Re: [Boost-users] [Serialization] Diamond shaped inheritance andserialization methods specialization

Subject: Re: [Boost-users] [Serialization] Diamond shaped inheritance andserialization methods specialization
From: Robert Bigaignon (bob.bigaignon_at_[hidden])
Date: 2010-08-02 09:04:51

On Thu, Jul 29, 2010 at 12:20 PM, Robert Ramey <ramey_at_[hidden]> wrote:
> Robert Bigaignon wrote:
>> Hi all,
>>
>> I'm having trouble with Boost's serialization library that I'm trying
>> to integrate with an existing class hierarchy that used MFC toolkit as
>> the archiver.
>>
>> It seems that when a serializable class defines a specialized save (or
>> load or serialize for that matter) in addition to the generic method,
>> the serialization library gets confused and ends up not identify
>> properly that the hierarchy of serializable classes uses multiple
>> inheritance and in particular diamond-shaped inheritance.
>>
>> I've reproduced the problem using the test_diamond code example
>> provided with Boost 1.43 using MS VC 9.0 by adding a save method
>> specialized for archive::text_oarchive to the derived1 class defined
>> in test_diamond.cpp. The following diff shows the implementation of
>> this specialized method:
>
> Without looking at this particular case, I would note that 1.44 - next
> release contains changes and a new test test_diamond_complex
> which improves support in this situation.
>
> Robert Ramey
>

Hi list, Robert,

A colleague of mine tried to compile my code with GCC, and it turned
out that the C++ I wrote is not legal. The explicit template
specialization was not in the correct namespace.

The standard says:
> An explicit specialization shall be declared in the namespace of which the template is a member, or, for member templates, in the namespace of which the enclosing class or enclosing class template is a member.
> An explicit specialization of a member function, member class or static data member of a class template shall be declared in the namespace of which the class template is a member

So it seems that my issue comes from the compiler, which did not warn
me about writting illegal C++ and, as a result, ended up with some
crippled executable.

--
Robert Bigaignon