[Freeswitch-users] playback from hadoop

[Andrew Thompson]

On Tue, Nov 10, 2009 at 11:29:37PM +0800, mark morreny wrote:
> Hi
> 
> Thanks for the tips.  May I ask how to split the file from hadoop to the
> shell?  Is it like copying the file to certain dir?
> 
> I can't find any mod_shell_stream related info from the wiki.  Does anyone
> know how to use it?
>

mod_shell_stream is undocumented, but from reading the code I gather it
works like this:

Module calls fork() and in the child process it runs an arbitrary shell
command (specified in its config file?). The parent process then reads
raw audio data from the child process and uses it as an audio source.

So basicially you could write the shell command in anything, so long as
it outputs raw audio to FS.

Or maybe I read the code wrong when I skimmed over it. If you do get it
working, please contribute some documentation to the wiki.

Andrew