Re: audio/flite: remove flush in play_wave()

[adr <adr%sdf.org@localhost>]

On Tue, 20 Feb 2024, Brian Buhrow wrote:
> 	Hello.  So If I understand this right, the notion of drain and flush are reversed in modern
> versions of the Alsa code with respect to the traditional Sun API
> and the difference between flite-1.3, which is what I'm using, and Flite-2.2 which is what
> you'r patching against, is there is an extra audio flush in the wave thread of the new
> version of Flite.  Do I have this right?  No wonder you've been confused!  And, no wonder my
> patch looks goofy!
> -thanks
> -Brian

Alsa uses snd_pcm_drain() as sun drain, and snd_pcm_drop() as sun
flush. I don't know the origen of the meaning of those terms chosen
by the flite developers.

The problem I have to solve now is why eflite (latest version,
older versions don't work correctly with new flite) doesn't behave
like an asynchronous system. I think that this could be related
with the locks chosen and how netbsd multiplexes the audio devices.
The same case as with espeak-ng, but note that flite doesn't provide
the thread mechanisms, it claims to be thread safe and in this case
is in eflite where the threads are managed.

Let's see what I find...

adr