17-Mar-86 06:20:30-EST,5677;000000000001
Received: from CUVMA by CU20B with HASP; 17 Mar 86 06:20:28 EST
Received: from USU(JRD) by CUVMA (Mailer X1.23a) id 5876;
          Sun, 16 Mar 86 21:36:58 EST
Date: 16 MAR 86 19:39-MST
From: JRD@USU
To: SY.FDC@CU20B
Subject: MS Kermit for non-IBM machines.

Frank and Howie,
Subject: MS Kermit for non-IBM machines.

        The list below summarizes the modules extracted last week from
Columbia's Kermsrv and which have been tested with the latest MS Kermit
main body code.

        Changes have been made to the msx modules, except for msxtek, to:
   1) add the global entry point VTSTAT for use by Status in mssset.asm,
        this procedure just returns;
   2) ensure the terminal emulation flag, flags.vtflg, is cleared at
        initialization time;
   3) add register save/restore in procedure getbaud (things are dirty there);
   4) revise name of header file from msdefs.h to mssdef.h where necessary;
        the same change was made to msy modules.

Module msxrb, for the Rainbow, has had more extensive changes made. These are
removal of the old memory management calls to malloc and replacement by calls
to new procedure sbrk. The screen buffer code has been modified accordingly,
but it has not yet been tested on a real Rainbow. Since malloc was used
only by the Rainbow and IBM versions of Kermit it is retained for historical
reasons but sbrk is used instead.

        Each Kermit below, again except as noted, was assembled and linked
without error. Each was run on my Zenith 151 clone of an IBM PC. As the
notes indicate below most Kermits caused problems when the serial port was
touched due to hardware differences between machines. A few (the NEC
machines AP3 and APC plus APR) are hostile to my hard disk. The Wang version
had a register corruption problem which caused loss of control during a Status
display; that was fixed by the register save/restore code.

MSXAP3   ASM    20364   3-13-86   7:00p initialization crash, eats hard disks.
MSXAPC   ASM    46509   3-12-86   9:03p same comment as msxap3.
MSXAPR   ASM    21033   3-16-86   1:55p same comment as msxap3.
MSXDM2   ASM    15475   3-12-86   8:56p works ok through part of connect mode.
MSXDMB   ASM      121   3-11-86   6:43p not needed. Left over from old days.
MSXEZP   ASM    32364   3-16-86   2:27p works in connect mode.
MSXGEN   ASM    15017   3-13-86   9:09p works in connect mode.
MSXGRI   ASM    31037   3-16-86   2:35p initialization crash.
MSXHP1   ASM    10724   3-16-86   2:03p works until connect mode entered.
MSXHPX   ASM    16369   3-12-86   8:58p works in connect mode.
MSXIBM   ASM    22611   3-12-86  10:42p Old version of IBM code. Obsolete.
MSXM24   ASM    31902   3-16-86   2:41p works in connect mode.
MSXMBC   ASM    26001   3-16-86   2:46p works until connect mode entered.
MSXRB    ASM    51790   3-16-86   1:36p works until connect mode entered.
MSXTEK   ASM     7135   3-11-86   6:57p Not touched, very non-standard.
MSXTIP   ASM    36342   3-16-86   2:52p initialization crash.
MSXV90   ASM    32707   3-16-86   2:18p works until connect mode entered.
MSXWNG   ASM    14740   3-16-86   2:09p works until connect mode entered.
MSXZ10   ASM    15871   3-15-86   3:53p init problem on IBM due to Bios call.
MSXZ100  ASM    15429   3-12-86   9:10p init problem on IBM due to Bios call.

MSYEZP   ASM    48666   3-12-86   9:26p removed internal def of lnwrap.
MSYIBM   ASM    35433   3-11-86   8:05p Old verson of IBM code. Obsolete.
MSYM24   ASM    76877   3-12-86   9:28p removed internal def of lnwrap.
MSYTIP   ASM    44401   3-12-86   9:31p
MSYV90   ASM    14123   3-12-86   9:19p
MSYV9T   ASM    61582   3-12-86   9:32p not tested. update ref to mssdef.h,
                                        needs sbrk to avoid segment Graph.
MSYZ10   ASM    16016   3-11-86   7:12p
MSYZ100  ASM    16017   3-13-86   7:43p


The updated modules being returned to you (in Howie's dir) are as follows:

msxap3.asm
msxapc.asm
msxapr.asm
msxdm2.asm
msxezp.asm, msyezp.asm
msxgen.asm
msxgri.asm
msxhp1.asm
msxhpx.asm
msxm24.asm, msym24.asm
msxmbc.asm
msxrb.asm
msxtip.asm, msytip.asm
msxv90.asm, msyv90.asm
msxwng.asm
msxz10.asm, msyz10.asm
msxz100.asm, msyz100.asm

Module msyv9t.asm (Victor 9000 Tek 4010) needs updating to use my sbrk
procedure, in mssker, to replace their segment Graph. I did not do so
because of the extensive hardware specific constructions in that module.

Here is a comment added to PRTCHR in msxgen which applies to modules based
upon it, and to comms difficulties in general (i.e., just after booting
DOS thinks the serial port has a received char, it does not really, because
of a short coming in DOS's initialization procedures).
; Note added by [jrd]: The test for char-at-input-port is int 21h function
; 44h (ioctl) sub function 6 (get input status). On many systems an FFH will
; be reported (meaning Ready) even though no char is available; the Ready
; indication is misleading. In such cases the system will wait for a char
; and will appear to be hung. A preferrable method is to use the ROM Bios
; call int 14H function 3 (get port status) and if the lsb of the returned
; 8 bits in AH is 1 then a char is availble at the port; this assumes that
; the machine emulates this ROM Bios operation.
; Lastly, if the current code is used and the system hangs then reboot
; and say  ECHO Hello >COM1  before running Kermit; this should make MSDOS
; truely aware of the port's actual status. Dark grey magic. Good luck! [jrd]

        The updated main body code is being sent along separately.

        Regards,
        Joe D.