From: Robert Ramey (ramey_at_[hidden])
Date: 2019-11-30 18:26:47
On 11/30/19 9:15 AM, Vinnie Falco via Boost wrote:
> On Sat, Nov 30, 2019 at 9:08 AM Robert Ramey via Boost
> <boost_at_[hidden]> wrote:
>> ojason_archive .. and ijason_archive
>
> I looked at the code, but it was above my pay grade of understanding.
> Can someone please tell me how this Boost.Serialization integration of
> the JSON format handles the fact that keys in objects are unordered?
The sequence of elements in a serialization archive is determined by the
C++ data structures definition. So the sequence of elements in the input
archive is pre-determined.
serialization does not load any arbitrary data into a specified data
structure. That would be impossible. It loads data formated by the
output archive class with code from a compatible input archive class.
This is why JSON parser optimized for serialization would never be a
good choice for an arbitrary JSON parser. And a general purpose JSON
parser would not be compatible with the serialization library.
Robert Ramey
>
> Thanks
>
> _______________________________________________
> Unsubscribe & other changes: http://listarchives.boost.org/mailman/listinfo.cgi/boost
>