Re: [boost] [serialization] or [spirit] JSON Archives

Subject: Re: [boost] [serialization] or [spirit] JSON Archives
From: Fernando Pelliccioni (fpelliccioni_at_[hidden])
Date: 2010-06-15 10:41:03

On Tue, Jun 15, 2010 at 11:33 AM, Sebastian Redl <
sebastian.redl_at_[hidden]> wrote:

>
> On Tue, 15 Jun 2010 11:12:58 -0300, Fernando Pelliccioni
> <fpelliccioni_at_[hidden]> wrote:
> > On Tue, Jun 15, 2010 at 4:41 AM, Stephan Menzel
> > <stephan.menzel_at_[hidden]>wrote:
> >
> >> On Tue, Jun 15, 2010 at 9:20 AM, Juraj Ivančić
> <juraj.ivancic_at_[hidden]>
> >> wrote:
> >> >
> >> > Have you seen Boost Property Tree library? I have not tried it, but
> it
> >> > does support JSON format to some extent. It might already have all
> you
> >> > need.
> >>
> >> Indeed. A JSON parser in spirit has been done multiple times. The one
> >> in property tree is one example. It should also contain a serializer.
> >> TinyJSON is another. And yet, all of them are spirit classic. So at
> >> least, upon inventing this particular wheel the 3rd time it should be
> >> done in Qi.
> >>
> >> Cheers,
> >> Stephan
> >>
> >>
> >
> > Hi,
> >
> > The parser should be written Qi, there is no other option for me.
> > On the other hand, the objective is not only a parser but a Serializer,
> so
> > I
> > think you should follow the design of Boost.Serialization.
> > I have not checked yet, but I think that Property Tree does not comply
> with
> > these premises.
>
> True. PTree's parser is written in Spirit.Classic (currently, though I
> eventually want to rewrite it), and is definitely not an Archive. PTree's
> JSON serializer does not use Spirit at all.
>
> Sebastian
>

The question is:
Could I make a JSON serializer using the Archive model ?