SRFI 238: Codesets Arthur A. Gleckler (03 Nov 2022 22:10 UTC)
Re: SRFI 238: Codesets John Cowan (04 Nov 2022 02:01 UTC)
Re: SRFI 238: Codesets Arthur A. Gleckler (04 Nov 2022 02:22 UTC)
Re: SRFI 238: Codesets John Cowan (04 Nov 2022 03:03 UTC)
Re: SRFI 238: Codesets Arthur A. Gleckler (04 Nov 2022 03:38 UTC)
Re: SRFI 238: Codesets John Cowan (04 Nov 2022 03:55 UTC)
Re: SRFI 238: Codesets John Cowan (04 Nov 2022 04:03 UTC)
Re: SRFI 238: Codesets Arthur A. Gleckler (04 Nov 2022 04:06 UTC)
Re: SRFI 238: Codesets Lassi Kortela (04 Nov 2022 07:01 UTC)
Re: SRFI 238: Codesets Lassi Kortela (04 Nov 2022 07:04 UTC)
Re: SRFI 238: Codesets Marc Nieper-Wißkirchen (04 Nov 2022 07:01 UTC)
Re: SRFI 238: Codesets Lassi Kortela (04 Nov 2022 12:52 UTC)

Re: SRFI 238: Codesets Marc Nieper-Wißkirchen 04 Nov 2022 07:01 UTC

Am Fr., 4. Nov. 2022 um 05:06 Uhr schrieb Arthur A. Gleckler
<xxxxxx@speechcode.com>:
>
> On Thu, Nov 3, 2022 at 9:02 PM John Cowan <xxxxxx@ccil.org> wrote:
>
>>
>> Yes, it can.  But then the examples
>>
>> (codeset-symbol 'errno 1) => EPERM
>> (codeset-number 'errno 'EPERM) => 1
>>
>> aren't going to work.  The whole idea here is that `errno` is the name of a bidirectional
>> mapping between symbols and numbers, and it can't be.
>
> I still don't understand.  Can't there be a bidirectional mapping between symbols and numbers on each machine?  The mapping doesn't change from run to run on the same machine, at least not between reboots.

If the list can change between reboots, this would be bad if
precompiled or pre-expanded libraries are used.  A macro transformer
that ran yesterday could have inserted the wrong code for today in its
output.