Re: Vendoring code in a (Non?)GNU ELPA package

[Stefan Kangas]

Daniel Semyonov <daniel@dsemy.com> writes:

> I don't see why it wouldn't be possible, though it would change the
> behavior of `play-sound' slightly, so maybe a new function should be
> introduced instead.
>
> I could try to implement an `async-play-sound-internal' built-in
> function (for GNU/Linux only at first), using the available subroutines
> in "src/sound.c", though it might take me a while as I have near zero
> experience with Emacs' C code, and this will be lower level than the
> implementation in my module.

We could also change the signature from

    (play-sound SOUND)

to

    (play-sound SOUND &optional ASYNC)

or something along those lines.  Though, on the face of things, I don't
think I understand why `play-sound` should block.  Is there a good
reason for that, or is it just an implementation accident?  And are
there any important reasons to preserve its blocking behavior?

> BTW, is using threads (pthreads) acceptable?  I saw many files under
> "src/" that include "pthread.h" but I couldn't find any file where a
> thread is created (though maybe I missed something).
> If not, I'm not really sure how to implement this (I only ever did async
> programming in C using pthreads, and since I don't have any sort of
> formal background in C, I never learned or tried any other method).

Threads sound like an unnecessary complication, and I'm not sure why we
would need them.  Aren't there non-blocking calls that we could use?

Or would that require pulling in pulseaudio?

BTW, if this new feature only works on GNU/Linux at first, then that is
fine according to our guidelines.