March 2006 - Boost-users - lists.preview.boost.org

Re: [Boost-users] problem with serialization of pure virtual
by Nat Goodspeed 01 Mar '06

01 Mar '06

> -----Original Message----- > From: boost-users-bounces(a)lists.boost.org [mailto:boost-users- > bounces(a)lists.boost.org] On Behalf Of Tobias Combe > Sent: Tuesday, February 28, 2006 7:18 AM > To: boost-users(a)lists.boost.org > Subject: [Boost-users] problem with serialization of pure virtual > > When i use "ar.register<derived>();" in the template of the base-class > to register the derived classes, i get a syntax error. [Nat] What compiler are you using?

2 1

serialization: found a (satisfactory) way to avoid the abstract class problem linking problem
by RIVASSEAU Jean Noel 01 Mar '06

01 Mar '06

Hello Robert, I found a way to avoid the link problem when dealing with a base abstract class, that maintains the base class abstract : I just manually instantiated the save/load functions for the abstract base class. They should have been instantiated by BOOST_CLASS_EXPORT, but this apparently does not work with abstract base classes; however manual instantiation works just fine. So I now have only one more problem (unfortunately, very serious), that happens with the serialization of std::vector containing pointers. Jean-Noël

1 0

What can I do after I finished all the steps of "getting started" ?
by 郭华长 01 Mar '06

01 Mar '06

thanks _________________________________________________________________ 与联机的朋友进行交流，请使用 MSN Messenger: http://messenger.msn.com/cn

1 0

Re: [Boost-users] serialization: errata: even manual instantiationisimpossible
by RIVASSEAU Jean Noel 01 Mar '06

01 Mar '06

// my_class_serialization.cpp or my_class.cpp or .. *.cpp #include <boost/archive/xml_iarchive.hpp> .... // all archives you might whant to use here. template<class Archive> serialize(Archive &ar, my_class &t....){ ... serialization for class my_class } #include "my_class.hpp" // explict instantiation here for all the archives which interest me template void serialize<boost::archive::xml_iarchive, my_class>; ... // explicit instantiation for all other archives you might use here. I notice you still include the archive headers in my_class_serialization.cpp. In that case, you do not have in fact to manually instantiate templates, since my_class_serialization.cpp includes "my_class.hpp" which includes export.hpp and maybe BOOST_CLASS_EXPORT after the archive headers included in class_serialization.cpp, so template instantiation will occure anyway. Anyway let's close this topic. I understood what the point was. Jean-Noël

1 0

Re: [Boost-users] Serialization : unregistered_cast & otherquestions
by RIVASSEAU Jean Noel 01 Mar '06

01 Mar '06

This is true. But this is a minor in my opinion. The root cause of this situation is that linkers complain about multiply defined functions - even when those functions are identical. Maybe someday we'll have smarter linkers - but until then this works quite well. It does require that one spend a little time explicitly considering code instantiation. But this isn't necessarily bad. Letting the compile/linker do all the work will lead to code bloat and compilation time problems. I definitely agree with you - smarter linkers would be a big gift. Right now dealing with templates is generally a pain because of linker issues, and requires that one takes a lot of precaution to insure everything compiles/links smoothly. I had the same problems when using the BGL last year, and also had to resort to manual template instantiation in the end. Jean-Noël

1 0