9-Jul-85 10:07:53-EDT,2007;000000000001
Return-Path: <SYSKERMIT%vax1.central.lancaster.ac.uk@ucl-cs.ARPA>
Received: from ucl-cs by CU20B.ARPA with TCP; 9 Jul 85 10:07:19 EDT
Received: from vax1.central.lancaster.ac.uk by 44d.Cs.Ucl.AC.UK 
           via Janet with NIFTP  id a001544; 9 Jul 85 12:29 BST
Date:		 9-JUL-1985 12:22:20
From:		SYSKERMIT%vax1.central.lancaster.ac.uk@ucl-cs.arpa
To:		"SY.FDC" <SY.FDC%cu20b.arpa%cs.ucl.ac.uk@ucl-cs.arpa>
Subject : PERQ KERMIT bugs

Dear Frank,

   I'm catching up here on a year of saved up messages for you! Below are
some changes and bug notes on the PERQ KERMIT (PQ) sent by Ian Clifton
of Daresbury Laboratories.

             Alan Phillips
               Lancaster University

---------------------------------------------------------------------------

1) In KermitGlobals.Pas, a couple of identifiers, "EscapeChr" and
"DstryCmdFile" needed to be changed to "EscapeChar" & "DstryCmdFiles" to match
the rest of the program. 

2) There is a module "Sleep" which I guess is part of some versions of
Perq Operating System. PQ-Kermit seems to work with all references
to this module ( & also "Nap" which is presumably part of it) commented out.

3) There were a couple of bugs in the way PQ-Kermit handles the setting-up
of the control character quote (See Section 7, "Initial Connection" of the
Protocol manual.
 i) PQ-Kermit compares the QCTL slot [of the data field of the ACK to
a Send-Init] to 0 when it should compare with ASCII(blank).
 ii) PQ-Kermit also applies control-chatacter prefix decoding [to the ACK of a
Send-Init] when it shouldn't. So when a remote Kermit sends a "#" in the QCTL
slot of it`s ACK to PQ-Kermit`s Send-Init, this character & the one following
are decoded as a single control char! I added al extra Boolean global variable
to PQ-Kermit so that the relevant parts of the program can switch prefix
decoding off temporarily. 

-------------------------------------------------------------------------------
*** end of mail ***