comments

diff --git a/src/Encoder.php.bak b/src/Encoder.php.bak
new file mode 100644 (file)
index 0000000..ab4311c
--- /dev/null
+++ b/src/Encoder.php.bak
@@ -0,0 +1,339 @@
+<?php
+
+namespace PhpXmlRpc;
+
+use PhpXmlRpc\Helper\Logger;
+use PhpXmlRpc\Helper\XMLParser;
+
+/**
+ * A helper class to easily convert between Value objects and php native values
+ * @todo implement an interface
+ * @todo add class constants for the options values
+ */
+class Encoder
+{
+    /**
+     * Takes an xmlrpc value in object format and translates it into native PHP types.
+     *
+     * Works with xmlrpc requests objects as input, too.
+     *
+     * Given proper options parameter, can rebuild generic php object instances (provided those have been encoded to
+     * xmlrpc format using a corresponding option in php_xmlrpc_encode())
+     * PLEASE NOTE that rebuilding php objects involves calling their constructor function.
+     * This means that the remote communication end can decide which php code will get executed on your server, leaving
+     * the door possibly open to 'php-injection' style of attacks (provided you have some classes defined on your server
+     * that might wreak havoc if instances are built outside an appropriate context).
+     * Make sure you trust the remote server/client before enabling this!
+     *
+     * @author Dan Libby (dan@libby.com)
+     *
+     * @param Value|Request $xmlrpcVal
+     * @param array $options if 'decode_php_objs' is set in the options array, xmlrpc structs can be decoded into php
+     *                       objects; if 'dates_as_objects' is set xmlrpc datetimes are decoded as php DateTime objects
+     *
+     * @return mixed
+     */
+    public function decode($xmlrpcVal, $options = array())
+    {
+        switch ($xmlrpcVal->kindOf()) {
+            case 'scalar':
+                if (in_array('extension_api', $options)) {
+                    $val = reset($xmlrpcVal->me);
+                    $typ = key($xmlrpcVal->me);
+                    switch ($typ) {
+                        case 'dateTime.iso8601':
+                            $xmlrpcVal = array(
+                                'xmlrpc_type' => 'datetime',
+                                'scalar' => $val,
+                                'timestamp' => \PhpXmlRpc\Helper\Date::iso8601Decode($val)
+                            );
+                            return (object)$xmlrpcVal;
+                        case 'base64':
+                            $xmlrpcVal = array(
+                                'xmlrpc_type' => 'base64',
+                                'scalar' => $val
+                            );
+                            return (object)$xmlrpcVal;
+                        default:
+                            return $xmlrpcVal->scalarval();
+                    }
+                }
+                if (in_array('dates_as_objects', $options) && $xmlrpcVal->scalartyp() == 'dateTime.iso8601') {
+                    // we return a Datetime object instead of a string since now the constructor of xmlrpc value accepts
+                    // safely strings, ints and datetimes, we cater to all 3 cases here
+                    $out = $xmlrpcVal->scalarval();
+                    if (is_string($out)) {
+                        $out = strtotime($out);
+                    }
+                    if (is_int($out)) {
+                        $result = new \DateTime();
+                        $result->setTimestamp($out);
+
+                        return $result;
+                    } elseif (is_a($out, 'DateTimeInterface')) {
+                        return $out;
+                    }
+                }
+                return $xmlrpcVal->scalarval();
+
+            case 'array':
+                $arr = array();
+                foreach($xmlrpcVal as $value) {
+                    $arr[] = $this->decode($value, $options);
+                }
+                return $arr;
+
+            case 'struct':
+                // If user said so, try to rebuild php objects for specific struct vals.
+                /// @todo should we raise a warning for class not found?
+                // shall we check for proper subclass of xmlrpc value instead of presence of _php_class to detect
+                // what we can do?
+                if (in_array('decode_php_objs', $options) && $xmlrpcVal->_php_class != ''
+                    && class_exists($xmlrpcVal->_php_class)
+                ) {
+                    $obj = @new $xmlrpcVal->_php_class();
+                    foreach ($xmlrpcVal as $key => $value) {
+                        $obj->$key = $this->decode($value, $options);
+                    }
+                    return $obj;
+                } else {
+                    $arr = array();
+                    foreach ($xmlrpcVal as $key => $value) {
+                        $arr[$key] = $this->decode($value, $options);
+                    }
+                    return $arr;
+                }
+
+            case 'msg':
+                $paramCount = $xmlrpcVal->getNumParams();
+                $arr = array();
+                for ($i = 0; $i < $paramCount; $i++) {
+                    $arr[] = $this->decode($xmlrpcVal->getParam($i), $options);
+                }
+                return $arr;
+
+            /// @todo throw on unsupported type
+        }
+    }
+
+    /**
+     * Takes native php types and encodes them into xmlrpc PHP object format.
+     * It will not re-encode xmlrpc value objects.
+     *
+     * Feature creep -- could support more types via optional type argument
+     * (string => datetime support has been added, ??? => base64 not yet)
+     *
+     * If given a proper options parameter, php object instances will be encoded into 'special' xmlrpc values, that can
+     * later be decoded into php objects by calling php_xmlrpc_decode() with a corresponding option
+     *
+     * @author Dan Libby (dan@libby.com)
+     *
+     * @param mixed $phpVal the value to be converted into an xmlrpc value object
+     * @param array $options can include 'encode_php_objs', 'auto_dates', 'null_extension' or 'extension_api'
+     *
+     * @return Value
+     */
+    public function encode($phpVal, $options = array())
+    {
+        $type = gettype($phpVal);
+        switch ($type) {
+            case 'string':
+                /// @todo should we be stricter in the accepted dates (ie. reject more of invalid days & times)?
+                if (in_array('auto_dates', $options) && preg_match('/^[0-9]{8}T[0-9]{2}:[0-9]{2}:[0-9]{2}$/', $phpVal)) {
+                    $xmlrpcVal = new Value($phpVal, Value::$xmlrpcDateTime);
+                } else {
+                    $xmlrpcVal = new Value($phpVal, Value::$xmlrpcString);
+                }
+                break;
+            case 'integer':
+                $xmlrpcVal = new Value($phpVal, Value::$xmlrpcInt);
+                break;
+            case 'double':
+                $xmlrpcVal = new Value($phpVal, Value::$xmlrpcDouble);
+                break;
+            // Add support for encoding/decoding of booleans, since they are supported in PHP
+            case 'boolean':
+                $xmlrpcVal = new Value($phpVal, Value::$xmlrpcBoolean);
+                break;
+            case 'array':
+                // PHP arrays can be encoded to either xmlrpc structs or arrays, depending on whether they are hashes
+                // or plain 0..n integer indexed
+                // A shorter one-liner would be
+                // $tmp = array_diff(array_keys($phpVal), range(0, count($phpVal)-1));
+                // but execution time skyrockets!
+                $j = 0;
+                $arr = array();
+                $ko = false;
+                foreach ($phpVal as $key => $val) {
+                    $arr[$key] = $this->encode($val, $options);
+                    if (!$ko && $key !== $j) {
+                        $ko = true;
+                    }
+                    $j++;
+                }
+                if ($ko) {
+                    $xmlrpcVal = new Value($arr, Value::$xmlrpcStruct);
+                } else {
+                    $xmlrpcVal = new Value($arr, Value::$xmlrpcArray);
+                }
+                break;
+            case 'object':
+                if (is_a($phpVal, 'PhpXmlRpc\Value')) {
+                    $xmlrpcVal = $phpVal;
+                } elseif (is_a($phpVal, 'DateTimeInterface')) {
+                    $xmlrpcVal = new Value($phpVal->format('Ymd\TH:i:s'), Value::$xmlrpcDateTime);
+                } elseif (in_array('extension_api', $options) && $phpVal instanceof \stdClass && isset($phpVal->xmlrpc_type)) {
+                    // Handle the 'pre-converted' base64 and datetime values
+                    if (isset($phpVal->scalar)) {
+                        switch ($phpVal->xmlrpc_type) {
+                            case 'base64':
+                                $xmlrpcVal = new Value($phpVal->scalar, Value::$xmlrpcBase64);
+                                break;
+                            case 'datetime':
+                                $xmlrpcVal = new Value($phpVal->scalar, Value::$xmlrpcDateTime);
+                                break;
+                            default:
+                                $xmlrpcVal = new Value();
+                        }
+                    } else {
+                        $xmlrpcVal = new Value();
+                    }
+
+                } else {
+                    $arr = array();
+                    foreach($phpVal as $k => $v) {
+                        $arr[$k] = $this->encode($v, $options);
+                    }
+                    $xmlrpcVal = new Value($arr, Value::$xmlrpcStruct);
+                    if (in_array('encode_php_objs', $options)) {
+                        // let's save original class name into xmlrpc value:
+                        // might be useful later on...
+                        $xmlrpcVal->_php_class = get_class($phpVal);
+                    }
+                }
+                break;
+            case 'NULL':
+                if (in_array('extension_api', $options)) {
+                    $xmlrpcVal = new Value('', Value::$xmlrpcString);
+                } elseif (in_array('null_extension', $options)) {
+                    $xmlrpcVal = new Value('', Value::$xmlrpcNull);
+                } else {
+                    $xmlrpcVal = new Value();
+                }
+                break;
+            case 'resource':
+                if (in_array('extension_api', $options)) {
+                    $xmlrpcVal = new Value((int)$phpVal, Value::$xmlrpcInt);
+                } else {
+                    $xmlrpcVal = new Value();
+                }
+                break;
+            // catch "user function", "unknown type"
+            default:
+                // giancarlo pinerolo <ping@alt.it>
+                // it has to return an empty object in case, not a boolean.
+                $xmlrpcVal = new Value();
+                break;
+        }
+
+        return $xmlrpcVal;
+    }
+
+    /**
+     * Convert the xml representation of a method response, method request or single
+     * xmlrpc value into the appropriate object (a.k.a. deserialize).
+     *
+     * @todo is this a good name/class for this method? It does something quite different from 'decode' after all
+     *       (returning objects vs returns plain php values)... In fact it belongs rather to a Parser class
+     *
+     * @param string $xmlVal
+     * @param array $options
+     *
+     * @return Value|Request|Response|false false on error, or an instance of either Value, Request or Response
+     */
+    public function decodeXml($xmlVal, $options = array())
+    {
+        // 'guestimate' encoding
+        $valEncoding = XMLParser::guessEncoding('', $xmlVal);
+        if ($valEncoding != '') {
+
+            // Since parsing will fail if
+            // - charset is not specified in the xml prologue,
+            // - the encoding is not UTF8 and
+            // - there are non-ascii chars in the text,
+            // we try to work round that...
+            // The following code might be better for mb_string enabled installs, but makes the lib about 200% slower...
+            //if (!is_valid_charset($valEncoding, array('UTF-8'))
+            if (!in_array($valEncoding, array('UTF-8', 'US-ASCII')) && !XMLParser::hasEncoding($xmlVal)) {
+                if ($valEncoding == 'ISO-8859-1') {
+                    $xmlVal = utf8_encode($xmlVal);
+                } else {
+                    if (extension_loaded('mbstring')) {
+                        $xmlVal = mb_convert_encoding($xmlVal, 'UTF-8', $valEncoding);
+                    } else {
+                        Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': invalid charset encoding of xml text: ' . $valEncoding);
+                    }
+                }
+            }
+        }
+
+        // What if internal encoding is not in one of the 3 allowed? We use the broadest one, ie. utf8!
+        if (!in_array(PhpXmlRpc::$xmlrpc_internalencoding, array('UTF-8', 'ISO-8859-1', 'US-ASCII'))) {
+            /// @todo emit a warning
+            $parserOptions = array(XML_OPTION_TARGET_ENCODING => 'UTF-8');
+        } else {
+            $parserOptions = array(XML_OPTION_TARGET_ENCODING => PhpXmlRpc::$xmlrpc_internalencoding);
+        }
+
+        $xmlRpcParser = new XMLParser($parserOptions);
+        $xmlRpcParser->parse($xmlVal, XMLParser::RETURN_XMLRPCVALS, XMLParser::ACCEPT_REQUEST | XMLParser::ACCEPT_RESPONSE | XMLParser::ACCEPT_VALUE | XMLParser::ACCEPT_FAULT);
+
+        if ($xmlRpcParser->_xh['isf'] > 1) {
+            // test that $xmlrpc->_xh['value'] is an obj, too???
+
+            Logger::instance()->errorLog($xmlRpcParser->_xh['isf_reason']);
+
+            return false;
+        }
+
+        switch ($xmlRpcParser->_xh['rt']) {
+            case 'methodresponse':
+                $v = $xmlRpcParser->_xh['value'];
+                if ($xmlRpcParser->_xh['isf'] == 1) {
+                    /** @var Value $vc */
+                    $vc = $v['faultCode'];
+                    /** @var Value $vs */
+                    $vs = $v['faultString'];
+                    $r = new Response(0, $vc->scalarval(), $vs->scalarval());
+                } else {
+                    $r = new Response($v);
+                }
+                return $r;
+
+            case 'methodcall':
+                $req = new Request($xmlRpcParser->_xh['method']);
+                for ($i = 0; $i < count($xmlRpcParser->_xh['params']); $i++) {
+                    $req->addParam($xmlRpcParser->_xh['params'][$i]);
+                }
+                return $req;
+
+            case 'value':
+                return $xmlRpcParser->_xh['value'];
+
+            case 'fault':
+                // EPI api emulation
+                $v = $xmlRpcParser->_xh['value'];
+                // use a known error code
+                /** @var Value $vc */
+                $vc = isset($v['faultCode']) ? $v['faultCode']->scalarval() : PhpXmlRpc::$xmlrpcerr['invalid_return'];
+                /** @var Value $vs */
+                $vs = isset($v['faultString']) ? $v['faultString']->scalarval() : '';
+                if (!is_int($vc) || $vc == 0) {
+                    $vc = PhpXmlRpc::$xmlrpcerr['invalid_return'];
+                }
+                return new Response(0, $vc, $vs);
+            default:
+                return false;
+        }
+    }
+}

diff --git a/src/Helper/Charset.php b/src/Helper/Charset.php
index 8fdecb1..eca3e46 100644 (file)
--- a/src/Helper/Charset.php
+++ b/src/Helper/Charset.php
@@ -103,16 +103,21 @@ class Charset
 
     /**
      * Convert a string to the correct XML representation in a target charset.
+     * This involves:
+     * - character transformation for all characters which have a different representation in source and dest charsets
+     * - using 'charset entity' representation for all characters which are outside of the target charset
      *
      * To help correct communication of non-ascii chars inside strings, regardless of the charset used when sending
      * requests, parsing them, sending responses and parsing responses, an option is to convert all non-ascii chars
      * present in the message into their equivalent 'charset entity'. Charset entities enumerated this way are
      * independent of the charset encoding used to transmit them, and all XML parsers are bound to understand them.
-     * Note that in the std case we are not sending a charset encoding mime type along with http headers, so we are
-     * bound by RFC 3023 to emit strict us-ascii.
+     *
+     * Note that when not sending a charset encoding mime type along with http headers, we are bound by RFC 3023 to emit
+     * strict us-ascii for 'text/xml' payloads (but we should review RFC 7303, which seems to have changed the rules...)
      *
      * @todo do a bit of basic benchmarking (strtr vs. str_replace)
-     * @todo make usage of iconv() or recode_string() or mb_string() where available
+     * @todo make usage of iconv() or mb_string() where available
+     * @todo support aliases for charset names, eg ASCII, LATIN1, ISO-88591 (see f.e. polyfill-iconv for a list)
      *
      * @param string $data
      * @param string $srcEncoding

diff --git a/src/Helper/XMLParser.php b/src/Helper/XMLParser.php
index 6233225..59eec67 100644 (file)
--- a/src/Helper/XMLParser.php
+++ b/src/Helper/XMLParser.php
@@ -12,6 +12,8 @@ use PhpXmlRpc\Value;
  * @todo implement an interface to allow for alternative implementations
  *       - make access to $_xh protected, return more high-level data structures
  *       - add parseRequest, parseResponse, parseValue methods
+ * @todo if iconv() or mb_string() are available, we could allow to convert the received xml to a custom charset encoding
+ *       while parsing, which is faster than doing it later by going over the rebuilt data structure
  */
 class XMLParser
 {

diff --git a/src/Request.php b/src/Request.php
index 82f2b79..66cc788 100644 (file)
--- a/src/Request.php
+++ b/src/Request.php
@@ -283,6 +283,7 @@ class Request
         // This allows to send data which is native in various charset, by extending xmlrpc_encode_entities() and
         // setting xmlrpc_internalencoding
         if (!in_array(PhpXmlRpc::$xmlrpc_internalencoding, array('UTF-8', 'ISO-8859-1', 'US-ASCII'))) {
+            /// @todo emit a warning
             $options = array(XML_OPTION_TARGET_ENCODING => 'UTF-8');
         } else {
             $options = array(XML_OPTION_TARGET_ENCODING => PhpXmlRpc::$xmlrpc_internalencoding);

diff --git a/src/Server.php b/src/Server.php
index d5134ba..199a7a6 100644 (file)
--- a/src/Server.php
+++ b/src/Server.php
@@ -524,6 +524,7 @@ class Server
         // This allows to send data which is native in various charset,
         // by extending xmlrpc_encode_entities() and setting xmlrpc_internalencoding
         if (!in_array(PhpXmlRpc::$xmlrpc_internalencoding, array('UTF-8', 'ISO-8859-1', 'US-ASCII'))) {
+            /// @todo emit a warning
             $options = array(XML_OPTION_TARGET_ENCODING => 'UTF-8');
         } else {
             $options = array(XML_OPTION_TARGET_ENCODING => PhpXmlRpc::$xmlrpc_internalencoding);