changed semantic of [def] statements:

- [def] is used to define a variable (like before)
  but if we assign a string such as [def var "string"] the string is evaluated to bool
  in a different way: empty string is false, not empty string is true
- added [def?] statement - similar like [def] but define a variable only if such
  a variable is not already defined
- added [let] statement - similar like [def] but with lazy evaluation (it is an alias to a function)
  if it is called to assign a string e.g. [let var "string"] then this has the
  same meaning like [def] - an alias is only created when we assign a function e.g. [let var my_function]
  both [def] and [let] are using the same Vars object
- added [let?] - make an alias but only if such a variable is not already defined
- a std::map of variables moved outside of Generator
  added method void SetVariables(Vars & variables);
- fixed: a result status was not correctly propagated when evaluating [def] statements,
  this was in Call(...) function: last_res from variables was set in Find()
  but later was overwritten by Call(...) called for parameters (recursively)

src/Makefile

@@ -20,13 +20,13 @@ $(libname): $(o)
 
 
 %.o: %.cpp
-	$(CXX) -c $(CXXFLAGS) -I$(GLOBAL_WORKING_DIR)/pikotools $<
+	$(CXX) -c $(CXXFLAGS) -I$(GLOBAL_WORKING_DIR)/pikotools/src $<
 
 
 depend:
 	# !! IMPROVE ME
 	# as Ezc is a different project we rather shoudn't use '-I$(global_relative_working_dir)/pikotools' here?
-	makedepend -Y. -I$(global_relative_working_dir)/pikotools -f- *.cpp > Makefile.dep
+	makedepend -Y. -I$(global_relative_working_dir)/pikotools/src -f- *.cpp > Makefile.dep
 	echo -n "o = " > Makefile.o.dep
 	ls -1 *.cpp | xargs -I foo echo -n foo " " | sed -E "s/([^\.]*)\.cpp[ ]/\1\.o/g" >> Makefile.o.dep