Info-IBMPC@C.ISI.EDU (Info-IBMPC Digest) (03/06/87)
Info-IBMPC Digest Thursday, 5 March 1987 Volume 6 : Issue 15
This Week's Editor: Billy Brackenridge
Today's Topics:
Archives Available via Modem
Sorry about #13
NARROW.E
Floating-point "bug"
New WordStar
Non-DOS disk
Bibliography Software
Crosstalk and the Sanyo PC Compatible
Today's Queries:
Bug in Wordstar V3.31
Qmodem
Aztec C Problem?
Procomm V. 2.4.2 and DOS 3.2
Partitioning hard disks larger than 32 MB
Unblocked I/O
----------------------------------------------------------------------
Date: 5 Mar 1987 17:16:45 PST
Subject: Archives Available via Modem
From: Billy <BRACKENRIDGE@C.ISI.EDU>
I have moved a copy of the info-ibmpc archives to a BBS running on
an IBM PC with a 60MB disk. All back digests and the entire
source file lending library are available for ascii or xmodem
downloading.
We are running a modified version of the DLX bulletin board system. DLX
is a commercial product written by our own info-ibmpc editor Richard
Gillmann. As far as I know it is the only BBS that allows multiple
simultaneous users on a single processor PC under DOS.
I have added a key word search feature to the standard DLX. The
titles of all articles ever published in info-ibmpc are available
for key word search. If you remember some article but forgot when
it appeared, this feature will give you the answer. Don't expect
the editor to answer anymore queries of the form "a few months ago
you ran an article on X". I will just send you a file with the DLX
phone number.
Now that the system is announced, I feel I have the right to
announce a little vapor ware. I hope to make one line of this BBS
available to amateur packet radio.
We also hope to add network support to DLX. First we will support
FTP access to the info-ibmpc archives and lending library. Tops-20
should go away forever sometime this summer. At that time our PC
will be the only repository for info-ibmpc files.
I also hope to make the system available via telnet login.
DLX should be available 24 hours a day. Eventually we will have a
maintenance schedule but for the first few weeks it may be down at
random hours.
Login and try out the system. It is self explanatory and there are
lots of help files and commands. The modems (Mitsuba super modems)
speak 2400, 1200, and 300 baud. So far these modems are cheap and
have been working fine.
Phone numbers for DLX are:
(213)827-2515
(213)827-2635 (This number will be working when Craig Rogers
returns my modem).
------------------------------
Date: 5 Mar 1987 17:29:52 PST
Subject: Sorry about #13
From: Billy <BRACKENRIDGE@C.ISI.EDU>
Issue #13 was twice the size it should have been. It has issue #14
appended. Many hosts threw it away because it was too large. If
you didn't get issue 13 or 14 send me a note and I will forward them.
------------------------------
Date: 5 Mar 1987 17:31:22 PST
Subject: NARROW.E
From: Billy <BRACKENRIDGE@C.ISI.EDU>
I clobbered NARROW.E I have ordered it retrieved from a backup tape. It
should be restored by Friday afternoon. To those who tried to FTP it.
Please try again later.
------------------------------
Date: Thu, 5 Mar 87 10:37:18 EST
From: Bernard_Tiffany@um.cc.umich.edu
Subject: Floating-point "bug"
The MSC example in Info-IBMPC Digest V6 #13 claimed that there is a
floating-point "bug" in MSC version 3.0 and 4.0. This "bug" occurs
in all computer languages where the conversion from a
floating-point number to an integer number takes place. The
computed value 247.99999999996 is truncated to 247, so one solution
is to try to add a small value like below:
main()
{
int j1, j2, j3, k;
k = 800;
j1 = k * .31+.001;
j2 = ((double)k) * .31+.001;
j3 = (int) (((double)k) * .31 + .001);
printf ("%d, %d, %d\n", j1, j2, j3);
}
Here, 248 instead of 247 is printed.
------------------------------
Date: Thu 5 Mar 87 13:39:34-PST
From: Liquid Len <Asbed@eclc.usc.edu>
Subject: New WordStar
I was a beta tester for the new WordStar Professional Release 4, and I've
written a book on the product, so I'm not an objective reviewer. However,
I think the new WS will solve your printer problems.
In addition to the standard four "user-defined" print commands, ^PQ,
^PE, ^PR, and ^PW, you can CHANGE these commands at will in your file
with four new dot commands: .XQ, .XE, .XR, and .XW. So, even if you've
installed the four commands to do other printing functions, you can change
their functions on the fly. A printer command imbedded in a dot command
line can be up to 24 bytes long.
How this helps!
Vincent Alfieri, Ph.D.
213.257.9866
------------------------------
Date: Thu 5 Mar 87 17:23:28-EST
From: Paul G. Weiss <PGW@XX.LCS.MIT.EDU>
Subject: Non-DOS disk
To: CYAMAMOT%GUMBY@ECLC.USC.EDU
I too have been having this "problem" with more than one machine (an
IBM-AT has well as a COMPAQ II). I call it a "problem" rather than a
problem because it is only an inconvenience - I haven't seen any other
manifestation of the "problem" other than the message from CHKDSK.
If it really bugs you, you can get rid of it by backing up the entire disk
and redoing the whole FDISK-FORMAT routine.
------------------------------
Date: Thu 5 Mar 87 09:18:22-PST
From: Steve Dennett <DENNETT@SRI-NIC.ARPA>
Subject: Bibliography Software
To: reintom@ROCKEFELLER.ARPA
Tom,
you might want to look at "Pro-Cite" from Personal Bibliographic Systems
(P.O. Box 4250, Ann Arbor, MI 48106; 313/996-1580). It is basically a
specialized database designed for bibliographic citations.
Steve Dennett
------------------------------
Date: Thu, 5 Mar 87 8:25:24 EST
From: Kenneth Van Camp -FSAC- <kvancamp@ARDEC.ARPA>
To: hicks@WALKER-EMH.ARPA
Subject: Crosstalk and the Sanyo PC Compatible
>Can someone tell us where to patch crosstalk
>so that it recognizes the serial port on the Sanyo? (To top it
>off, I have no idea what the address for the Sanyo serial port is!
>BUT I CAN MAKE A PATCH IF SOMEONE WILL TELL ME WHERE TO DO IT!!)
Forget it. The Sanyo doesn't even use the same UART that the IBM PC
uses. If you patch in a different address, Crosstalk still won't work
because the control lines are different. There is a version of Kermit
for the Sanyo (available at Columbia cu20b) that works quite nicely;
they also store the source code, so you can use that if you need an
example of how to program the Sanyo serial port. If you like the script
facilities of Crosstalk, a program called Intellicom provides a nice
script facility (although not quite up to Crosstalk's par); I'll get
the name of the company and price if anybody requests it.
--Ken Van Camp <kvancamp@ARDEC.ARPA>
------------------------------
Date: 5 Mar 87 08:15 CST
From: SCTC.AFCC@AFCC-3.ARPA
Subject: Bug in Wordstar V3.31
ANy help with the following problem would be appreciated. You can respond
directly to me via e-mail at SCTC@@AFCC-3 or via the net. Thanks.
There is a nasty bug in the most recent Zenith version of WordStar (v3.31).
While printing a document in background mode and editing another in the
foreground mode, a control-C page-down command drops you out of your
document, out of WordStar, and back to the DOS prompt. Any changes to the
foreground document are lost, and the printout of the background document
halts.
This is not supposed to happen with the DOS "BREAK" feature set to OFF. I
have a screen dump in hard copy showing the dropout to DOS and the state of
the BREAK toggle.
As you know, when BREAK=ON the keyboard is continuously monitored for a
control-C signal to interrupt and abandon a currently executing process,
even if in a runaway condition. When BREAK=OFF, the control-C scan is much
more limited, occurring mainly whenever the system is idle. Since I use
WordStar so often, I set BREAK=OFF in my AUTOEXEC.BAT file on initial
boot-up so I don't have to worry about it. Only it seems I still have to
worry...
------------------------------
Date: Thu, 05 Mar 1987 10:38 5
From: ejs%acorn@oak.lcs.mit.edu
Subject: Qmodem
What terminal does QMODEM emulate? If there are choices in this regard,
what are they? I've tried to configure my host machine as a vt52,
vt100, h19, and raw ansi and none of these seems to work with qmodem
when the later is in "terminal mode." Any help would be appreciated.
------------------------------
Date: Wed, 4 Mar 87 15:13:18 EST
From: Lee A Butler (Space Telescope|mike) <butler@BRL.ARPA>
To: INFO-IBMPC@c.isi.edu
Subject: Aztec C Problem?
While writing a library for Aztec C, I generated the following short
program and compiled it with the large code, large data model. It
appears that the second "for" loop overwrites the array filled in the
first loop. Anyone who can tell me why, point out a bug in my code, or
confirm a bug in the compiler? I'll include the assembly generated by
the compiler for this.
Lee A. Butler
Space Telescope Science Institute 3700 San Martin Dr. Baltimore MD 21218
Arpanet: butler@stsci.arpa | butler@brl.arpa
Usenet: seismo!stsci.arpa!butler |
{noao,astrovax,cfa,charm,nrao1}!stsci!butler
SPAN: SCIVAX::BUTLER (6405::BUTLER)
Phone: (301) 338-4531
---------- begin C source ----------
#define WOW 32766
int big[WOW];
int step[12];
char array[16];
main()
{
register unsigned int i;
for(i=0 ; i < 16 ; i++)
array[i]=i;
for(i=0 ; i < (unsigned)WOW ; i++)
big[i]= i;
for(i=0 ; i < 16 ; i++)
array[i]=i;
}
---------- end C source ----------
---------- begin Assembly output ----------
largecode
codeseg segment para public 'code'
dataseg segment para public 'data'
dataseg ends
assume cs:codeseg,ds:dataseg,es:dataseg,ss:dataseg
extrn $cswt:far
extrn $begin:far
;#define WOW 32766
;int big[WOW];
global big_:word,-4
;int step[12];
global step_:word,24
;char array[16];
global array_:word,16
;main()
;{
public main_
main_ proc far
push bp
mov bp,sp
ifdef $2
add sp,$2
endif
ifdef $3
push di
endif
ifdef $4
push si
endif
; register unsigned int i;
;
; for(i=0 ; i < 16 ; i++)
xor si,si
$7:
cmp si,16
bhis $6
; array[i]=i;
mov ax,si
mov byte ptr array_ [si],al
$5:
inc si
jmp $7
$6:
;
; for(i=0 ; i < (unsigned)WOW ; i++)
xor si,si
$10:
cmp si,32766
bhis $9
; big[i]= i;
mov bx,si
shl bx,1
mov word ptr big_ [bx],si
$8:
inc si
jmp $10
$9:
;
; for(i=0 ; i < 16 ; i++)
xor si,si
$13:
cmp si,16
bhis $12
; array[i]=i;
mov ax,si
mov byte ptr array_ [si],al
$11:
inc si
jmp $13
$12:
;}
$14:
ifdef $4
pop si
endif
ifdef $3
pop di
endif
mov sp,bp
pop bp
ret
$4 equ 1
main_ endp
codeseg ends
dataseg segment para public 'data'
dataseg ends
end
---------- end Assembly output ----------
------------------------------
Date: Wed, 04 Mar 87 23:08 EST
From: EKZMS%CUNYVM.BITNET@wiscvm.wisc.edu
Subject: Procomm V. 2.4.2 and DOS 3.2
Does anyone know how to get Procomm 2.4.2 to work with DOS 3.2?
Please reply to:
EKZMS%CUNYVM.BITNET@WISCVM.WISC.EDU
Thanks.
------------------------------
Date: Thu, 5 Mar 87 09:32 N
From: <ZEELAND%HLERUL5.BITNET@wiscvm.wisc.edu>
Subject: Partitioning hard disks larger than 32 MB
Does anybody have a public domain utility program that will let me
divide a 40 MB hard disk into two or more logical drives, so I can
use all the space on the hard disk.
Thanks ,
Karel van Zeeland.
------------------------------
Date: 5 Mar 87 13:45:57 EST (Thursday)
Subject: Unblocked I/O
From: Marty <Leisner.Henr@Xerox.COM>
Are there mechanisms to perform no delay reads on file handles in ms-dos?
I'm using Aztec C 3.4 and they don't seems to support no-delay reads.
I looked at the dos manual, and it doesn't seem to support no-delay read
requests.
It seems I really would like to be able to implement a select call on
a bunch of file descriptors so I know which ones are ready for input.
Essentially, I want to ask: which file descriptors have pending
input?
Anyone have any ideas how to implement this? Anyone already implemented it?
marty leisner
xerox corp.
leisner.henr@xerox.com (preferred location)
martyl@rocksvax.uucp (alternate address)
[Oh don't you wish! IBM did it to us when they wrote their disk BIOS
to busy wait. All Unix clones bypass the BIOS. I guess Microsoft
had to do busy wait I/O for some contractual reason. The Wendin tool
boxes have non busy wait I/O. 3Comm server software claims to do
this, but as it doesn't do anything else well, I doubt this is a
winner. Like the folks who wrote aztec C. Much of the code I have
written is sitting with conditional assemblies waiting for the day
DOS supports non busy wait I/O. Even Microsoft Pascal supports
no wait I/O "on operating systems which support this feature". -wab]
------------------------------
End of Info-IBMPC Digest
************************
-------