Info-IBMPC@C.ISI.EDU (Info-IBMPC Digest) (03/19/87)
Info-IBMPC Digest Wednesday, March 18, 1987 Volume 6 : Issue 18
This Week's Editor: Richard Gillmann
Today's Topics:
Long Pathnames/Command Lines (3 msgs)
C Cursor control (2 msgs)
Advice about text editors: Brief vs. Epsilon (2 msgs)
Tektronix 4010 Emulators (2 msgs)
Some additions to the knet specs
COM3 and COM4
APL Character ROM - Logo incompatibility
VAXmate Design
Logic design development tools
Today's Queries:
Graphics Editor for EGA Wanted
AT Clone Query
Blossom software required
286 accelerator cards and AST RAMPAGE card problem
Clock rate compatibitily Query
Problem with old hard disk
MIRRORS/PC Query
Hard disk upgrade for AT&T 6300+ Query
JRAM Problems
Problem with Hard Disk
INT 24H Disk Error Query
----------------------------------------------------------------------
Date: Mon, 16 Mar 87 10:46:58 +0100 (Central European Time)
From: XBR1D770%DDATHD21.BITNET@wiscvm.wisc.edu (R. Hoenig, THD)
Subject: Large Paths
To: info-ibmpc@c.isi.edu
Volume 6, Issue 16 of the digest contained a question with the subject
"Large PATHs Small Environment"
There are two limitations for the size of PATH (and other environment
variables). The first is the environment size. You can increase the
environment size by patching the command.com or - if you work with
DOS 3.x - by using the parameter /E:nn of the command.com in the
shell-statement of config.sys (or on command line if you load a second
copy of the command.com). With DOS 3.1 you can increase till 1044 byte
by /E:62 (unit is paragraph), with DOS 3.2 maximum is 32 KB (e.g. /E:32768).
The second limitation is, that the command line for the command.com
is limited to 128 byte.
If you want to set the path with the set- or the path-command, you
can only define a longer path by using SUBST. But if you use a programm,
that changes the environment directly, you can also enter a longer
path. We have a programm in our library, that lets you define a path
till full environment size.
Reiner Hoenig, EARN: XBR1D770@DDATHD21.BITNET
Institut for Bussiness Administration PSI: PSI%45615130514::BR1::D770
Technical University Darmstadt Tel.: 049 06151 163063
Hochschulstrasse 1 " " 165360
D-6100 Darmstadt
Fed. Rep. Germany
------------------------------
Date: 16 Mar 87 16:01:33 EST (Monday)
Subject: Long command lines
To: INFO-IBMPC@C.ISI.EDU
From: marty <Leisner.Henr@Xerox.COM>
The talk about long search path's has raisen another sleeping demon
inside me:
How to get long command lines to programs?
I'm currently using Allen Holub's shell and it supports a 2k
environment buffer called CMDLINE. The idea is a program can examine
its environment for the command line and handle it accordingly (or do
it in the c startup routine before main [assuming you use C]).
Somewhat clumsy.
Aztec C supports an interesting feature -- all Aztec programs run from
Aztec's make supports a long command line (much longer than 128
bytes). I have the Aztec source code but haven't been able to
decipher what they're doing special. When I have a change I'm gonna
look into it. Anyone out there knows how Aztec does it? There
doesn't seem to be anything special in the targetted program which
tells the differnce between whether it was envoked by dos or by
Aztec's make.
marty leisner
leisner.henr@xerox.com
------------------------------
Date: Mon, 16 Mar 1987 21:48 PST
From: Gene Hart <MLD3600%UWACDC.BITNET@wiscvm.wisc.edu>
Subject: Problem with Undeletable Directory Structure
To: <INFO-IBMPC@C.ISI.EDU>
I created a directory on my hard disk with the name
\junk1\junk2\junk3\junk4\junk5\junk6\junk7\junk8\junk9\junk10\junk11\junk12
How can I get rid of it? Things are behaving in a peculiar fashion.
cd \junk1\junk2\junk3\junk4\junk5\junk6\junk7\junk8\junk9\junk10
works fine.
cd \junk1\junk2\junk3\junk4\junk5\junk6\junk7\junk8\junk9\junk10\junk11
gives the error message "Invalid directory".
If I am in the subdirectory \junk1\...\junk10 and enter the command
cd junk11
then I get the error "Invalid directory".
If, when in this subdirectory, I enter the command
dir
then I see .,..,and junk11. All are directories.
The commmand
chkdsk/f
gives the error "Cannot CHDIR to c:\junk1\...\junk10\junk11"
"Tree past this point not processed." And it doesn't free lost clusters.
The Norton Utilies will not run on my hard disk. It bombs with an
"insufficient memory" error.
What can I do to get rid of this monster? I am running PCDOS 3.2. Could
it be a problem with my hard disk or is it in DOS?
[Try renaming these subdirectories to shorter names using one of the
public domain utilities that can rename a subdirectory. I believe there
is a limit to pathnames of 64 or 128 characters. -rag]
------------------------------
Date: Tue, 17 Mar 87 05:03:54 PST
From: Primixsys!bagpiper@csvax.caltech.edu
Subject: Screen and printer routines in Microsoft c
To: info-ibmpc@c.isi.edu
I have not used Microsoft C much, but there are no such routines that I
know of in the standard library. To the first two questions I would advise you
to write your own routines. You did not mention what OS you were using but I
will assume you are using MS/PC-Dos and not Xenix, ect. Many MS-Dos C's
include a function int int86(int intnum,struct *regsin, struct *regsout) that
will call bios or dos interrupts. To position the cursor you will have to set
DH equal to the row number, DL equal to the column number, BH equal to the page
number, intnum equal to 10h, and ah = 2(that is interrupt 10h, function 2). It
is very useful to get a book on basic dos routines and familiarize yourself
with them. A good book is "The Programmer's Guide to the IBM PC" by Peter
Norton (also covers the PCjr, XT, and the AT somewhat). There are other good
books around, but this one happens to be the one I am using right now.
As far as redirecting O/P from the screen to the printer, all you have to
do is to replace stdout with stdprn in your O/P statements unless you are doing
complicated O/P (such as graphics) in which case you will have to write a
printer O/P routine. You could write a memory resident routine to catch all
calls to the Dos and Bios video O/P routines and redirect them to the printer.
Good Luck Hacking,
Michael Hunter
cit-vax!bagpiper@csvax.caltech.edu
Occidental College Box 241
Los Angeles, CA 90041
------------------------------
Date: Tue 17 Mar 1987 10:30:31 EST
From: <DIGITS@LL.ARPA>
Subject: C Cursor control
To: info-ibmpc@c.isi.edu
This is a reply to Ron Gershon's question about cursor control from
'C'. He states that there are no built in routines in Microsoft 'C' to
easily control cursor movement and from what I have seen, this is
correct. However, all is not lost if you are using ANSI.SYS. You
can imbed ANSI commands into a 'printf' statement that will allow
cursor control and much more; a FAST way to clear the screen for
instance.
If you have any trouble making this scheme work,, let me know and I
will forward some examples but this is so easy that you should have
no problems at all.
Lou DiPalma
MIT-Lincoln Labs
------------------------------
Date: Mon, 16 Mar 87 12:26:57 est
From: Joshua Marantz <josh@VX.LCS.MIT.EDU>
To: Info-IBMPC@C.ISI.EDU
Subject: Advice about text editors: Brief vs. Epsilon
I feel very strongly that Lugaru's Epsilon is far and away the best PC editor.
Unfortunately, UnderWare, the makes of Brief, have a much better advertising
scheme than Lugaru's, and as a result, my employer committed to Brief,
purchasing licences for every engineer. I refused, however, and made
them get me a copy of Epsilon.
Actually Brief is also an excellent editor. It is fully extensible,
with multiple windows, multiple buffers, you can push into DOS and
compile with a keystroke, and most importantly, it has a good Undo
facility. The undo facility is the only thing its got that Epsilon hasn't.
Epsilon, more than makes up for it in a variety of ways. Like Brief,
it is fully extensible with multiple windows and multiple buffers.
But unlike Brief, it allows you to put DOS in an Epsilon window,
execute a well behaved program (such as a compiler, linker, or Make
program), and WHILE THE PROGRAM IS RUNNING, CONTINUE TO EDIT. The
program's output becomes part of your Epsilon buffer, and you can cut
& paste with it, page-up & page-down to review the output, and so on,
just like you had redirected the output to a file and then editted the
file. It effectively makes the PC a multi-tasking machine. All the
poor Brief users have to look over my shoulder as I compile in the
background while continuing to edit.
My guess is that Epsilon is completely interrupt driven -- whenever
you type a character, the DOS program is interrupted, Epsilon
processes the character, and returns control to the DOS program.
Whenever the DOS program writes a character (well-behaved programs
only, remember) Epsilon grabs the interrupt and inserts the character
in its DOS buffer. Extremely well done.
Another important consideration is that Epsilon searching is orders of
magnitude faster than Brief's. Read in a 42K byte file, go to the
beginning, and search for a string 19 characters long that occurs only
at the end of the file. Brief does this in 38 seconds. Epsilon does
this INSTANTLY.
In general, Epsilon seems to be a much more tightly coded piece of
software. Compared to Brief, the searching is tremendously fast,
while other operations are noticebly but not significantly faster than
in Brief. Writing and reading files, refresh speed, etc.
Both Brief and Epsilon give you a full programming language to write
extensions in. Brief's looks like Lisp, Epsilon's like C. Brief's
default editting environment is unique (and incompatible with the rest
of the world). Epsilon's looks like Emacs (essentially compatible with
the Editor of choice on VAX/VMS, */UNIX, and TOPS-20). I'm sure you
could make Epsilon look like Wordstar if you wanted to. I've written
macros to make Epsilon look like Brief, to give folks a migration path
to the superior Editor. I suspect you could also make Brief look like
Emacs or Wordstar.
Both Brief and Epsilon list for $199, but discounts are usually available
at distributors. It is worth every penny.
Note that I am not in any way connected with Lugaru software, Inc.
-Joshua D. Marantz
Viewlogic Systems, Inc.
P.S. I have a fairly major correction. I claimed that searching for
a 19 character string in a 42K byte file took 38 seconds in Brief.
The problem was that my string contained 4 * characters, which Brief
interprets as wildcards. Using \ to quote the *, the search took
about 2 seconds. Much more reasonable, but not up to par with
Epsilon, where no visible time elapsed. It can be argued that it is a
user interface bug that Brief defaults to regular expression searching
without indicating so. Epsilon has a different command for regular
expression searching. Of course this is all controlled by the user,
since both Brief and Epsilon are completely extensible.
-Joshua Marantz
Viewlogic Systems, Inc.
------------------------------
Date: Wed, 18 Mar 87 12:50:48 PST
From: Jim Anderson <bilbo.jta@CS.UCLA.EDU>
To: info-ibmpc@c.isi.edu
Subject: Epsilon C mode Query
Has anyone tackled the problem of making Epsilon do a decent job of indenting
switch statements?
------------------------------
Date: Mon, 16 Mar 87 14:39:07 pst
From: reynolds@ames-prandtl.arpa (Don Reynolds)
To: info-ibmpc@c.isi.edu
Subject: TEKTRONIX 4010 Emulators
Yet another addition to the list could be Zap, by Paul Fishwick. In addition
to supporting speeds over 9600 baud (for those of us on direct connect modem
eliminators), it supports TEK 4010 vector plot mode, point plot mode, and text
mode while retaining VT52 or VT100/VT102 capability. Last address I have is
Cheshire Cat Software
P.O. Box 5
Devon, PA 19333
Support by Software Systems, (800)821-2492. Not copy protected, of course.
[Usual disclaimers apply -- I personally use ProComm except for Kermit 2.29
for faster file transfers to IBM compatible mainframes that don't understand
parity.]
Best,
Don
------------------------------
Date: Tue, 17 Mar 87 19:28:29 CET
From: Eberhard W. Lisse <LISSE%DACTH51.BITNET@wiscvm.wisc.edu>
To: info-ibmpc@c.isi.edu
Subject: public tek emulator available
I have a public domaine tek4010 emulator which runs on the EGA/AGA/CGA and
with HGCIBM or SIMCGA on the Hercules board.
I'll send it to R. Eberhardt who asked for one in V 6 I 16.
I have sent it to SIMTEL20.ARPA who never aknowledged it and am willing
to send it UUencoded to anybody asking. (With no ACK I mean I don't know
if they have stored it or not, beeing on BITNET only)
Package includes:
Turbo Source to UUdecode
HGCIBM.ARC
SIMCGA.ARC
XARC.COM
TEK.ARC
so anybody can extract it.
el
------------------------------
Date: Mon, 16-Mar-87 17:39:10 est
From: David Farber <farber@UDEL.EDU>
Subject: Some additions to the knet specs
To: info-ibmpc%c.isi.edu@UDEL.EDU
cdp!caulkins writes:
One change to the knet specs: The version you have (9.03 with
solid state nodes) will support up to 15 users. Actually the
electrical properties will support 30 or more, but traffic and contention
put the limit at the lower number. The actual limit is, of course,
situation and user-patience dependent.
Dave C
------------------------------
From: <kubitron@ATHENA.MIT.EDU>
To: INFO-IBMPC@C.ISI.EDU
Subject: COM3 and COM4
Date: Tue, 17 Mar 87 18:46:57 EST
If there are no standard port addresses for COM3 and COM4, how does
software such as BITCOM recognize modems attached to these ports?
I have a modem which is jumper selected to COM3 (to avoid conflicts).
Although BITCOM has no trouble dealing with it, the port address for COM3
is NOT stored at 40:4-5 as would be expected. These locations contain a
zero word.
There must be some standard method of locating the addresses of COM3
and COM4. Too many software packages accept any of 4 serial ports, and
too many hardware cards allow jumper selection of 4 logical serial
addresses.
What's the scoop?!?
---KUBI---
------------------------------
Date: Tue, 17 Mar 87 23:03:59 EST
From: Psych Bitnet Files <YFPY0071%YORKVM1.BITNET@wiscvm.wisc.edu>
Subject: APL Character ROM - Logo incompatibility
To: IBM PC Digest <INFO-IBMPC@C.ISI.EDU>
I have a peculiar problem of incompatibility between STSC APL and IBM
Logo:
To get APL characters with the CGA, STSC supplies a replacement
character generator ROM for the chip on the CGA. The current release
of STSC APL will work without the chip, but with a loss of some 'features'
like full screen editing & screen buffer.
I also use IBM Logo (from Logo Computer Systems Inc), and when
I installed the APL chip, Logo prints the first line or so of its
sign-on display and causes my machine to hang, requiring cold boot.
I have querried IBM, STSC, and LCSI about the problem, to no avail.
LCSI maintains that they are not doing anything funny like looking
for an IBM copyright in the ROM, but the guy who wrote the IBM version
is no longer there, so they are not sure. STSC tech support can't
think of any reason why Logo should crash, etc.
I have tried running Logo under debug, but it was not enlightening.
Can anyone think of a way to try to pinpoint the problem?
Or a better debugging package (pref. PD)?
Final note: I've heard that a Montreal company sells a little board
called Double ROM that plugs into the ROM socket on the CGA and
offers software selection of the character generator chips between
APL & IBM-ascii. Has anyone tried this? Does it work as advertised?
If there are replies of general interest, I will summarize them to
the list.
-- Michael Friendly
Psychology Computer Coordinator
------------------------------
Date: 16-Mar-1987 1134
From: waters%mosaic.DEC@decwrl.DEC.COM
To: info-ibmpc@c.isi.edu, stone@sumex-aim.stanford.edu
Subject: VAXmate Design
When the VAXmate was announced, I sent a detailed description
to INFO-IBMPC digest. Unfortunately, the editor (I don't
remember who) insisted that it was a DEC Rainbow Compatible PC
and therefore, the information should be in a Rainbow Digest.
I just couldn't get the point across that the VAXmate is an
IBM PC/AT compatible system.
As for the VAXmate design, everything was done in-house. No
outside help was solicited for the BIOS or any of the software.
The VAXmate BIOS has a number of extended features for
buffered communications support, ethernet support, and more.
The DEC release of MS-DOS 3.10 has a number of bug fixes over the
IBM release of MS-DOS. In addition, it supports multiple DOS
partitions on a single hard disk (up to four 32Mb partitions on
each 128Mb hard disk installed).
The best part about the VAXmate is the built in Ethernet and the
connectivity to a VAX/VMS system via MS-NET/DECnet. Files can
be transparently shared with a VMS system as if the file were on
your local hard disk.
Lester Waters
Digital Equipment Corporation
Personal Computing Systems Group
WATERS%MOSAIC.DEC@DECWRL.COM (Arpa)
...!decvax!decwrl!dec-rhea!dec-mosaic!waters
(UUCP)
------------------------------
Date: 18 Mar 87 16:18:24 GMT
From: hplabs!rutgers!uw-beaver!dataio.Data-IO.COM!weil@cae780 (Steve Weil)
To: uw-entropy!uw-beaver!rutgers!hplabs!cae780!tekgvs!ericf@uw-entropy
Subject: Logic design development tools
> Can someone recommend any logic design/reduction/development tools for
> the IBM PC?
Eric,
Data I/O - FutureNet has just released a second generation logic synthesis
tool called Dash-Gates. It is built on the ABEL(tm) syntax but adds a
great many new features.
I will list a few of them here, but suggest you contact a sales engineer
for more details and literature.
Device independant design (you don't need to specify a device
until you are ready to generate a fusemap.
Interactive, screen-oriented package. No long edit-complile loop.
Designs may be partitioned into multiple PLDs.
Uses the ESPRESSO reduction algorithm for smaller circuit size.
New syntax for specifying don't care information, reduction
now uses this information.
Completely new state machine algorithm handles synchronous
outputs correctly. Also allows don't cares in state values
for smaller state machines.
Factoring algorithm, breaks logic into multiple levels suitable
for gate arrays or discrete gates.
Dash schematics of final design automatically generated.
New event-driven simulator simulates feedback properly. Simulates
equations (not fuses) so design may be checked before
reduction or partitioning.
Steve Weil
Data I/O - FutureNet
(206) 881-6444 x569
uw-entropy!dataio!weil
------------------------------
Date: Sun, 15 Mar 87 21:28:42 est
From: bloch-gilles@YALE.ARPA
Subject: Graphics Editor for EGA Wanted
To: info-ibmpc@b.isi.edu
I am working on buiding a system where the PC
1- is interfaced with several videodics
2- can overlay text and graphics on the video
3- can store and playback digitilized voices
I have experience with all this except for the digitilized voice.
I would like to find a graphics editor for the EGA, where I can very
simply print a file on the screen from the C language (also how to get
the IBM Seminar papers mentioned in a previous digest about the EGA
card).
Thanx, gilles
------------------------------
Date: Mon, 16 Mar 1987 09:01 CST
To: <INFO-IBMPC@C.ISI.EDU>
From: Dave Benson <C09800DB@WUVMD>
Subject: AT Clone Query
Doe anyone have information or experience with an AT Clone named JS286?
I understand it is made in California. I have not seen anything about
it, but a local vendor is supporting it. Has it been written up anywhere?
------------------------------
Date: Mon, 16 Mar 87 16:52 IST
From: Amir Prat <C44%TAUNIVM.BITNET@wiscvm.wisc.edu>
Subject: Blossom software required
To: ibm pc info bulletin <INFO-IBMPC@B.ISI.EDU>
I just got a letter from 'DEPARTMENT OF HEALTH & HUMAN SERVICES' announcing
a new version of BLOSSOM statistical programs for use with Lotus 1-2-3.
The letter says that the software is found on several bulletin boards. I
tried to locate it in several but could not find it.
I know it is found in the NIH central computer. I will appriciate any
help finding the BLOSSOM, or if someone has allready got it one can send
it over.
Thank you
Amir Prat
bitnet : c44@taunivm
------------------------------
Date: Mon 16 Mar 87 12:45:34-PST
From: Ivan Auger <LAWRENCE.Auger@BIONET-20.ARPA>
Subject: 286 accelerator cards and AST RAMPAGE card problem
To: info-ibmpc@C.ISI.EDU
I have been unable to make AST's REMM.SYS driver to work with either
an Orchid Turbo EGA card or a 286 Express (PC Technologies) card.
I keep on getting a parity error message. I have taken all the
precautionary measure such as making the RAMPAGE start at 640KB,
disabling the cache, etc... The funny thing is that AST's INSTALL
program does detect the RAMPAGE card, and it finds that it has 2meg
on it. Has anybody encountered this problem ? (I'm using V6.2 of the
AST Superpak Utility, the version of REMM.SYS is 3.1, PC-DOS 3.1 and
an IBM PC-XT).
Ivan Auger
(518)473-3382
lawrence.auger@bionet-20.arpa and/or augeri@csv.rpi.edu
------------------------------
Date: Tue, 17 Mar 1987 08:38 CET
To: <INFO-IBMPC@C.ISI.EDU>
From: Karl Keyte
Subject: Clock rate compatibitily Query
Is anyone aware of any of the major IBM PC software products (compilers,
text processing systems, etc..) having requirements to run at a
particular clock frequency?
Karl
------------------------------
Date: Mon, 16 Mar 87 15:18:02 PST
From: Dixon_Low%SFU.Mailnet@MIT-MULTICS.ARPA
To: info-ibmpc@USC-ISIB.ARPA
Subject: Problem with old hard disk
This is regarding my last question about trying to install an old
IBM XT hard drive ( 10 meg. Miniscribe II from a real IBM XT ).
I might have given an incorrect model number on the controller or
maybe was just unclear.
The disk drive did not come with any controller, and I would like to
run it using my Western Digital floppy / hard drive controller card
that is in the AT compatible. The controller is a 16 bit type, for the
AT only.
Thanks.
------------------------------
Date: Tue, 17 Mar 87 16:05 O
From: "Gershon Kunin" <P82041%BARILAN.BITNET@BERKELEY.EDU>
To: INFO-IBMPC@c.isi.edu
Subject: MIRRORS/PC Query
I just saw an ad for a program called MIRRORS/PC from a company called
MASTERSOFT. This program is supposed to convert files from the format
of one word processor to another. Examples of supported formats are
Wordstar, Wordstar 2000, Volkswriter, Word Perfect and Display Write
III. In addition, ASCII-EBCDIC conversion is also supported. Does
anyone have any experience with this or other such software? Before
shelling out $150, I'd like to hear that it works as advertised.
Gershon Kunin
Bitnet: P82041@BARILAN
------------------------------
Date: Tue, 17 Mar 87 14:01:00 cst
From: cody@anl-mcs.ARPA (Jim Cody)
To: INFO-IBMPC@C.ISI.EDU
Subject: Hard disk upgrade for AT&T 6300+ Query
Can anyone tell me about non-AT&T hard disk upgrades for the
6300+ when used with Unix? Am now using an Olivetti 20Mb disk.
------------------------------
Date: Tue, 17 Mar 87 20:14:31 EST
From: Dean Carpenter <ST701979%BROWNVM.BITNET@wiscvm.wisc.edu>
To: Info IBMPC <INFO-IBMPC@B.ISI.EDU>
Subject: JRAM Problems
Just an update on that problem I was (still am) having with the older 512K
JRAM cards. I have since tried them in several other systems, and all
exhibit the same symptoms. That is, after about 5-10 minutes of use, the
ramdisk suddenly develops disk read errors and suchlike. The oddest of
these suchlikes is that the files on the ramdisk become corrupted ...
By corrupted I mean that when comparing a good file from the hard disk to
a copy on the ramdisk, 9E hex has been changed to 9F hex, etc etc. Only
certain bits are being munged, but they seem to vary with individual poweron
cycles. I think something must have somehow fried my system - now I am having
trouble rebooting the thing. Admittedly now I'm not using the *EXCELLENT*
JBOOT software for the JRAM cards that also traps the NMI interrupt and
provides a few other services. But still, it`s weird after running one
program in vanilla DOS pressing C-A-D and hanging the system badly enough
to need the Big Red Switch. Even the NMI button on the card that came with
my copy of the IBM Professional Debug Facility has no effect.
Well, sorry about the rambling, but this has stumped everyone I can tell about
it, (that's a lot). Guess I'll be spending my $40 to have Tall Trees look at
the boards and call me to say that they aren't even any good as paperweights.
Dean Carpenter ST701979 at BROWNVM.BITNET
------------------------------
Date: Wed, 18 Mar 87 14:48 EST
To: INFO-IBMPC@C.ISI.EDU
From: UEG101%URIMVS.BITNET@wiscvm.wisc.edu
Subject: Problem with Hard Disk
a quick question for those experienced with odd-ball upgrades..
one of the depts. here at URI purchased an XT hard drive and
controler from good oled IBM to put in a vanilla PC. The PC, of
course, did not have enough power to support the drive, so they
purchased a 130W power supply. This was of no help. Is it impossible
to use an XT controler with a PC, even if the power supply is
adequate? Thanks in advance.
Paul R. Zonfrillo
Univeristy of Rhode Island, Administrative Computer Ctr.
UEG101@urimvs (bitnet)
------------------------------
Date: 1987 Mar 18 22:27 EST
From: Bob Babcock <PEPRBV%CFAAMP.BITNET@wiscvm.wisc.edu>
To: Info-IBMPC Digest <Info-IBMPC@C.ISI.EDU>
Subject: INT 24H Disk Error Query
I would like to have a large C program I have written recover
gracefully from disk errors on DOS function calls which would
otherwise give the "abort, retry, ignore" message. I know how to
intercept INT 24H, but this sort of error will probably happen in
a library routine rather than my code. Thus, the error handler
doesn't know exactly what operation is in progress, and the
library routine doesn't know that the error is being intercepted.
I would like to be able to do something like return to a top level
menu, at least without leaving any open files, and ideally without
losing track of any allocated core. Is this possible without
modifying the library (I do have source code, but I prefer not to
touch it). My C compiler is Computer Innovations C86, but I think
the problem is generic.
------------------------------
End of Info-IBMPC Digest
************************
-------