Re: [Boost-users] [Serialization]class(type) name in xml archive?

From: Robert Ramey (ramey_at_[hidden])
Date: 2005-10-19 01:01:56

The library implementation strives to avoid including redundant data in the 
archive.
That's why things are the way they are.

Note that you've expressed a desire to do this in order to support the edit 
of
XML archives.  In the past I've suggested that this is a bad idea and my
opinion hasn't changed.

But if you insist you want to do it.  Make a derivation of xml_archive which
does the following.

Overrides the serialization of class id, load, object_id etc to maintain a 
table
which relates object_id -> class_id -> class_name. Then overide 
serialization
of save to lookup the class_id, and class name in the table an add them to 
the file.

This won't be terrible efficient - but heck - its XML so its not going to be
efficient anyway.

Good luck.

Robert Ramey

Sean Huang wrote:
> ----- Original Message -----
> From: "Robert Ramey" <ramey_at_[hidden]>
> To: <boost-users_at_[hidden]>
> Sent: Thursday, October 13, 2005 4:33 PM
> Subject: Re: [Boost-users] [Serialization]class(type) name in xml
> archive?
>
>
>> That may be helpful - but that's the best I can do.  Note that the
>> class_id # in the xml output indicates which type is being handled.
>> I would hope that some sort of XML viewer could make sense of it.
>>
> Do you think it is reasonable to add an option so that users could
> choose to include class names?
> I do not mind adding a lookup table at the end of the archive that
> maps class_id to class_name. But I do not see an easy way to do that.
> Can you shed some light on it?
>
> Thanks,
>
> Sean