PC Online 1997 August/PCO0897.ISO/filesbbs/dos/bdos_xr5.arj/DOC/BTPE_402.DOC

Text File | 1997-06-09 | 8.0 KB | 213 lines

BTPE402 (Binkley Term hacked by Paul Edwards)

This source code is Binkleyterm 2.60 as modified by Paul Edwards,
Fidonet 3:711/934, to make it compile under the Watcom, Borland
CSET and GNU/EMX compilers under OS/2.  Also, it compiles under
Watcom 10.0b (DOS) and Borland C++ 3.1 for DOS.

Please read the license agreement, written by Bit Bucket Software.
To get all the original files, look for BSRC_260.ZIP somewhere
near you (you can FREQ them from 3:711/934 if you want to).

Any changes I have made are Public Domain.

The makefile called "makefile.wat" is set up to use Watcom 10.0b
The makefile called "makefile.bcc" is set up to use BCOS2 1.5
The makefile called "makefile.ibm" is set up to use CSET++/2 2.0
The makefile called "makefile.emx" is set up to use EMX 0.9b
The makefile called "makefile.w16" is set up to use Watcom 10.0b for DOS
The makefile called "makefile.bcd" is set up to use BC++ 3.1 for DOS

The executables provided here have been compiled with watcom with
optimization on.

The main features of BTPE cf Binkley 2.60 are:
A few new compiler configurations supported
1/100 second resolution in logs.
AT commands get printed out
Responses aren't case-changed.
Case-sensitive comparison of "ERROR" etc so as to not get a match on
  Netcomm's "Error Correction" string, also more stringent matching on
  the string, so as to not match on Netcomm's debug info "OK, trying xxx".
Fixed bug in Wazoo session with product >255 causing trap.
Protect against control characters in the log.
Added "Afterconnect" (same syntax as Aftercall)
Ability to debug every character going to or from the modem by setting
  environment variable BINKDBG to some file (OS/2 only)
32-bit OS/2 comms routines, required by EMX, and used in the others too
Most compiler warnings flushed


Compiling:

First of all, you need a utility called "patch".  This is an
INVALUABLE utility, you MUST GET IT!!!  You won't be disappointed
in it, it is worth the hassle of finding and installing.  It will
be a hassle to find and install it though, most likely.  Look for
an archive called "GNUPATCH.*" somewhere near you.  Failing that,
I have modified and recompiled it myself, and you can freq "PATCH.*"
from 3:711/934 to get it.  However, you will still need to stuff
around and get a copy of EMX 0.9b to get the proper runtime DLL,
or RSX.EXE for DOS/Windows, or EMX.EXE for DOS.

The source is distributed as changes to Binkley 2.60.  The file
PATCH.400 is what you need.

You then need to have the original Binkley 2.60 source code extracted
into directories as such:
c:\scratch\binkley\include
                  \src

It is important that the name "binkley" and "include" and "src" are
as above, but the first bit of the directory name (e.g. c:\scratch)
can be anything.

Now go into c:\scratch and do the following:

patch <patch.400

This will patch all the source code with the BTPE changes.

Then, go into the "src" directory, and choose one of the makefiles.
You may want to use "compile.cmd" or "compile.bat".



History:

Version 4.02,  1996-05-12

Hydra was not working, under Watcom or EMX.  I didn't use Hydra
so I didn't ever know about that.  Anyway, that is because it
uses a couple of routines, ComOutSpace (through ComTxBlockTimeout)
and ComTxRemain.  These were new to Binkley 2.60, so I wasn't able
to use the BTPE 3.06 code to implement them.  I had provided dummy
functions for those two routines, but had never actually used them,
because I'd never used Hydra!  Anyway, I have now, and the dummy
routines worked for my test.  There is still a task to implement
those routines properly.


Version 4.01,  1996-05-05

The EMX port reportedly didn't work.  I had only done minimal
local testing of it.  More testing showed that it failed to
read the nodelist, which I remembered from an older version was
to do with structures not being packed.  I brought that code
change across too, and ran it live on my system to make sure it
worked.  Also, I added "+FCON" as a fax string, because of the
way I changed the operation of the response strings, "+FCON"
no longer matches "+FCO".


Version 4.00,  1996-04-27

Binkley 2.60 was released, so I switched to using that, and
merged most of my changes in.  Changed Afterconn to Afterconnect.
BTW, I discovered that the delay that I had thought was due to an
MNP modem delaying a character transmit was in fact Binkley flushing
the comm line straight after a connect.  Binkley 2.60 stopped doing
that, and hey presto!  Used the spawn.asm from msged/sq (ie the one
modified by Matthew Parker), which enabled me to use Watcom to compile
for DOS (no need for masm anymore - yippee!).


Version 3.06,  1996-03-31

A variety of changes.  Improved negotiation when communicating
with brain-dead systems, made afterconn more robust, made the
search for "CONNECT" etc case-sensitive to avoid conflict with
modems that put out debug info such as "Error correction".


Version 3.05,  1995-11-04

Very minor changes.  I made the log messages start at the same
offset of the original Binkley, by adding a couple of spaces.
I also made the debugging mildly faster.


Version 3.04,  1995-08-18

I added the "Afterconn" keyword, so that after you connect, you
can (e.g.) take the modem offhook with "+++", issue a command
to check the line conditions, then reconnect.  Also, the
"Aftercall" keyword is working fine.  The delay thought to be
in 3.03 seems like it was simply a case of MNP compression
delaying a single character on one modem, and V42 got rid of
that.


Version 3.03,  1995-06-24

Due to a bug in EMX 0.9a, I worked around some code in 3.02.  In
this version, I backed out that workaround, so it is imperative
you use EMX 0.9a with FIX03 (or later) if you want to compile
using this compiler.  I also added the "Aftercall" keyword, but
it doesn't seem to work.  Also, I got rid of the unnecessary
delays during session negotiation, although there still seems
to be one left under some circumstances, and it's proving very
elusive to find out what's causing it.


Version 3.02,  1995-03-20

Added support for using EMX 0.9a.  Also made sure it compiled
under Borland C++ 3.1 for DOS.  Also allowed a facility to
allow debugging of every character sent to/from the modem.
Also fixed support for the CSET compiler.


Version 3.01,  1995-03-11

Fixed a problem with it not recognizing the outbound mail because
it was being case-sensitive.


Version 3.00,  1995-03-05

Made it use the same binkley.lng, binkley.scd and binkley.day as
normal Binkley 2.50.  Because of this major change in external
interface, I increased the version by such an extravagant amount,
despite the fact that 2.0 was only released a short time ago!


Version 2.00,  1995-03-04

Made it actually work (I managed to FREQ a file from another BBS).


Version 1.03,  1995-01-31

Made the dummy functions actually compile in properly!


Version 1.02,  1995-01-30

Changed the dummy functions to make them call Dos16SemWait etc
so it might actually be technically correct.


Version 1.01,  1993-11-15

Got it to link by creating dummy functions for the missing Dos* functions.
Got it to read the nodelist and other things by changing various ints to
shorts.  Note that this is still bad practice.  It should have been done
properly with chars only, and not assuming how big an int etc is.  I
haven't bothered.  It does start and finish, but doesn't recognize the
outbound mail (haven't figured out why yet).  Also haven't figured out how
to make Borland link from a response file yet, so kludged up the makefile
instead!


Version 1.00,  1993-11-09

Compiles on both BCOS2 and CSET++/2.  However, I'm not sure what to replace
functions DosSemClear, DosSemWait, DosSemRequest, DosAsync with, so it
doesn't actually link.  I thought I might get a bit of help if I made the
code available now.  Also, I haven't figured out how to incorporate the
changes into the make file, so you have to modify the makefile for Borland
or IBM yourself.  As it is, it compiles for Borland.