Re: [Boost-users] 1.33 pointer to object serialization

From: Robert Ramey (ramey_at_[hidden])
Date: 2005-07-19 20:48:09

> gps_position* p_g = new gps_position(35, 59, 24.567f);
> oa << p_g;  // causes the error's below
>

Why didn't your just change the above to?:

const gps_position* p_g = new gps_position(35, 59, 24.567f);
oa << p_g;  // causes the error's below

Chris Coleman wrote:
> Ok,
>
> I've since found some information on Object Tracking in another boost
> list that I missed previously. (apologies)  This suggests the
> behaviour
> is now correct, and that 1.32 implementation was overly relaxed.
>
> As I understand it the << operator require a const reference to the
> object to be serialized inorder to prevent the same object being
> written
> or read, to or from an archive multiple times.  So you can now no
> longer
> serialize a pointer to an object, even if its a const ptr?
>
> My solution has been to dereference the object as below, and it
> compiles
> and seems to work.  However, I'm not convinced this is the correct way
> of doing things.
>
> oa << *memory->pull_slice(slice_id);
> ofs.close();
>
> where pull_slice returns a const * to the object to be serialized.
> Is this best?
>
>
> Also, it's not really my place to say, but as a user it would
> probably be really useful if these changes were emphasised in the
> release notes or somewhere appropriate / obvious to avoid too people
> spamming the list with unnecessary questions as I have just done...
> <slopes of feeling stupid>
>
> Cheers
> Chris
>
>
> Chris Coleman wrote:
>
>> Hi,
>>
>> I'm having trouble with serializing an object through a pointer to
>> it.
>>
>> The code below works in 1.32, but in 1.33 it gives the errors below.
>> I
>> am using Mac OS X 10.4.2, gcc 4.0.
>>
>> I've had a look at the headers, but my template programming isn't
>> good enough to make too much sense of them.  It seems that there is
>> a check
>> to ensure T is not a pointer on line 549 of oserializer.hpp, which is
>> the reason the static assert is failing?
>>
>> Is it correct that objects shouldn't be serializable by pointers to
>> them, or not?
>>
>> Cheers
>> Chris
>>
>>
>> gps_position* p_g = new gps_position(35, 59, 24.567f);
>> oa << p_g;  // causes the error's below
>>
>>
>> g++ -o test test.cpp -lboost_serialization-d-1_33
>> /usr/local/include/boost/archive/detail/oserializer.hpp: In function
>> 'void boost::archive::save(Archive&, T&) [with Archive =
>> boost::archive::text_oarchive, T = gps_position*]':
>> /usr/local/include/boost/archive/basic_text_oarchive.hpp:78:
>> instantiated from 'void
>> boost::archive::basic_text_oarchive<Archive>::save_override(T&, int)
>> [with T = gps_position*, Archive = boost::archive::text_oarchive]'
>> /usr/local/include/boost/archive/detail/interface_oarchive.hpp:85:
>> instantiated from 'Archive&
>> boost::archive::detail::interface_oarchive<Archive>::operator<<(T&)
>> [with T = gps_position*, Archive = boost::archive::text_oarchive]'
>> hash_test.cpp:130:   instantiated from here
>> /usr/local/include/boost/archive/detail/oserializer.hpp:560: error:
>> invalid application of 'sizeof' to incomplete type
>> 'boost::STATIC_ASSERTION_FAILURE<false>'
>>
>>
>>
>> This message has been checked for viruses but the contents of an
>> attachment may still contain software viruses, which could damage
>> your computer system: you are advised to perform your own checks.
>> Email communications with the University of Nottingham may be
>> monitored as permitted by UK legislation.
>>
>> _______________________________________________
>> Boost-users mailing list
>> Boost-users_at_[hidden]
>> http://lists.boost.org/mailman/listinfo.cgi/boost-users
>>
>>
>
>
> This message has been checked for viruses but the contents of an
> attachment
> may still contain software viruses, which could damage your computer
> system:
> you are advised to perform your own checks. Email communications with
> the
> University of Nottingham may be monitored as permitted by UK
> legislation.