[Xastir] KISS TNC tx delays

[Matt Werner]

----- Original Message ----- 
From: "Curt, WE7U" <archer at eskimo.com>
To: "Matt Werner" <kb0kqa at arrl.net>
Cc: <xastir at xastir.org>
Sent: Friday, February 13, 2004 9:43 AM
Subject: Re: [Xastir] KISS TNC tx delays


> On Fri, 13 Feb 2004, Matt Werner wrote:
>
> > This is actually a thread I brought up around Dec 1, 2003.
> >
> > At the time I noticed that the STA light on my 9612 in KISS mode would
light
> > up and stay lit for a long period of time before an actual transmission
> > occured (maybe 10 seconds).  This isn't as big of an issue for regular
> > position transmits, but is more of a problem for messages and a large
> > problem when my station is acting as a RELAY.
> >
> > Curt worked with me (THANKS) and got the delay down to less than 2
seconds.
> >
> > The delay is back again.

First off, I just did a CVS update now and it appears that my interface.c
code is up to date.

> Nine seconds sounds about right, unless you update to the latest
> CVS, checked in 17 hours ago.  It had been noticed that the tnc-stop
> file wasn't given enough time to get flushed to the TNC when closing
> a port.  Three separate 2-second delays were added, which is
> probably the version you have now.

The tnc-stop file isn't used in KISS, is it?

> Erik Burrows just contributed code that would wait for the output
> queue to be flushed and then close the port, and also took out the
> 2-second delays.  The end result is the fastest we can go, without
> waiting unnecessarily.  That code is what is in CVS now.

I still have the large delay.  Am I the only one running KISS that
experiences this delay?

The delay, about 9 seconds still, makes xastir a pain when one tries to run
RELAY or send messages.  The large messages hardly get out on the RF before
xastir is sending it again.  Any time it relay's a mobile, it has such a
large delay that it chimes in well after the WIDE has put it out if it was
heard by someone else.

9 seconds works fine if you are just sending out your position, but is
harder to deal with in other situations.

I'm curious what's changed since we had it down to a two second delay
before?  Two seconds wasn't the best but wasn't bad either.

I'm willing to test some code that has the delays minimized or eliminated so
that we can get to the bottom of this..

Just a little frustrated...  :)

73 - Matt