Path: utzoo!attcan!uunet!lll-winken!lll-tis!mordor!joyce!ames!pasteur!ucbvax!UCONNVM.BITNET!SEWALL
From: SEWALL@UCONNVM.BITNET (Murph Sewall)
Newsgroups: comp.sys.apple
Subject: Re: Kermit
Message-ID: <8805271119.aa16326@SMOKE.BRL.ARPA>
Date: 27 May 88 15:49:40 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 72

An interesting return path; it looks nothing like the one in the .sig
(I guess that means it will bounce - ain't UUCP a hoot 8-)

>              ...Anyone want to explain "WARNING KER383.2 IS DELETED"?
>Is that supposed to mean that it _will_ be deleted by the time KER383.1 is
>done EXECuting?

KER383.1 writes the 'T' files containing serial drivers and the KER383.INSTALL
files then calls KER383.2 which loads the main program (the 'B' file
into memory).  There isn't room on a 5.25 inch floppy for KER383.1,
KER383.2, the serial drivers AND KERMIT383 (binary).  So KER383.2 is
DELETED by KER383.INSTALL to make room to BSAVE KERMIT383,A$1000,L$6F00
(maybe L$6E00, Ted changes that as he makes improvements, I use L$6F00
because that's as large as any version has gotten so far).

If the EZ Install crashes while running, you can try the BSAVE by hand.
That works if the Install hung AFTER it got the program code into memory.
It's not too hard to load a serial driver by hand, BSAVE and write a
KERMIT.INIT file with a word processor (or use KERMIT.INIT from a previous
version).

>I downloaded my copy of Kermit 3.83 from Apple2-L and was able to get it
>working after a few tries. It would be nice if EXEC files were limited to
>25K sections.

I think the problem is getting the (borrowed) EXECUTIONER encryption to
work properly.  If that code is broken into two files, EXECing the second
one seems to crash the code that's reading the encryption.

Ted had a (much smaller) 6byte encryption version, but that proved to
travel badly to some sites (the old ASCII-->EBCDIC-->ASCII problem), so
the current INSTALL is in 4byte.  Once you have Kermit to download it
with, you shouldn't need to use word processors to clean up the received
files (it's getting that first working copy of Kermit that's the head-
ache).

>Kermit seemed to work OK at 1200 baud, but was very noisy at 2400 baud

The problem is the phone line.  Sometimes simply hanging up and dialing
again will clear the condition.  The baud rate the problems occur at
depend on the noise frequency.  I call a local BBS at 2400 and have no
noise at all; whereas others calling at 1200 complain at length.

I've used Kermit up to 9600 baud and it works.  Grant Delaney evidently
has it running at 19,200 on his IIgs.

>(normally I get noise free connections at 2400 baud) and locked up a couple
>of times. This required escaping to the Kermit prompt and reconnecting. I'm
>using a //c, so there's no question that interrupts were enabled (no SSC
>DIP switch.)

Using SET FLOW XON?  Ted still hasn't found the bug that causes a lockup
when the Apple keyboard is in use while the host is transmitting characters.
The cure is ^Q (send an XON; apparently Kermit "forgets" it's sent one
more XOFF than it's sent XONs).

>RE: Mouse text in Kermit
>Occasionally when using AE some noise will cause the screen to display Mouse
>text. This is caused by the reception of an ESC (27) and can be turned off
>with a Ctrl-X (24).

The cause can't simply be ESC, as cursor control in VT-52/VT-100 is built
on sequences starting with ESC.  I know that k-esc (default= ^@), E
will restore the characters to normal inverse; I'll have to try ^X the
next time it happens and see if that works too.

---------------------
Disclaimer: The "look and feel" of this message is exclusively MINE!
            (subject to change without notice; void where prohibited)

ARPA:   sewall%uconnvm.bitnet@mitvma.mit.edu       Murphy A. Sewall
BITNET: SEWALL@UCONNVM                          School of Business Admin.
UUCP:   ...ihnp4!psuvax1!UCONNVM.BITNET!SEWALL  University of Connecticut