Re: [scheme-requests-for-implementation/srfi-208] Add sample C implementation and FFI (#1) Wolfgang Corcoran-Mathe (30 Oct 2020 23:54 UTC)

Re: [scheme-requests-for-implementation/srfi-208] Add sample C implementation and FFI (#1) Wolfgang Corcoran-Mathe 30 Oct 2020 23:54 UTC

On 2020-10-30 21:11 +0100, Marc Nieper-Wißkirchen wrote:
> For which machine ABIs does the C implementation work?

As far as I know (and I am most certainly not an expert in this area),
it should work on any architecture which uses the same endianness for
both integers and doubles.  (cf. John's "tiny handful of processors"
remark, https://srfi-email.schemers.org/srfi-208/msg/15413728/)

I'd rather not include any ifdef hackery to try to handle such systems,
but if there are simple ways to increase the portability of the sample
implementation, please let me know.

--
Wolfgang Corcoran-Mathe  <xxxxxx@sigwinch.xyz>

"It was the best of types, it was the worst of types."
--Atsushi Igarashi, et al