Re: No feature for dynamic modules?

[Ulrich Mueller]

>>>>> On Sat, 15 Aug 2020, Eli Zaretskii wrote:

>> Looks like there is no feature associated with dynamic modules. One can
>> of course test if (e.g.) function module-load is bound, but it feels a
>> little clumsy.
>> 
>> Shouldn't emacs-module.c provide a feature?

> We don't normally provide features related to C-level optional
> features.  What's wrong with testing a known function for being
> fboundp?  Quite a few of optional features need that, for example
> native JSON support.

Sure, that WFM.

However, some features _are_ provided in the C source:

   $ grep '^[ \t]*Fprovide' src/*.{c,m} | wc -l
   24

These include "dbusbind", "inotify", "lcms2", and "multi-tty". So is
there an exact rule when a C-level feature should be provided, or are
these historical accidents?