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

[Carsten Bormann <[e-mail]>]

On 2024-01-27, at 17:06, Jeremy O'Donoghue <[e-mail]> wrote:
> 
> As a minimum there needs to be guidance about the use of in-built language numeric parsers and formatting in the text.

If it helps, this is what I did in Ruby (which allows all-zeros elision neither in the integer part nor in the fractional part) [1], slightly edited:

    if /p/i =~ text_value
      Float(text_value.sub(/x\./i, "x0.").sub(/\.p/i, ".0p”))
    else
      Integer(text_value) 
    end

This uses case-insensitive matches (/p/i) for the presence of the character p or P (which distinguishes a hex float from a hex integer), for a dot right after the x or X, or a dot right before the p or P.

The equivalent for decimal floats is similar, but more intricately installed in the ABNF, please see [1].

Do you think this approach needs to be hinted at in the document?

Grüße, Carsten

[1]: https://github.com/cabo/edn-abnf/commit/8a5f56#diff-b8a7b57