Re: [Json] Update Standard to support ECMA-262 BigInt

[David Kemp <[e-mail]>]

Also as an observer, RFC8259 doesn't care what you put in numbers either,
as long as they are digits with no leading zeroes and an optional
fractional part.

"This specification allows implementations to set limits on the range and
precision of numbers accepted."

If an implementation supports bignums (or even int64s) then nothing in the
JSON
spec prohibits it from doing so.  If the implementation supports only IEEE
754 doubles,
then it's the implementation that needs to be updated, not the standard.
If you're
relying on implementations outside of your control, then putting numbers in
strings
is a workaround to prevent them from getting molested along the way.

That said, applications that handle massive datasets, or at least massive
amounts
of serialization, are probably better off using a serialization designed
for that purpose
such as Protobuf or CBOR.

Dave


On Thu, Apr 29, 2021 at 7:22 AM Pete Cordell <[e-mail]> wrote:

> As no one else has answered this, my understanding, more as an observer
> than as a key working group member, is that you would have to put such
> large numbers in a string and RFC 8259 doesn't care what you put in
> strings.  Things like dates, durations, IP addresses, geolocations and
> large numbers in strings are beyond its scope.
>