Question with $PLAY_COMMAND

Chris Brannon cmbrannon79 at gmail.com
Fri Mar 5 13:38:56 CET 2010


Steve Holmes wrote:
> The last time I looked when I was getting the problem, I saw that
> [null] was being substituted for the PLAY_COMMAND variable

Well, it turns out that there is a problem with the code that chooses
play_command.  I totally missed it yesterday!
Here is src/modules/generic.c, line 396:

		}else{
		  assert("Unknown audio output method requested");
		}

The assert call aborts the program if a condition is false.
In this case, the condition of the assert is just a string, and that's
always true, under C's definition of truth.  The process doesn't terminate
when it decides that your audio output method is unknown.

I'm going to send a patch for this one, later.
One approach is to default to "play" for play_command.
I think that the play program is commonly available.
It should be a reasonable default.
The other option is to print a warning and call abort.  That seems to be
the intent of the original code.

-- Chris



More information about the Speechd mailing list