[PATCH] server - More tightly integrate the server into a user session

[Luke Yelavich]

On Wed, Jul 15, 2009 at 06:14:29PM EST, hanke at volny.cz wrote:
> > This commit allows for tighter integration of the speech
> > server into a user's session. When the user logs in, the
> > SPEECHD_PORT environment variable is set
> > to a unique port number, which is the base port set
> > in intl/def.h, plus the user's UID.
> 
> Luke, please, how does this interact with the SPEECHD_PORT
> settings done by spd-conf?

I haven't got to that yet, I'll be looking at that soon. Hopefully when I am 
finished the work I have planned, spdconf will no longer be needed in the 
majority of cases.

> I think the correct place to put this variable initialization
> is in ~/.profile according to Freedesktop. It should be read
> by the text console as well as by X session. spd-conf already
> does that, but it would need a little modification to respect
> the better approach based on user id.

One rule in Ubuntu and Debian packaging, is that we don't touch any user home 
directories. There is no other way that I know of, to ensure the environment 
variable gets set on all logins. On Ubuntu, that shell fragment does get run 
even when logging into an X session.

> > When the server gets loaded, it then checks
> > for the presence of
> > the SPEECHD_PORT environment variable, and uses that
> > port for communication
> > if found.
> 
> Great!
> 
> > The environment variable will overwrite any
> > setting in the config
> > file, but will not override the port command-line flag.
> 
> Yes, this is correct.
> 
> > This integration is disabled by default, but can be
> > enabled at compile time
> > with the --enable-session-integration command-line
> > flag given to the
> > configure script.
> 
> What would be the reason to disable it?

I thought that some people building from source, or some distros, may not want 
this enabled. I can easily just enable it by default if there is concensus to 
do so.

THanks for your questions and feedback.

Luke