support for uninitialized fields & empty sets and tables.

The only snag is... with the default output format of the log-file writer, the input reader cannot tell if a table or set is empty or uninitialized (both cases use the same character by default). In this case, by default it is assumed that the field/vector is uninitalized.

scripts/base/frameworks/input/__load__.bro

@@ -1 +1,3 @@
 @load ./main
+@load ./readers/ascii
+

src/InputMgr.cc

@@ -505,10 +505,10 @@ void InputMgr::SendEntry(const InputReader* reader, const LogVal* const *vals) {
 				position++;
 			}
 			
-			if ( val == 0 ) {
+			/* if ( val == 0 ) {
 				reporter->InternalError("conversion error");
 				return;
-			}
+			} */
 
 			r->Assign(j,val);
 
@@ -871,7 +871,9 @@ int InputMgr::GetLogValLength(const LogVal* val) {
 		}
 
 	case TYPE_VECTOR: {
-		for ( int i = 0; i < val->val.vector_val.size; i++ ) {
+		int j = val->val.vector_val.size;
+		for ( int i = 0; i < j; i++ ) {
+			reporter->Error("size is %d", val->val.vector_val.size);
 			length += GetLogValLength(val->val.vector_val.vals[i]);
 		}
 		break;
@@ -945,7 +947,8 @@ int InputMgr::CopyLogVal(char *data, const int startpos, const LogVal* val) {
 
 	case TYPE_VECTOR: {
 		int length = 0;
-		for ( int i = 0; i < val->val.vector_val.size; i++ ) {
+		int j = val->val.vector_val.size;
+		for ( int i = 0; i < j; i++ ) {
 			length += CopyLogVal(data, startpos+length, val->val.vector_val.vals[i]);
 		}
 		return length;
@@ -995,6 +998,10 @@ Val* InputMgr::LogValToVal(const LogVal* val, BroType* request_type) {
 		return 0;
 	}
 
+	if ( !val->present ) {
+		return 0; // unset field
+	}
+	
 
 	switch ( val->type ) {
 	case TYPE_BOOL:
@@ -1033,13 +1040,6 @@ Val* InputMgr::LogValToVal(const LogVal* val, BroType* request_type) {
 		break;
 
 	case TYPE_TABLE: {
-		if ( val->val.set_val.size == 0 ) {
-			// empty table
-			TypeList* set_index = new TypeList(base_type(TYPE_ANY));
-			// iim quite sure this does not work... we probably need the internal set type for this...
-			reporter->InternalError("Implement me.");
-			return new TableVal(new SetType(set_index, 0));
-		} else {
 		// all entries have to have the same type...
 		BroType* type = request_type->AsTableType()->Indices()->PureType();
 		TypeList* set_index = new TypeList(type->Ref());
@@ -1050,13 +1050,10 @@ Val* InputMgr::LogValToVal(const LogVal* val, BroType* request_type) {
 			t->Assign(LogValToVal( val->val.set_val.vals[i], type ), 0);
 		}
 		return t;
-		}	 
 		break;
 		}
 
 	case TYPE_VECTOR: {
-			assert ( val->val.vector_val.size > 1 ); // implement empty vector...
-
 		// all entries have to have the same type...
 		BroType* type = request_type->AsVectorType()->YieldType();
 		VectorType* vt = new VectorType(type->Ref());

src/InputReaderAscii.cc

@@ -34,31 +34,19 @@ InputReaderAscii::InputReaderAscii()
 
 	//keyMap = new map<string, string>();
 	
-	separator_len = BifConst::LogAscii::separator->Len();
+	separator.assign( (const char*) BifConst::InputAscii::separator->Bytes(), BifConst::InputAscii::separator->Len());
-	separator = new char[separator_len];
+	if ( separator.size() != 1 ) {
-	memcpy(separator, BifConst::LogAscii::separator->Bytes(),
-	       separator_len);
-	if ( separator_len != 1 ) {
 		Error("separator length has to be 1. Separator will be truncated.");
 	}
 
-	set_separator_len = BifConst::LogAscii::set_separator->Len();
+	set_separator.assign( (const char*) BifConst::InputAscii::set_separator->Bytes(), BifConst::InputAscii::set_separator->Len());
-	set_separator = new char[set_separator_len];
+	if ( set_separator.size() != 1 ) {
-	memcpy(set_separator, BifConst::LogAscii::set_separator->Bytes(),
-	       set_separator_len);
-	if ( set_separator_len != 1 ) {
 		Error("set_separator length has to be 1. Separator will be truncated.");
 	}
 
-	empty_field_len = BifConst::LogAscii::empty_field->Len();
+	empty_field.assign( (const char*) BifConst::InputAscii::empty_field->Bytes(), BifConst::InputAscii::empty_field->Len());
-	empty_field = new char[empty_field_len];
-	memcpy(empty_field, BifConst::LogAscii::empty_field->Bytes(),
-	       empty_field_len);
 	
-	unset_field_len = BifConst::LogAscii::unset_field->Len();
+	unset_field.assign( (const char*) BifConst::InputAscii::unset_field->Bytes(), BifConst::InputAscii::unset_field->Len());
-	unset_field = new char[unset_field_len];
-	memcpy(unset_field, BifConst::LogAscii::unset_field->Bytes(),
-	       unset_field_len);
 	
 }
 
@@ -66,10 +54,6 @@ InputReaderAscii::~InputReaderAscii()
 {
 	DoFinish();
 
-	delete [] separator;
-	delete [] set_separator;
-	delete [] empty_field;
-	delete [] unset_field;	
 }
 
 void InputReaderAscii::DoFinish()
@@ -172,7 +156,10 @@ bool InputReaderAscii::GetLine(string& str) {
 LogVal* InputReaderAscii::EntryToVal(string s, FieldMapping field) {
 
 	LogVal* val = new LogVal(field.type, true);
-	//bzero(val, sizeof(LogVal));
+
+	if ( s.compare(unset_field) == 0 ) { // field is not set...
+		return new LogVal(field.type, false);
+	}
 
 	switch ( field.type ) {
 	case TYPE_ENUM:
@@ -245,18 +232,12 @@ LogVal* InputReaderAscii::EntryToVal(string s, FieldMapping field) {
 
 		unsigned int pos = 0;
 		
+		if ( s.compare(empty_field) == 0 ) 
+			length = 0;
+
 		LogVal** lvals = new LogVal* [length];
 
 		if ( field.type == TYPE_TABLE ) {
-			// construct a table from entry...
-			// for the moment assume, that entries are split by ",".
-
-			/* Fix support for emtyp tables if ( s == "-" ) {
-				// empty 
-				val->val.set_val.size = 0;
-				break;
-			} */
-
 			val->val.set_val.vals = lvals;
 			val->val.set_val.size = length;
 		} else if ( field.type == TYPE_VECTOR ) {
@@ -266,18 +247,20 @@ LogVal* InputReaderAscii::EntryToVal(string s, FieldMapping field) {
 			assert(false);
 		}
 
+		if ( length == 0 )
+			break; //empty
+
 		istringstream splitstream(s);
 		while ( splitstream ) {
 			string element;
 
-			if ( pos >= length ) {
-				Error(Fmt("Internal error while parsing set. pos %d > length %d", pos, length));
-				break;
-			}
-
 			if ( !getline(splitstream, element, set_separator[0]) )
 				break;
 
+			if ( pos >= length ) {
+				Error(Fmt("Internal error while parsing set. pos %d >= length %d. Element: %s", pos, length, element.c_str()));
+				break;
+			}
 
 			LogVal* newval = EntryToVal(element, field.subType());
 			if ( newval == 0 ) {

src/InputReaderAscii.h

@@ -59,17 +59,13 @@ private:
 	//map<string, string> *keyMap;
 	//
 	// Options set from the script-level.
-	char* separator;
+	string separator;
-	int separator_len;
 
-	char* set_separator;
+	string set_separator;
-	int set_separator_len;
 
-	char* empty_field;
+	string empty_field;
-	int empty_field_len;
 
-	char* unset_field;
+	string unset_field;
-	int unset_field_len;
 
 };