nico wrote:
Hello.
Context: ------- I use the boost serialization library to save and load objects of a system. I defined practices around this lib, and so I always serialize from a base class (every class serializable inherits from ISerializable).
I don't see that this practice is necessary. I'm not sure it's desirable either. The serialization library goes to great lengths to make this unnecessary. Having said that, I don't know that there's anything wrong with it. It's off topic - but just thought I'd mention it.
As a consequence, the true_type (i.e. the most derived type) is different from the this_type (i.e. ISerializable) and the true_type is stored in the archives.
correct.
My question: ----------- How to retreive this true_type (as the string written in the archive), only from the archive object?
I developp just a little bit: ---------------------------- Let's have this class tre:
ISerializable <|-- B <|-- D
If I do : B* b = new D(); b->SaveToFile(path); // <= this will do the serialization ar & this (this being a ISerializable*)
I obtain an archive where it is written the true_type "D" (whatever the type of the archive : txt, bin or xml).
With the b object and this code : const boost::serialization::extended_type_info & true_type = * boost::serialization::type_info_implementation<ISerializable>::type ::get_const_instance().get_derived_extended_type_info(*b);
I have what I want in true_type.get_key(), i.e : "D".
I can verify that "D" is written in every archive storing b.
My question again : how, only with an archive object (construct from the archive file without error), car I retrieve this key?
I think I understand the question - maybe. As the library is written, the inquiry into the the type of object is part of the loading process and off hand I don't see how one could "look ahead" without actually loading the object. You might be able to write some code which would "look ahead" but I have no idea what it would entail. why do you want to do this? What is the context? Robert Ramey
Tanks for any help.
Nicolas.
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost