port-or-generator argument in reference implementation Shiro Kawai (04 Aug 2020 21:39 UTC)
Re: port-or-generator argument in reference implementation Amirouche Boubekki (05 Aug 2020 06:35 UTC)
Re: port-or-generator argument in reference implementation Arthur A. Gleckler (05 Aug 2020 06:37 UTC)
Re: port-or-generator argument in reference implementation Amirouche Boubekki (07 Aug 2020 03:08 UTC)
Re: port-or-generator argument in reference implementation Amirouche Boubekki (07 Aug 2020 16:27 UTC)
Re: port-or-generator argument in reference implementation Arthur A. Gleckler (07 Aug 2020 18:19 UTC)

Re: port-or-generator argument in reference implementation Amirouche Boubekki 07 Aug 2020 03:07 UTC

Le mer. 5 août 2020 à 09:38, Shiro Kawai <xxxxxx@gmail.com> a écrit :
>
> Thanks, indeed json-tokens is written to accept a character generator.
>
> It is reasonable to add errata to allow json-generator to take a character generator.
>
> Regarding json-generator protocol: The entry of json-generator describes the protocol of the generator what json-generator *returns*, not the generator json-fold etc. takes, correct?  I was confused initially that json-fold can take a generator that json-generator returns.  Actually, what json-fold takes is a character generator.
>
> The only "protocol" port-or-generator must obey is, if we ever mention it, that json-number-of-character-limit should be handled in that generator.

That is another mistake, json-fold takes a generator of chars so the
protocol is just "generate chars".

Thanks!

>
> On Tue, Aug 4, 2020 at 8:37 PM Arthur A. Gleckler <xxxxxx@speechcode.com> wrote:
>>
>> On Tue, Aug 4, 2020 at 11:35 PM Amirouche Boubekki <xxxxxx@gmail.com> wrote:
>>
>>>
>>> json-generator passes its arguments to json-tokens which accepts a
>>> port or generator. Indeed there is a mistake in the specification
>>> json-generator should mention port-or-generator as argument.
>>
>>
>> Thanks for your quick reply.  Would you like to send me a pull request?

I will.