Re: [Json] Limitations on number size?

[Tatu Saloranta <[e-mail]>]

On Tue, Jul 9, 2013 at 11:11 AM, Markus Lanthaler
<[e-mail]>wrote:

> > On Tuesday, July 09, 2013 6:38 PM, Nico Williams wrote:
> > On Tue, Jul 9, 2013 at 11:23 AM, Bjoern Hoehrmann <[e-mail]>
> > wrote:
> > > I think it is well-understood that if you need particularily small,
> > > big, precise or otherwise unusual numbers in JSON then it's best to
> > > encode them as strings so you can do the string-to-number conversion
> > > at a higher level than whatever JSON library you might be using
> > today.
> >
> > <semifacetious>
> > If that's what we should always do then why can't parsers do it
> > automatically?
> > </semifacetious>
>
> Some parsers have options to support that. PHP, e.g. has a
> JSON_BIGINT_AS_STRING flag
>


Ditto for Java: usually coercion from JSON String to number is supported;
and ability to force output of all or some numeric types as Strings as well.
Additional settings to enable (or not) use of unlimited-length numeric
types.

Things are actually bit easier with data-binding, because expected target
type is the metadata that informs whether to truncate this (or error out),
or to keep full fidelity.

-+ Tatu +-