_result dom_element_inner_html_read(dom_object *obj, zval *retval);
 zend_result dom_element_inner_html_write(dom_object *obj, zval *newval);
 zend_result dom_element_class_list_read(dom_object *obj, zval *retval);
+zend_result dom_modern_element_substituted_node_value_read(dom_object *obj, zval *retval);
+zend_result dom_modern_element_substituted_node_value_write(dom_object *obj, zval *newval);
 
 /* entity properties */
 zend_result dom_entity_public_id_read(dom_object *obj, zval *retval);

ext/dom/element.c

@@ -1842,4 +1842,219 @@ PHP_METHOD(Dom_Element, closest)
 	dom_element_closest(thisp, intern, return_value, selectors_str);
 }
 
+zend_result dom_modern_element_substituted_node_value_read(dom_object *obj, zval *retval)
+{
+	DOM_PROP_NODE(xmlNodePtr, nodep, obj);
+
+	xmlChar *content = xmlNodeGetContent(nodep);
+
+	if (UNEXPECTED(content == NULL)) {
+		php_dom_throw_error(INVALID_STATE_ERR, true);
+		return FAILURE;
+	} else {
+		ZVAL_STRING(retval, (const char *) content);
+		xmlFree(content);
+	}
+
+	return SUCCESS;
+}
+
+zend_result dom_modern_element_substituted_node_value_write(dom_object *obj, zval *newval)
+{
+	DOM_PROP_NODE(xmlNodePtr, nodep, obj);
+
+	php_libxml_invalidate_node_list_cache(obj->document);
+	dom_remove_all_children(nodep);
+	xmlNodeSetContentLen(nodep, (xmlChar *) Z_STRVAL_P(newval), Z_STRLEN_P(newval));
+
+	return SUCCESS;
+}
+
+static void dom_element_get_in_scope_namespace_info(php_dom_libxml_ns_mapper *ns_mapper, HashTable *result, xmlNodePtr nodep, dom_object *intern)
+{
+	HashTable prefix_to_ns_table;
+	zend_hash_init(&prefix_to_ns_table, 0, NULL, NULL, false);
+	zend_hash_real_init_mixed(&prefix_to_ns_table);
+
+	/* https://www.w3.org/TR/1999/REC-xpath-19991116/#namespace-nodes */
+	for (const xmlNode *cur = nodep; cur != NULL; cur = cur->parent) {
+		if (cur->type == XML_ELEMENT_NODE) {
+			/* Find the last attribute */
+			const xmlAttr *last = NULL;
+			for (const xmlAttr *attr = cur->properties; attr != NULL; attr = attr->next) {
+				last = attr;
+			}
+
+			/* Reversed loop because the parent traversal is reversed as well,
+			 * this will keep the ordering consistent. */
+			for (const xmlAttr *attr = last; attr != NULL; attr = attr->prev) {
+				if (attr->ns != NULL && php_dom_ns_is_fast_ex(attr->ns, php_dom_ns_is_xmlns_magic_token)
+					&& attr->children != NULL && attr->children->content != NULL) {
+					const char *prefix = attr->ns->prefix == NULL ? NULL : (const char *) attr->name;
+					const char *key = prefix == NULL ? "" : prefix;
+					xmlNsPtr ns = php_dom_libxml_ns_mapper_get_ns_raw_strings_nullsafe(ns_mapper, prefix, (const char *) attr->children->content);
+					/* NULL is a valid value for the sentinel */
+					zval zv;
+					ZVAL_PTR(&zv, ns);
+					zend_hash_str_add(&prefix_to_ns_table, key, strlen(key), &zv);
+				}
+			}
+		}
+	}
+
+	xmlNsPtr ns;
+	zend_string *prefix;
+	ZEND_HASH_MAP_REVERSE_FOREACH_STR_KEY_PTR(&prefix_to_ns_table, prefix, ns) {
+		if (ZSTR_LEN(prefix) == 0 && (ns == NULL || ns->href == NULL || *ns->href == '\0')) {
+			/* Exception: "the value of the xmlns attribute for the nearest such element is non-empty" */
+			continue;
+		}
+
+		zval zv;
+		object_init_ex(&zv, dom_namespace_info_class_entry);
+		zend_object *obj = Z_OBJ(zv);
+
+		if (ZSTR_LEN(prefix) != 0) {
+			ZVAL_STR_COPY(OBJ_PROP_NUM(obj, 0), prefix);
+		} else {
+			ZVAL_NULL(OBJ_PROP_NUM(obj, 0));
+		}
+
+		if (ns != NULL && ns->href != NULL && *ns->href != '\0') {
+			ZVAL_STRING(OBJ_PROP_NUM(obj, 1), (const char *) ns->href);
+		} else {
+			ZVAL_NULL(OBJ_PROP_NUM(obj, 1));
+		}
+
+		php_dom_create_object(nodep, OBJ_PROP_NUM(obj, 2), intern);
+
+		zend_hash_next_index_insert_new(result, &zv);
+	} ZEND_HASH_FOREACH_END();
+
+	zend_hash_destroy(&prefix_to_ns_table);
+}
+
+PHP_METHOD(Dom_Element, getInScopeNamespaces)
+{
+	zval *id;
+	xmlNode *nodep;
+	dom_object *intern;
+
+	ZEND_PARSE_PARAMETERS_NONE();
+
+	DOM_GET_THIS_OBJ(nodep, id, xmlNodePtr, intern);
+
+	php_dom_libxml_ns_mapper *ns_mapper = php_dom_get_ns_mapper(intern);
+
+	array_init(return_value);
+	HashTable *result = Z_ARRVAL_P(return_value);
+
+	dom_element_get_in_scope_namespace_info(ns_mapper, result, nodep, intern);
+}
+
+PHP_METHOD(Dom_Element, getDescendantNamespaces)
+{
+	zval *id;
+	xmlNode *nodep;
+	dom_object *intern;
+
+	ZEND_PARSE_PARAMETERS_NONE();
+
+	DOM_GET_THIS_OBJ(nodep, id, xmlNodePtr, intern);
+
+	php_dom_libxml_ns_mapper *ns_mapper = php_dom_get_ns_mapper(intern);
+
+	array_init(return_value);
+	HashTable *result = Z_ARRVAL_P(return_value);
+
+	dom_element_get_in_scope_namespace_info(ns_mapper, result, nodep, intern);
+
+	xmlNodePtr cur = nodep->children;
+	while (cur != NULL) {
+		if (cur->type == XML_ELEMENT_NODE) {
+			/* TODO: this could be more optimized by updating the same HashTable repeatedly
+			 * instead of recreating it on every node. */
+			dom_element_get_in_scope_namespace_info(ns_mapper, result, cur, intern);
+		}
+
+		cur = php_dom_next_in_tree_order(cur, nodep);
+	}
+}
+
+PHP_METHOD(Dom_Element, rename)
+{
+	zend_string *namespace_uri, *qualified_name;
+	ZEND_PARSE_PARAMETERS_START(2, 2)
+		Z_PARAM_STR_OR_NULL(namespace_uri)
+		Z_PARAM_STR(qualified_name)
+	ZEND_PARSE_PARAMETERS_END();
+
+	zval *id;
+	dom_object *intern;
+	xmlNodePtr nodep;
+	DOM_GET_THIS_OBJ(nodep, id, xmlNodePtr, intern);
+
+	xmlChar *localname = NULL, *prefix = NULL;
+	int errorcode = dom_validate_and_extract(namespace_uri, qualified_name, &localname, &prefix);
+	if (UNEXPECTED(errorcode != 0)) {
+		php_dom_throw_error(errorcode, /* strict */ true);
+		goto cleanup;
+	}
+
+	if (nodep->type == XML_ATTRIBUTE_NODE) {
+		/* Check for duplicate attributes. */
+		xmlAttrPtr existing = xmlHasNsProp(nodep->parent, localname, namespace_uri && ZSTR_VAL(namespace_uri)[0] != '\0' ? BAD_CAST ZSTR_VAL(namespace_uri) : NULL);
+		if (existing != NULL && existing != (xmlAttrPtr) nodep) {
+			php_dom_throw_error_with_message(INVALID_MODIFICATION_ERR, "An attribute with the given name in the given namespace already exists", /* strict */ true);
+			goto cleanup;
+		}
+	} else {
+		ZEND_ASSERT(nodep->type == XML_ELEMENT_NODE);
+
+		/* Check for moving to or away from the HTML namespace. */
+		bool is_currently_html_ns = php_dom_ns_is_fast(nodep, php_dom_ns_is_html_magic_token);
+		bool will_be_html_ns = namespace_uri != NULL && zend_string_equals_literal(namespace_uri, DOM_XHTML_NS_URI);
+		if (is_currently_html_ns != will_be_html_ns) {
+			if (is_currently_html_ns) {
+				php_dom_throw_error_with_message(
+					INVALID_MODIFICATION_ERR,
+					"It is not possible to move an element out of the HTML namespace because the HTML namespace is tied to the HTMLElement class",
+					/* strict */ true
+				);
+			} else {
+				php_dom_throw_error_with_message(
+					INVALID_MODIFICATION_ERR,
+					"It is not possible to move an element into the HTML namespace because the HTML namespace is tied to the HTMLElement class",
+					/* strict */ true
+				);
+			}
+			goto cleanup;
+		}
+	}
+
+	php_libxml_invalidate_node_list_cache(intern->document);
+
+	php_dom_libxml_ns_mapper *ns_mapper = php_dom_get_ns_mapper(intern);
+
+	/* Update namespace uri + prefix by querying the namespace mapper */
+	/* prefix can be NULL here, but that is taken care of by the called APIs. */
+	nodep->ns = php_dom_libxml_ns_mapper_get_ns_raw_prefix_string(ns_mapper, prefix, xmlStrlen(prefix), namespace_uri);
+
+	/* Change the local name */
+	if (xmlDictOwns(nodep->doc->dict, nodep->name) != 1) {
+		xmlFree((xmlChar *) nodep->name);
+	}
+	const xmlChar *copy = xmlDictLookup(nodep->doc->dict, localname, -1);
+	if (copy != NULL) {
+		nodep->name = copy;
+	} else {
+		nodep->name = localname;
+		localname = NULL;
+	}
+
+cleanup:
+	xmlFree(localname);
+	xmlFree(prefix);
+}
+
 #endif

ext/dom/php_dom.c

@@ -89,6 +89,7 @@ PHP_DOM_EXPORT zend_class_entry *dom_modern_xpath_class_entry;
 #endif
 PHP_DOM_EXPORT zend_class_entry *dom_namespace_node_class_entry;
 PHP_DOM_EXPORT zend_class_entry *dom_adjacent_position_class_entry;
+PHP_DOM_EXPORT zend_class_entry *dom_namespace_info_class_entry;
 /* }}} */
 
 static zend_object_handlers dom_object_handlers;
@@ -840,6 +841,8 @@ PHP_MINIT_FUNCTION(dom)
 	DOM_REGISTER_PROP_HANDLER(&dom_namespace_node_prop_handlers, "parentElement", dom_node_parent_element_read, NULL);
 	zend_hash_add_new_ptr(&classes, dom_namespace_node_class_entry->name, &dom_namespace_node_prop_handlers);
 
+	dom_namespace_info_class_entry = register_class_Dom_NamespaceInfo();
+
 	dom_documentfragment_class_entry = register_class_DOMDocumentFragment(dom_node_class_entry, dom_parentnode_class_entry);
 	dom_documentfragment_class_entry->create_object = dom_objects_new;
 	dom_documentfragment_class_entry->default_object_handlers = &dom_object_handlers;
@@ -1066,6 +1069,7 @@ PHP_MINIT_FUNCTION(dom)
 	DOM_REGISTER_PROP_HANDLER(&dom_modern_element_prop_handlers, "previousElementSibling", dom_node_previous_element_sibling_read, NULL);
 	DOM_REGISTER_PROP_HANDLER(&dom_modern_element_prop_handlers, "nextElementSibling", dom_node_next_element_sibling_read, NULL);
 	DOM_REGISTER_PROP_HANDLER(&dom_modern_element_prop_handlers, "innerHTML", dom_element_inner_html_read, dom_element_inner_html_write);
+	DOM_REGISTER_PROP_HANDLER(&dom_modern_element_prop_handlers, "substitutedNodeValue", dom_modern_element_substituted_node_value_read, dom_modern_element_substituted_node_value_write);
 	zend_hash_merge(&dom_modern_element_prop_handlers, &dom_modern_node_prop_handlers, NULL, false);
 	DOM_OVERWRITE_PROP_HANDLER(&dom_modern_element_prop_handlers, "textContent", dom_node_text_content_read, dom_node_text_content_write);
 	zend_hash_add_new_ptr(&classes, dom_modern_element_class_entry->name, &dom_modern_element_prop_handlers);

ext/dom/php_dom.stub.php

@@ -1384,6 +1384,16 @@ public function closest(string $selectors): ?Element {}
         public function matches(string $selectors): bool {}
 
         public string $innerHTML;
+
+        public string $substitutedNodeValue;
+
+        /** @return list<NamespaceInfo> */
+        public function getInScopeNamespaces(): array {}
+
+        /** @return list<NamespaceInfo> */
+        public function getDescendantNamespaces(): array {}
+
+        public function rename(?string $namespaceURI, string $qualifiedName): void {}
     }
 
     class HTMLElement extends Element
@@ -1410,6 +1420,9 @@ class Attr extends Node
 
         /** @implementation-alias DOMAttr::isId */
         public function isId(): bool {}
+
+        /** @implementation-alias Dom\Element::rename */
+        public function rename(?string $namespaceURI, string $qualifiedName): void {}
     }
 
     class CharacterData extends Node implements ChildNode
@@ -1688,6 +1701,20 @@ public function count(): int {}
         public function getIterator(): \Iterator {}
     }
 
+    /**
+     * @not-serializable
+     * @strict-properties
+     */
+    readonly final class NamespaceInfo
+    {
+        public ?string $prefix;
+        public ?string $namespaceURI;
+        public Element $element;
+
+        /** @implementation-alias Dom\Node::__construct */
+        private function __construct() {}
+    }
+
 #ifdef LIBXML_XPATH_ENABLED
     /** @not-serializable */
     final class XPath