[PATCH] pico sends audio by chunk

[Luke Yelavich]

On Sun, Feb 02, 2014 at 06:38:07PM CET, Laurent Pelecq wrote:
> Hello,
> 
> I can't make the pico module works. It seems that it sends the audio
> data by chunk of 128 bytes, but module_tts_output expects the full
> text all at once. With Alsa, the device is initialized and
> uninitialized for each chunk.

The ALSA code is known to be unreliable of late. Do things behave any 
differently if you use another audio output, say via libao with ALSA?

> The following patch works for me. It accumulates the samples in a
> buffer before playing it.

This doesn't apply against git master. Please rebase against master, and 
resubmit.

Thanks

Luke