running speech-dispatcher-0.7.1 in system mode

[Steve Holmes]

On Mon, Oct 11, 2010 at 11:43:29AM -0500, William Hubbs wrote:
> The way you set up things for future sessions, at least on gentoo, is to
> drop files ending in .sh in /etc/profile.d.  These are sourced as part
> of /etc/profile, so they are run when someone logs in.
> 
> > We would need to dynamically override the contents
> > of /etc/profile. Maybe distros could solve it by including
> > a file that gets overwritten by the /etc/init.d script on
> > start and stop calls. This is hackish, but could actually
> > work if there is no better way.

Yes, ArchLinux does the exact same thing.

> How about something like this in /etc/profile.d.  A file that checks to
> see if there is a system wide version of speechd running, and if there
> is, it sets the speechd_address environment variable.

Now, there's a cool idea; I like that.  That would prevent a problem
with the environment variable being set when a user tries to autospawn
a new process.