Re: no scm_foreign_object_p ?

[Matt Wette <matt.wette@gmail.com>]

On 1/23/21 12:04 PM, divoplade wrote:
> Hello,
>
> Le samedi 23 janvier 2021 à 14:29 -0500, olivier.dion--- via General
> Guile related discussions a écrit :
>> On Sat, 23 Jan 2021, Matt Wette <matt.wette@gmail.com> wrote:
>>> I started using foreign objects but want something akin to
>>> number?,
>>> string? etc.
>>> Is there something that works here?
>>> I was hoping to see scm_foreign_object_p(type, object).
>> The only thing I see in the manual is
>> `scm_assert_foreign_object_type`.
> That's what I noticed too; if you want something finer you need to
> write scheme code.
>
> Since foreign pointers are GOOPS objects and foreign pointer types are
> classes, maybe the is-a? predicate could work, with the foreign pointer
> type as the class?
>
> divoplade
>
>

GOOPS should not be part of the solution.  I'm thinking of punting 
foreign-object
and see if using some combination of structs and bytevectors will work 
out better.