To:       Beta
Fm:       Howie N2WX
Re:       1.1.2 notes
Dist:     Beta

     Please +READ C2987 CC245 for the .HEX of the release described herein.

     1.1.2:  Changes in this release are virtually all attributable to feedback
from Beta testers.  Therefore, if you like anything in it, pat yourselves on
the back.  This is as much your release as it is anyone's!

     Time to drop some names.  Especially thank the following individuals (in
no particular order) for their input is represented by the features and fixes
of this release: KV7B & D, WA7GXD, W3IWI, WB6YMH, AD7I, and KY2D.  (Is your
call missing from this list?  How about downloading a release and making some
comments?!)

     Here are notes releated to 1.1.2:

     FIXES occupy much of these notes.  The following has been repaired, I
think:

     o- Flow release causing data loss of xmit data

     o- Anamolous disconnect when in S16 link state

     o- Retry-out handling when version 2 of l2 works like table now

     o- ID doubling fixed (i think -- could not dupe the bug here)

     o- Sluggish carrier detect repaired, the TNC is much friendlier to other
stns now

     o- Monitor headers aren't displayed for received data from a link when
MCON ON and MON ON

     o- NULLS are handled in int routines, all combinations supported w/o loss

     o- CTEXT packetized at connect time instead of traveling through the
queues

     o- NULL artifact from true-break transparent escape is now captured

     o- rare corruption of data with "cmd:" repaired

     o- 10 second timers (BEACON, CHECK) work at 10 sec intervals now.

     o- and finally, a very troubling intermittent that caused innumerable
problems like more loss of link state, strange stuff in cmd parameters etc.

     CHANGES not result of bugs:

     o- ^Y handled much faster (inside async int routine)

     o- UA responses while in link setup mode are IGNORED if they differ from
the attempted path

     o- Number-only calls allowed to support links to AX.121NA switches...

     o- The date of assembly (mm/dd/yy) appended to the release number

         ex:   Release 1.1.2 12/03/85 - 16K RAM

     o- Interrupt handling speeded up at least two times.  Support for 9600
both sides, fulldup is gauranteed with standard (2.45Mhz clock).  Support for
19.2Kbaud also assured with appropriate hardware mods for 4.9Mhz and faster I/O
clocks.

     (My tests used two TNC2's, both running 9600 sync and async, FULLDUP ON
MAX 7 PACL 0 in trans and CONV mode (ECHO ON and OFF) could not drop characters
thru a link of six fulldup digipeaters and a 28K file.)

     o- The overflow buffer expanded to 256 characters (overflow after XOFF
sent)

     NEW FEATURES:

     o- Failure of a CONPERMed link results in a

     "*** LINK OUT OF ORDER, possible data loss  [opt. daytime stamp]" message

     o- the command      LCStream  ON|OFF    [default ON] added.

     When LCStream is ON, the character immediately following the streamswitch
character is converted to upper case before being acted upon.  When off, the
case is significant.

     o- the command      NOmode ON|OFF    [defualt OFF] added.

     When the parameter is ON, the TNC will -NEVER- jump on its own between
cmd: and conv or trans, or vice versa.  Only user commands (CONV, TRANS, or ^C)
may change the typein mode.

     If the NOMODE parm is OFF then mode jumping is handled according to
NEWMODE and the rev 1 TNC 2 manual.

     ------

     That's it for 1.1.2.  As usual, all comments are appreciated, and most
receive immediate attention.

     Thank you everyone for your invaluable, visible (and I hope, continuing)
input to this project.  73 Howie

     PS: releases prior to 1.1.2 are NOT gauranteed for 9600 baud, and should
not be relied upon in any mode at this data rate!