[daip] AIPS TV queries

[David Titterington]

> > 
> > 2) Always, even when 'kleenex' is used to finish, the TV-lock daemon will
> >    remain as an active process and will have to cleared explicity later
> >    using 'kill'.  What's the recommended procedure for closing down the
> >    TV device cleanly so that the ./TVSRV daemons don't accumulate on the
> >    server?
> 
> This is curious.  XAS has the power to kill the lock daemon and does
> so on command, on ESC typed in the XAS window, or on aborts so long as
> they are not at the level kill -9 which inhibits the abort handler.
> So far as I know it works reliably for us but almost everyone here
> uses the Inet form and they also use the default visual :0.0 ratrher
> than the fourth visual.  I can't think that the latter should matter,
> but I am grasping at straws here.

> I have tried with local TVs and always get a message from TVSERV
> saying that it was told to quit by XAS and then it is gone both as a
> process and as a socket file in /tmp.
> 
> Sigh...
> 
> Eric Greisen
> 

Thanks for looking at this, Eric.  It doesn't look as though you are
able to reproduce the behaviour we see with the Sun Rays.  I find the
same as you if I run from a workstation console, i.e. using either INET
sockets or 'tv=local', XAS dies and signals the TVSERV to close if ESC
is typed in the window:

 >TVSERVER told to shut down by XAS
 XAS: Quitting NOW.

 > echo $DISPLAY
 :0.0


On the Sun Rays, however, which are typically started with 'tv=local:0'
and DISPLAY set to, e.g. ':3.0' or 'localhost:3.0', when you abort the
TV with ESC, XAS dies silently (no message) and the TVSERV remains, with
the lock file still present in /tmp.

I'm not sure we're going to get to the bottom of this, but we'll let you
know if we have any more evidence.

Thanks again,
David.

David Titterington,
Mullard Radio Astronomy Observatory,
Cavendish Laboratory,
University of Cambridge,
Cambridge, UK.