Fix bug #70570: json_encode and var_dump ignore simplexml cdata

XML_CDATA_SECTION_NODE was simply not handled, handle it the same way as
we handle text nodes for consistency reasons.
We introduce the php_sxe_is_inclusive_text_node() helper for text-like
nodes. In DOM parlance, these 2 types are called inclusive text nodes.
Unfortunately, the fact that we handle CData the same as text now has a
technical BC break.

Previously this XML:
```
<?xml version="1.0" encoding="UTF-8"?>
<container>
    <C><![CDATA[hello]]><foo/><![CDATA[world]]></C>
</container>
```

resulted in this var_dump output:
```
object(SimpleXMLElement)#1 (1) {
  ["C"]=>
  object(SimpleXMLElement)#2 (1) {
    ["foo"]=>
    object(SimpleXMLElement)#3 (0) {
    }
  }
}
```

However, after this patch (as text is not an element and thus handled
specially) we get the following output:
```
object(SimpleXMLElement)#1 (1) {
  ["C"]=>
  string(10) "helloworld"
}
```

Author: nielsdos

ext/simplexml/simplexml.c

@@ -53,6 +53,11 @@ static void php_sxe_iterator_move_forward(zend_object_iterator *iter);
 static void php_sxe_iterator_rewind(zend_object_iterator *iter);
 static zend_result sxe_object_cast_ex(zend_object *readobj, zval *writeobj, int type);
 
+static zend_always_inline bool php_sxe_is_inclusive_text_node(const xmlNode *node)
+{
+	return node->type == XML_TEXT_NODE || node->type == XML_CDATA_SECTION_NODE;
+}
+
 /* {{{ _node_as_zval() */
 static void _node_as_zval(php_sxe_object *sxe, xmlNodePtr node, zval *value, SXE_ITER itertype, char *name, const xmlChar *nsprefix, int isprefix)
 {
@@ -746,7 +751,7 @@ static int sxe_prop_dim_exists(zend_object *object, zval *member, int check_empt
 			if (node) {
 				exists = 1;
 				if (check_empty == 1 &&
-					(!node->children || (node->children->type == XML_TEXT_NODE && !node->children->next &&
+					(!node->children || (php_sxe_is_inclusive_text_node(node->children) && !node->children->next &&
 					 (!node->children->content || !node->children->content[0] || xmlStrEqual(node->children->content, (const xmlChar *) "0")))) ) {
 					exists = 0;
 				}
@@ -928,7 +933,7 @@ static void _get_base_node_value(php_sxe_object *sxe_ref, xmlNodePtr node, zval
 	php_sxe_object *subnode;
 	xmlChar        *contents;
 
-	if (node->children && node->children->type == XML_TEXT_NODE && !xmlIsBlankNode(node->children)) {
+	if (node->children && php_sxe_is_inclusive_text_node(node->children) && !xmlIsBlankNode(node->children)) {
 		contents = xmlNodeListGetString(node->doc, node->children, 1);
 		if (contents) {
 			ZVAL_STRING(value, (char *)contents);
@@ -1026,7 +1031,7 @@ static int sxe_prop_is_empty(zend_object *object) /* {{{ */
 			if (node->children != NULL || node->prev != NULL || node->next != NULL) {
 				SKIP_TEXT(node);
 			} else {
-				if (node->type == XML_TEXT_NODE) {
+				if (php_sxe_is_inclusive_text_node(node)) {
 					const xmlChar *cur = node->content;
 					if (*cur != 0) {
 						is_empty = 0;
@@ -1147,7 +1152,7 @@ static HashTable *sxe_get_prop_hash(zend_object *object, int is_debug) /* {{{ */
 			if (node->children != NULL || node->prev != NULL || node->next != NULL || xmlIsBlankNode(node)) {
 				SKIP_TEXT(node);
 			} else {
-				if (node->type == XML_TEXT_NODE) {
+				if (php_sxe_is_inclusive_text_node(node)) {
 					const xmlChar *cur = node->content;
 
 					if (*cur != 0) {
@@ -1313,13 +1318,13 @@ PHP_METHOD(SimpleXMLElement, xpath)
 
 		for (i = 0; i < result->nodeNr; ++i) {
 			nodeptr = result->nodeTab[i];
-			if (nodeptr->type == XML_TEXT_NODE || nodeptr->type == XML_ELEMENT_NODE || nodeptr->type == XML_ATTRIBUTE_NODE) {
+			if (php_sxe_is_inclusive_text_node(nodeptr) || nodeptr->type == XML_ELEMENT_NODE || nodeptr->type == XML_ATTRIBUTE_NODE) {
 				/**
 				 * Detect the case where the last selector is text(), simplexml
 				 * always accesses the text() child by default, therefore we assign
 				 * to the parent node.
 				 */
-				if (nodeptr->type == XML_TEXT_NODE) {
+				if (php_sxe_is_inclusive_text_node(nodeptr)) {
 					_node_as_zval(sxe, nodeptr->parent, &value, SXE_ITER_NONE, NULL, NULL, 0);
 				} else if (nodeptr->type == XML_ATTRIBUTE_NODE) {
 					_node_as_zval(sxe, nodeptr->parent, &value, SXE_ITER_ATTRLIST, (char*)nodeptr->name, nodeptr->ns ? (xmlChar *)nodeptr->ns->href : NULL, 0);

ext/simplexml/tests/bug70570.phpt

@@ -0,0 +1,94 @@
+--TEST--
+Bug #70570 (json_encode and var_dump ignore simplexml cdata)
+--EXTENSIONS--
+simplexml
+--FILE--
+<?php
+
+$input = <<<XML
+<?xml version="1.0" encoding="UTF-8"?>
+<container>
+    <A><![CDATA[0]]></A>
+    <B><![CDATA[hello world]]></B>
+    <C><![CDATA[hello]]><foo/><![CDATA[world]]></C>
+    <D><foo/><![CDATA[hello]]><bar/></D>
+</container>
+XML;
+
+$xml = simplexml_load_string($input);
+
+var_dump($xml);
+
+foreach (["A", "B"] as $item) {
+    echo "--- Testing $item ---\n";
+    var_dump($xml->xpath("//$item/text()"));
+    var_dump($xml->$item);
+    var_dump(isset($xml->$item));
+    var_dump(empty($xml->$item));
+    var_dump((bool) ($xml->$item));
+}
+
+echo "--- Testing C ---\n";
+
+var_dump($xml->C); // Equivalent to how (normal) text nodes behave
+var_dump((string) $xml->C);
+var_dump($xml->C->foo);
+
+?>
+--EXPECT--
+object(SimpleXMLElement)#1 (4) {
+  ["A"]=>
+  string(1) "0"
+  ["B"]=>
+  string(11) "hello world"
+  ["C"]=>
+  string(10) "helloworld"
+  ["D"]=>
+  object(SimpleXMLElement)#2 (2) {
+    ["foo"]=>
+    object(SimpleXMLElement)#3 (0) {
+    }
+    ["bar"]=>
+    object(SimpleXMLElement)#4 (0) {
+    }
+  }
+}
+--- Testing A ---
+array(1) {
+  [0]=>
+  object(SimpleXMLElement)#2 (1) {
+    [0]=>
+    string(1) "0"
+  }
+}
+object(SimpleXMLElement)#2 (1) {
+  [0]=>
+  string(1) "0"
+}
+bool(true)
+bool(true)
+bool(true)
+--- Testing B ---
+array(1) {
+  [0]=>
+  object(SimpleXMLElement)#2 (1) {
+    [0]=>
+    string(11) "hello world"
+  }
+}
+object(SimpleXMLElement)#2 (1) {
+  [0]=>
+  string(11) "hello world"
+}
+bool(true)
+bool(false)
+bool(true)
+--- Testing C ---
+object(SimpleXMLElement)#2 (1) {
+  ["foo"]=>
+  object(SimpleXMLElement)#3 (0) {
+  }
+}
+string(10) "helloworld"
+object(SimpleXMLElement)#3 (0) {
+}