[PATCH 1/3] spd-conf revised to reflect the current state

[Hynek Hanke]

On 27.8.2010 13:28, Andrei Kholodnyi wrote:

> +# Copyright (C) 2008,2010 Brailcom, o.p.s.
>    
 > a short description might help to understand what this module is doing

Yes, agreed. Freel free to submit a patch:)

>> +        os.system("espeak \"Type in\"")
>>      
> why are several synths hardcoded inside? I'd rather expect a dynamic
> behavior instead.
>    
Falling back to espeak is a reasonable scenario when your Speech
Dispatcher is not working. It might not help, but often it will.

Supporting all synthesizers would equal reimplementing
Speech Dispatcher in spd-conf, making it a circular problem:)

> E.g. if I configured --without-festival, I do not expect to see any
> menus or tests associated with it.
>    
That's a separate thing. I agree with you. Freel free to submit
a patch. I think however that this is of very low importance.
An additional occasionally useless test will do no harm.


>>      def test_alsa(self):
>>      
>   the same for the audio outputs, it should be dynamic
>   we are probably missing an API to retrieve supported audio backends.
>    
This is mostly in case where your Speech Dispatcher is not working,
so any Speech Dispatcher API won't fix it. We probably shouldn't
test alsa if Speech Dispatcher is not compiled with ALSA anyway,
but as before, I consider it low importance.


>>   +            assert(0)
>>      
>   Do we really need this assert? IMO asserts are good for debugging, but
>   not for the stable sw.
>    
This is only used to check internal consistency. If you compile
with -O, it will not be in the final byte-compiled code. Which is
exactly what assert is meant for.

Best regards,
Hynek Hanke

-------------- next part --------------
An HTML attachment was scrubbed...
URL: 
<http://lists.freebsoft.org/pipermail/speechd/attachments/20100827/245baf7f/attachment.htm>