Performance and bugfix for WebSockets

12 files changed, 161 insertions, 68 deletions

diff --git a/src/bindings/swig/php/uscxmlNativePHP.php b/src/bindings/swig/php/uscxmlNativePHP.php
index d7bb3d0..3aba5be 100644
--- a/src/bindings/swig/php/uscxmlNativePHP.php
+++ b/src/bindings/swig/php/uscxmlNativePHP.php
@@ -2,12 +2,12 @@
 
 /* ----------------------------------------------------------------------------
  * This file was automatically generated by SWIG (http://www.swig.org).
- * Version 2.0.9
- * 
- * This file is not intended to be easily readable and contains a number of 
+ * Version 2.0.11
+ *
+ * This file is not intended to be easily readable and contains a number of
  * coding conventions designed to improve portability and efficiency. Do not make
- * changes to this file unless you know what you are doing--modify the SWIG 
- * interface file instead. 
+ * changes to this file unless you know what you are doing--modify the SWIG
+ * interface file instead.
  * ----------------------------------------------------------------------------- */
 
 // Try to load our extension if it's not already loaded.
@@ -1030,13 +1030,7 @@ class Interpreter {
 	}
 
 	static function tokenizeIdRefs($idRefs) {
-		$r=Interpreter_tokenizeIdRefs($idRefs);
-		if (is_resource($r)) {
-			$c=substr(get_resource_type($r), (strpos(get_resource_type($r), '__') ? strpos(get_resource_type($r), '__') + 2 : 3));
-			if (class_exists($c)) return new $c($r);
-			return new StringVector($r);
-		}
-		return $r;
+		return Interpreter_tokenizeIdRefs($idRefs);
 	}
 
 	static function spaceNormalize($text) {
diff --git a/src/uscxml/Convenience.h b/src/uscxml/Convenience.h
index 9f4c1a1..0177476 100644
--- a/src/uscxml/Convenience.h
+++ b/src/uscxml/Convenience.h
@@ -49,7 +49,7 @@ inline bool isNumeric( const char* pszInput, int nNumberBase) {
 }
 
 inline bool iequals(const std::string& a, const std::string& b) {
-	// this impementation beats boost::iequals 2700ms vs 2100ms for test-performance.scxml
+	// this impementation beats boost::iequals 2700ms vs 2100ms for test-performance.scxml - we don't care for non-ascii yet
 	unsigned int size = a.size();
 	if (b.size() != size)
 		return false;
@@ -59,6 +59,16 @@ inline bool iequals(const std::string& a, const std::string& b) {
 	return true;
 }
 
+inline bool equals(const std::string& a, const std::string& b) {
+	unsigned int size = a.size();
+	if (b.size() != size)
+		return false;
+	for (unsigned int i = 0; i < size; ++i)
+		if (a[i] != b[i])
+			return false;
+	return true;
+}
+
 // see http://www.cplusplus.com/forum/general/27544/
 
 // Little-endian operating systems:
diff --git a/src/uscxml/Interpreter.cpp b/src/uscxml/Interpreter.cpp
index 1314986..dfd9311 100644
--- a/src/uscxml/Interpreter.cpp
+++ b/src/uscxml/Interpreter.cpp
@@ -912,11 +912,11 @@ void InterpreterImpl::send(const Arabica::DOM::Node<std::string>& element) {
 		// namelist
 		if (HAS_ATTR(element, "namelist")) {
 			if (_dataModel) {
-				std::vector<std::string> names = tokenizeIdRefs(ATTR(element, "namelist"));
-				for (int i = 0; i < names.size(); i++) {
-					Data namelistValue = _dataModel.getStringAsData(names[i]);
-					sendReq.namelist[names[i]] = namelistValue;
-					sendReq.data.compound[names[i]] = namelistValue;
+				std::list<std::string> names = tokenizeIdRefs(ATTR(element, "namelist"));
+				for (std::list<std::string>::const_iterator nameIter = names.begin(); nameIter != names.end(); nameIter++) {
+					Data namelistValue = _dataModel.getStringAsData(*nameIter);
+					sendReq.namelist[*nameIter] = namelistValue;
+					sendReq.data.compound[*nameIter] = namelistValue;
 				}
 			} else {
 				LOG(ERROR) << "Namelist attribute at send requires datamodel to be defined";
@@ -1054,9 +1054,9 @@ void InterpreterImpl::invoke(const Arabica::DOM::Node<std::string>& element) {
 
 		// namelist
 		if (HAS_ATTR(element, "namelist")) {
-			std::vector<std::string> names = tokenizeIdRefs(ATTR(element, "namelist"));
-			for (int i = 0; i < names.size(); i++) {
-				invokeReq.namelist[names[i]] = _dataModel.evalAsString(names[i]);
+			std::list<std::string> names = tokenizeIdRefs(ATTR(element, "namelist"));
+			for (std::list<std::string>::const_iterator nameIter = names.begin(); nameIter != names.end(); nameIter++) {
+				invokeReq.namelist[*nameIter] = _dataModel.evalAsString(*nameIter);
 			}
 		}
 
@@ -1235,10 +1235,11 @@ bool InterpreterImpl::hasConditionMatch(const Arabica::DOM::Node<std::string>& c
 
 void InterpreterImpl::executeContent(const NodeList<std::string>& content, bool rethrow) {
 	for (unsigned int i = 0; i < content.getLength(); i++) {
-		if (content.item(i).getNodeType() != Node_base::ELEMENT_NODE)
+		const Arabica::DOM::Node<std::string>& node = content.item(i);
+		if (node.getNodeType() != Node_base::ELEMENT_NODE)
 			continue;
 		try {
-			executeContent(content.item(i), true);
+			executeContent(node, true);
 		}
 		CATCH_AND_DISTRIBUTE2("Error when executing content", content.item(i));
 	}
@@ -1246,10 +1247,11 @@ void InterpreterImpl::executeContent(const NodeList<std::string>& content, bool
 
 void InterpreterImpl::executeContent(const NodeSet<std::string>& content, bool rethrow) {
 	for (unsigned int i = 0; i < content.size(); i++) {
-		if (content[i].getNodeType() != Node_base::ELEMENT_NODE)
+		const Arabica::DOM::Node<std::string>& node = content[i];
+		if (node.getNodeType() != Node_base::ELEMENT_NODE)
 			continue;
 		try {
-			executeContent(content[i], true);
+			executeContent(node, true);
 		}
 		CATCH_AND_DISTRIBUTE2("Error when executing content", content[i]);
 	}
@@ -1598,9 +1600,9 @@ NEXT_ANCESTOR:
 	return ancestor;
 }
 
-Arabica::XPath::NodeSet<std::string> InterpreterImpl::getStates(const std::vector<std::string>& stateIds) {
+Arabica::XPath::NodeSet<std::string> InterpreterImpl::getStates(const std::list<std::string>& stateIds) {
 	Arabica::XPath::NodeSet<std::string> states;
-	std::vector<std::string>::const_iterator tokenIter = stateIds.begin();
+	std::list<std::string>::const_iterator tokenIter = stateIds.begin();
 	while(tokenIter != stateIds.end()) {
 		states.push_back(getState(*tokenIter));
 		tokenIter++;
@@ -1717,18 +1719,32 @@ NodeSet<std::string> InterpreterImpl::getTargetStates(const Arabica::DOM::Node<s
 	}
 
 	std::string targetId = ((Arabica::DOM::Element<std::string>)transition).getAttribute("target");
-
-	std::vector<std::string> targetIds = InterpreterImpl::tokenizeIdRefs(ATTR(transition, "target"));
-	for (int i = 0; i < targetIds.size(); i++) {
-		Arabica::DOM::Node<std::string> state = getState(targetIds[i]);
+	std::list<std::string> targetIds = InterpreterImpl::tokenizeIdRefs(ATTR(transition, "target"));
+	for (std::list<std::string>::const_iterator targetIter = targetIds.begin(); targetIter != targetIds.end(); targetIter++) {
+		Arabica::DOM::Node<std::string> state = getState(*targetIter);
 		assert(HAS_ATTR(state, "id"));
 		targetStates.push_back(state);
 	}
 	return targetStates;
 }
 
-std::vector<std::string> InterpreterImpl::tokenizeIdRefs(const std::string& idRefs) {
-	std::vector<std::string> ids;
+std::list<std::string> InterpreterImpl::tokenizeIdRefs(const std::string& idRefs) {
+	std::list<std::string> ids;
+
+	// appr. 3x faster than stringstream
+	size_t start = 0;
+	for (int i = 0; i < idRefs.size(); i++) {
+		if (isspace(idRefs[i])) {
+			if (i > 0 && start < i - 1) {
+				ids.push_back(idRefs.substr(start, i - start));
+			}
+			while(isspace(idRefs[++i])); // skip whitespaces
+			start = i;
+		} else if (i + 1 == idRefs.size()) {
+			ids.push_back(idRefs.substr(start, i + 1 - start));
+		}
+	}
+
 
 #if 0
 	if (idRefs.length() > 0) {
@@ -1740,11 +1756,13 @@ std::vector<std::string> InterpreterImpl::tokenizeIdRefs(const std::string& idRe
 	}
 #endif
 
+#if 0
 	// this version is somewhat fatser than the one above
 	std::stringstream ss (idRefs);
 	std::string item;
 	while(ss >> item)
 		ids.push_back(item);
+#endif
 
 	return ids;
 }
diff --git a/src/uscxml/Interpreter.h b/src/uscxml/Interpreter.h
index 5cc9053..d2e63e9 100644
--- a/src/uscxml/Interpreter.h
+++ b/src/uscxml/Interpreter.h
@@ -236,7 +236,7 @@ public:
 	}
 
 	Arabica::DOM::Node<std::string> getState(const std::string& stateId);
-	Arabica::XPath::NodeSet<std::string> getStates(const std::vector<std::string>& stateIds);
+	Arabica::XPath::NodeSet<std::string> getStates(const std::list<std::string>& stateIds);
 
 	Arabica::DOM::Document<std::string>& getDocument()       {
 		return _document;
@@ -286,7 +286,7 @@ public:
 	static bool isCompound(const Arabica::DOM::Node<std::string>& state);
 	static bool isDescendant(const Arabica::DOM::Node<std::string>& s1, const Arabica::DOM::Node<std::string>& s2);
 
-	static std::vector<std::string> tokenizeIdRefs(const std::string& idRefs);
+	static std::list<std::string> tokenizeIdRefs(const std::string& idRefs);
 	static std::string spaceNormalize(const std::string& text);
 
 	bool isInEmbeddedDocument(const Arabica::DOM::Node<std::string>& node);
@@ -540,7 +540,7 @@ public:
 	Arabica::DOM::Node<std::string> getState(const std::string& stateId) {
 		return _impl->getState(stateId);
 	}
-	Arabica::XPath::NodeSet<std::string> getStates(const std::vector<std::string>& stateIds) {
+	Arabica::XPath::NodeSet<std::string> getStates(const std::list<std::string>& stateIds) {
 		return _impl->getStates(stateIds);
 	}
 
@@ -628,7 +628,7 @@ public:
 		return InterpreterImpl::isDescendant(s1, s2);
 	}
 
-	static std::vector<std::string> tokenizeIdRefs(const std::string& idRefs) {
+	static std::list<std::string> tokenizeIdRefs(const std::string& idRefs) {
 		return InterpreterImpl::tokenizeIdRefs(idRefs);
 	}
 	static std::string spaceNormalize(const std::string& text) {
diff --git a/src/uscxml/Message.cpp b/src/uscxml/Message.cpp
index c79e08a..26cf640 100644
--- a/src/uscxml/Message.cpp
+++ b/src/uscxml/Message.cpp
@@ -61,6 +61,64 @@ Blob::Blob(void* _data, size_t _size, const std::string& _mimeType, bool adopt)
 	size = _size;
 }
 
+#if 0
+// there used to work base64 encoded images in a browser - can't check extensively just now
+static const std::string base64_chars =
+    "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
+    "abcdefghijklmnopqrstuvwxyz"
+    "0123456789+/";
+
+std::string Blob::base64() {
+
+	int in_len = size;
+	char const* bytes_to_encode = data;
+
+	std::string ret;
+	int i = 0;
+	int j = 0;
+	unsigned char char_array_3[3];
+	unsigned char char_array_4[4];
+
+	while (in_len--) {
+		char_array_3[i++] = *(bytes_to_encode++);
+		if (i == 3) {
+			char_array_4[0] = (char_array_3[0] & 0xfc) >> 2;
+			char_array_4[1] = ((char_array_3[0] & 0x03) << 4) + ((char_array_3[1] & 0xf0) >> 4);
+			char_array_4[2] = ((char_array_3[1] & 0x0f) << 2) + ((char_array_3[2] & 0xc0) >> 6);
+			char_array_4[3] = char_array_3[2] & 0x3f;
+
+			for(i = 0; (i <4) ; i++)
+				ret += base64_chars[char_array_4[i]];
+			i = 0;
+		}
+	}
+
+	if (i) {
+		for(j = i; j < 3; j++)
+			char_array_3[j] = '\0';
+
+		char_array_4[0] = (char_array_3[0] & 0xfc) >> 2;
+		char_array_4[1] = ((char_array_3[0] & 0x03) << 4) + ((char_array_3[1] & 0xf0) >> 4);
+		char_array_4[2] = ((char_array_3[1] & 0x0f) << 2) + ((char_array_3[2] & 0xc0) >> 6);
+		char_array_4[3] = char_array_3[2] & 0x3f;
+
+		for (j = 0; (j < i + 1); j++)
+			ret += base64_chars[char_array_4[j]];
+
+		while((i++ < 3))
+			ret += '=';
+
+	}
+
+	return ret;
+
+}
+#else
+std::string Blob::base64() {
+	return base64Encode((char* const)data, size);
+}
+#endif
+
 Data::Data(const char* _data, size_t _size, const std::string& mimeType, bool adopt) {
 	binary = boost::shared_ptr<Blob>(new Blob((void*)_data, _size, mimeType, adopt));
 }
diff --git a/src/uscxml/Message.h b/src/uscxml/Message.h
index 5de9a56..5f6703e 100644
--- a/src/uscxml/Message.h
+++ b/src/uscxml/Message.h
@@ -55,9 +55,7 @@ public:
 		return uscxml::md5(data, size);
 	}
 
-	std::string base64() {
-		return base64Encode((char* const)data, size);
-	}
+	std::string base64();
 
 	Blob* fromBase64(const std::string base64) {
 		std::string decoded = base64Decode(base64);
diff --git a/src/uscxml/interpreter/InterpreterDraft6.cpp b/src/uscxml/interpreter/InterpreterDraft6.cpp
index c9c00ef..772ad96 100644
--- a/src/uscxml/interpreter/InterpreterDraft6.cpp
+++ b/src/uscxml/interpreter/InterpreterDraft6.cpp
@@ -467,9 +467,8 @@ bool InterpreterDraft6::isEnabledTransition(const Node<std::string>& transition,
 		return false;
 	}
 
-	std::vector<std::string> eventNames = tokenizeIdRefs(eventName);
-
-	std::vector<std::string>::iterator eventIter = eventNames.begin();
+	std::list<std::string> eventNames = tokenizeIdRefs(eventName);
+	std::list<std::string>::iterator eventIter = eventNames.begin();
 	while(eventIter != eventNames.end()) {
 		if(nameMatch(*eventIter, event) && hasConditionMatch(transition)) {
 			return true;
diff --git a/src/uscxml/plugins/datamodel/ecmascript/JavaScriptCore/JSCDataModel.cpp b/src/uscxml/plugins/datamodel/ecmascript/JavaScriptCore/JSCDataModel.cpp
index 947c0c2..9827b91 100644
--- a/src/uscxml/plugins/datamodel/ecmascript/JavaScriptCore/JSCDataModel.cpp
+++ b/src/uscxml/plugins/datamodel/ecmascript/JavaScriptCore/JSCDataModel.cpp
@@ -400,9 +400,10 @@ Data JSCDataModel::getValueAsData(const JSValueRef value) {
 			if (!isNumeric(property.c_str(), 10))
 				isArray = false;
 			propertySet.insert(property);
-			JSStringRelease(stringValue);
+			//JSStringRelease(stringValue); // JSPropertyNameArrayRelease does the job it seems
 			free(buf);
 		}
+		JSPropertyNameArrayRelease(properties);
 		std::set<std::string>::iterator propIter = propertySet.begin();
 		while(propIter != propertySet.end()) {
 			if (isArray) {
@@ -537,14 +538,14 @@ void JSCDataModel::assign(const Element<std::string>& assignElem,
 		throw Event("error.execution", Event::PLATFORM);
 
 	// flags on attribute are ignored?
-	if (key.compare("_sessionid") == 0)
-		return; //throw Event("error.execution", Event::PLATFORM);
+	if (key.compare("_sessionid") == 0) // test 322
+		throw Event("error.execution", Event::PLATFORM);
 	if (key.compare("_name") == 0)
-		return; //throw Event("error.execution", Event::PLATFORM);
-	if (key.compare("_ioprocessors") == 0)
-		return; //throw Event("error.execution", Event::PLATFORM);
+		throw Event("error.execution", Event::PLATFORM);
+	if (key.compare("_ioprocessors") == 0)  // test 326
+		throw Event("error.execution", Event::PLATFORM);
 	if (key.compare("_invokers") == 0)
-		return; //throw Event("error.execution", Event::PLATFORM);
+		throw Event("error.execution", Event::PLATFORM);
 
 	if (HAS_ATTR(assignElem, "expr")) {
 		evalAsValue(key + " = " + ATTR(assignElem, "expr"));
diff --git a/src/uscxml/server/HTTPServer.cpp b/src/uscxml/server/HTTPServer.cpp
index 358c819..6b24cd8 100644
--- a/src/uscxml/server/HTTPServer.cpp
+++ b/src/uscxml/server/HTTPServer.cpp
@@ -92,17 +92,24 @@ HTTPServer::HTTPServer(unsigned short port, unsigned short wsPort, SSLConfig* ss
 
 	evhttp_set_allowed_methods(_http, allowedMethods); // allow all methods
 
-	if (_port > 0)
+	if (_port > 0) {
 		_httpHandle = evhttp_bind_socket_with_handle(_http, INADDR_ANY, _port);
-	if (_httpHandle)
-		LOG(INFO) << "HTTP server listening on tcp/" << _port;
+		if (_httpHandle) {
+			LOG(INFO) << "HTTP server listening on tcp/" << _port;
+		} else {
+			LOG(ERROR) << "HTTP server cannot bind to tcp/" << _wsPort;
+		}
+	}
 
 	_wsPort = wsPort;
-	if (_wsPort > 0)
+	if (_wsPort > 0) {
 		_wsHandle = evws_bind_socket(_evws, _wsPort);
-	if (_wsHandle)
-		LOG(INFO) << "WebSocket server listening on tcp/" << _wsPort;
-
+		if (_wsHandle) {
+			LOG(INFO) << "WebSocket server listening on tcp/" << _wsPort;
+		} else {
+			LOG(ERROR) << "WebSocket server cannot bind to tcp/" << _wsPort;
+		}
+	}
 #if (defined EVENT_SSL_FOUND && defined OPENSSL_FOUND && defined OPENSSL_HAS_ELIPTIC_CURVES)
 	_sslHandle = NULL;
 	_https = NULL;
@@ -136,11 +143,15 @@ HTTPServer::HTTPServer(unsigned short port, unsigned short wsPort, SSLConfig* ss
 		evhttp_set_bevcb(_https, sslBufferEventCallback, ctx);
 		evhttp_set_gencb(_https, sslGeneralBufferEventCallback, NULL);
 
-		if (_sslPort > 0)
+		if (_sslPort > 0) {
 			_sslHandle = evhttp_bind_socket_with_handle(_https, INADDR_ANY, _sslPort);
-		if (_sslHandle)
-			LOG(INFO) << "HTTPS server listening on tcp/" << _wsPort;
+			if (_sslHandle) {
+				LOG(INFO) << "HTTPS server listening on tcp/" << _wsPort;
+			} else {
+				LOG(ERROR) << "HTTPS server cannot bind to tcp/" << _wsPort;
 
+			}
+		}
 	}
 #endif
 
diff --git a/src/uscxml/util/Base64.h b/src/uscxml/util/Base64.h
index 7dfc83c..a07d24e 100644
--- a/src/uscxml/util/Base64.h
+++ b/src/uscxml/util/Base64.h
@@ -36,9 +36,7 @@ extern "C" {
 	} base64_decodestate;
 
 	USCXML_API void base64_init_decodestate(base64_decodestate* state_in);
-
 	USCXML_API int base64_decode_value(char value_in);
-
 	USCXML_API int base64_decode_block(const char* code_in, const int length_in, char* plaintext_out, base64_decodestate* state_in);
 
 /// ENDCODE
@@ -54,11 +52,8 @@ extern "C" {
 	} base64_encodestate;
 
 	USCXML_API void base64_init_encodestate(base64_encodestate* state_in);
-
 	USCXML_API char base64_encode_value(char value_in);
-
 	USCXML_API int base64_encode_block(const char* plaintext_in, int length_in, char* code_out, base64_encodestate* state_in);
-
 	USCXML_API int base64_encode_blockend(char* code_out, base64_encodestate* state_in);
 
 #ifdef __cplusplus
diff --git a/src/uscxml/util/Base64.hpp b/src/uscxml/util/Base64.hpp
index a106a12..e3cbe03 100644
--- a/src/uscxml/util/Base64.hpp
+++ b/src/uscxml/util/Base64.hpp
@@ -12,7 +12,7 @@ extern "C" {
 
 namespace uscxml {
 
-USCXML_API inline std::string base64Encode(const char* data, unsigned int len) {
+USCXML_API inline std::string base64Encode(const char* data, unsigned int len, bool withBlockEnd = true) {
 	base64_encodestate* ctx = (base64_encodestate*)malloc(sizeof(base64_encodestate));
 	base64_init_encodestate(ctx);
 	
@@ -22,10 +22,15 @@ USCXML_API inline std::string base64Encode(const char* data, unsigned int len) {
 	 * be approximated with this formula:
 	 */
 	
+	int written = 0;
 	char* out = (char*)malloc(len * 1.4 + 814);
-	base64_encode_block(data, len, out, ctx);
+	written += base64_encode_block(data, len, out, ctx);
+	if (withBlockEnd) {
+		written += base64_encode_blockend(out + written, ctx);
+		written--;  // drop the newline
+	}
+	std::string result(out, written);
 	free(ctx);
-	std::string result(out);
 	free(out);
 	return result;
 }
@@ -42,5 +47,8 @@ USCXML_API inline std::string base64Decode(const std::string& data) {
 	return result;
 }
 
+//	USCXML_API std::string base64Decode(const std::string& data);
+//	USCXML_API std::string base64Encode(const char* data, unsigned int len);
+
 }
 #endif /* end of include guard: BASE64_H_5FKG12HF */
diff --git a/src/uscxml/util/MD5.c b/src/uscxml/util/MD5.c
index 1bf05e4..f553c9c 100644
--- a/src/uscxml/util/MD5.c
+++ b/src/uscxml/util/MD5.c
@@ -52,6 +52,7 @@
  */
 
 #include "MD5.h"
+#include <string.h>
 
 #undef BYTE_ORDER	/* 1 = big-endian, -1 = little-endian, 0 = unknown */
 #ifdef ARCH_IS_BIG_ENDIAN