Re: [Json] JSON and int64s - any change in current best practice since I-JSON

[Joe Hildebrand <[e-mail]>]

If we were going to do any more work on this, I agree that there would have to be a unifying philosophy; I bet that would answer this question.

As it stands, unless there are a few more people who get excited about putting effort into this, it's probably not worth discussing each of the points on the laundry list until we come to conclusions.

— 
Joe Hildebrand

> On Jan 25, 2024, at 3:16 PM, Rob Sayre <[e-mail]> wrote:
> 
> On Thu, Jan 25, 2024 at 1:56 PM Joe Hildebrand <[e-mail]> wrote:
> Yes, unless we decided that significant digits mattered.  This was just copied from JSON5, I don't feel strongly about it myself, but apparently it's a thing that bites people trying to hand-type JSON sometimes?
> 
> That is a philosophical difference your effort should take a stand on. JSON has no comments, and it does not allow trailing commas because "that would mean someone is hand-writing it" (loose paraphrase). I think for the numbers you should check what David M. Gay's strtod things do. There, I think there is no debate, just reality (the GNU versions are close).
> 
> If you write Protocol Buffers or something, there are comments and trailing commas. This is because to disallow them would be a pain for the person adding the next field. You always have to add a field, because there is an inevitable transition phase, even if you intended to make one unused.
> 
> thanks,
> Rob
>