t-rayc@microsoft.UUCP (Raymond CHEN) (07/03/90)
[
Now that comp.sys.ibm.pc has split, this file should also be split
into pieces. Any volunteers?
Differences between this version and last month's version are
described in a companion article. There have been some significant
changes. --rjc
]
[@(#)faq.csip 4.8 Last revised on 7/2/90]
TABLE OF CONTENTS
CHAPTER 0: INTRODUCTION
0.1) What is this file?
0.2) I have a question that's not in this file. How should I ask it
to increase the chances that I'll get a response?
0.3) Somebody just said that a { Mac, Amiga } beats the socks of a PC.
I have my flame thrower ready. Any last words?
0.4) Somebody just asked a really simple question. What should I do?
CHAPTER 1: OBTAINING PROGRAMS FROM THE ARCHIVES
1.1) What is anonymous ftp? How do I use it?
1.2) How do I use the programs that are posted to comp.binaries.ibm.pc?
1.3) I've uudecoded it, and the file is on my PC. Now what?
1.4) The unpacking program said that the archive was corrupted.
1.5) What public archives exist, and what are their IP numbers?
1.6) I do not have ftp access. Can somebody email me the files?
CHAPTER 2: PROGRAMMING
2.1) How do I program the mouse/joystick? What does interrupt X do?
How do I program my EGA/VGA/whatever graphics card?
2.2) How do I do direct video access from C? Read/write an arbitrary byte?
2.3) How do I reboot my computer via software?
2.4) What is the difference between extended/expanded memory, and
how can I access it from my C/Pascal program?
2.5) I get the message "floating point formats not linked" when I
try to run my C program. Why?
CHAPTER 3: SOFTWARE AND VENDORS
3.1) Has anyone had any experience with {company X, product Y}?
3.2) I'm having trouble with product X. Can anyone help me?
3.3) What is the file format that application X uses?
3.4) Rumor has it that company X is working on a new version of product Y.
Does anybody have any information about it?
3.5) I don't like my keyboard's layout. How can I swap the X and Y keys?
CHAPTER 4: HARDWARE
4.1) What are the differences between chip X and chip Y?
4.2) Can I move floppies around freely between 1.2M and 360K drives?
How about formatting a 360K floppy as 1.2M? 720K as 1.44M?
4.3) How do I do a low-level reformat of my hard drive?
4.4) What does interleave mean? What interleave should I use?
4.5) What do ESDI, SCSI, RLL, MFM mean?
4.6) Can I use an RLL controller on my MFM drive?
4.7) Should I leave my PC turned on all the time? What about
mounting it sideways?
CHAPTER 5: MISCELLANEOUS
5.1) How do I undelete a file? Recover from an accidental reformat?
5.2) Where can I get curses? vi? emacs? How do I remove a TSR from memory?
How do I write my own TSR? Is there a program that does X?
5.3) What about GIF images? How do I view a GIF image?
5.4) How do I turn the ECHO off from a batch file?
5.5) How can I set a long path? How can I enlarge my environment?
How do I get/set an environment variable from a C/pascal program?
5.6) How do I install a new version of DOS on my hard drive?
APPENDIX A: ANSWERS TO MORE QUESTIONS and AVAILABILITY OF THIS FILE
--------------------------------------------------------------------------
CHAPTER 0: INTRODUCTION
0.1) What is this file?
This is a list of questions that seem to pop up repeatedly in this
newsgroup. The purpose of this article is to answer them once and
for all so as to improve the signal-to-noise ratio. There are also
bits and pieces of etiquette hidden in the answers, so even if you
aren't particularly fascinated by the topic, skim through the answer.
You might learn something.
The mentioning of a particular company is not meant to be interpreted
as an endorsement, merely a presentation of information which is
requested often. If you feel I am showing favoritism, disagree with
these answers, find any errors, or otherwise want to say something
regarding them, send email to raymond@math.berkeley.edu.
Special thanks go to Prof. Timo Salmi and Keith Petersen for providing
much of the raw material from which Chapter 1 was built. Thanks also
to all the people (too numerous to list) who contributed to this file.
0.2) I have a question that's not in this file. How should I ask it
to increase the chances that I'll get a response?
When asking a question, choose an appropriate subject. The volume of
this newsgroup is so high that articles with subjects like "Help" or
"A question" or "IMPORTANT READ THIS NOW" are likely to be ignored.
Moreover, spend some time proofreading your article. Use your
system's spellchecker to check words whose spellings you are unsure
of. Pretend your article is going to be published around the world.
After all, it is.
You should probably also check that your question hasn't been asked
to death before. Many of the files on my email server address
commonly-asked questions whose answers are too long to list here.
I also keep an index of previously-asked questions on my email server.
Instructions on using the email server are posted separately.
0.3) Somebody just said that a { Mac, Amiga } beats the socks of a PC.
I have my flame thrower ready. Any last words?
Oh, for crying out loud. About twice a year some immature bozo
posts a note like this and the flames don't die out for weeks. If
you consider yourself a mature human being, bite your lip and keep
quiet. Let the nincompoop think he's won. If you absolutely
positively can't keep quiet, flame the sucker by email. But
please, don't make the rest of us have to suffer through it.
While I have your attention, here are a few more subjects that
we're sick of seeing:
FCC to impose surtax on modem usage: This is simply not true.
Consult the file "fcc" on my mail server for a letter from
the FCC debunking this rumor.
Make Money Fast: They're all pyramid schemes, and they're illegal,
especially the one from Dave Rhodes.
Dying Boy's Last Wish: He broke the record in 1988, so stop it already.
Besides, the poor child's address keeps changing, ranging from London
to Atlanta.
0.4) Somebody just asked a really simple question. What should I do?
Send him the answer by email. Don't post the answer, since
(trust me) about a dozen people will do it anyway. If you
want, drop me a line so I can add the question to the FAQ archives.
CHAPTER 1: OBTAINING PROGRAMS FROM THE ARCHIVES
1.1) What is anonymous ftp? How do I use it?
Ftp stands for File Transfer Protocol, a standard for transferring
files among various computers. "Anonymous ftp" is a way of making
files publicly available for transfer via ftp. Usually, this is done
by logging in as userid "anonymous" and giving your electronic mail
address as the password. (Although rumor has it that nobody actually
saves the passwords typed to the ftp program.) It's only polite to
minimize your connect time, since each machine can service only a
small number of ftp users at a time.
Beyond this rather sketchy description, the details vary from machine
to machine. On many UNIX machines, you can get information by typing
"man ftp". In any case, check with your sysadmin for information
peculiar to your site.
Here is what a typical ftp session might look like.
ftp 26.2.0.74 The IP address for the SIMTEL20 archives
anonymous The special userid for anonymous ftp
raymond@math.berkeley.edu My email address is the password
cd pd1:<msdos.gif> Switch to the proper directory
tenex Magic word, see below
get gif_lib.zip Get the Turbo C GIF library (include source)
quit Clean up and leave
I could have gotten away with typing "ftp wsmr-simtel20.army.mil"
instead of using the IP address.
1.2) How do I use the programs that are posted to comp.binaries.ibm.pc?
Follow the instructions in order to cut and paste the files
in the proper order. The resulting file must then be
(1) "uudecode"d to a binary file,
(2) transferred to your PC.
The order in which the above steps are taken is not important,
although the usual order is as shown above. If your host machine
does not have the uudecode program (e.g., if it is not a UNIX
machine), you will probably have to do the uudecoding on your PC.
Here's a script to perform step 1 on UNIX machines:
#! /bin/sh
sed '/^END/,/^BEGIN/d' $* | uudecode
If you saved the pieces as "part01" "part02" and so on, then you
just type "combine part*" and everything gets decodeded automatically.
1.3) I've uudecoded it, and the file is on my PC. Now what?
Most files have to be unpacked before you can run them. The
appropriate unpacking program to use can be determined by inspecting
the file extension. Some of the more common ones appear in the
table below.
Exten- Program Full name of program package Archive sites that
sion to use and directory on SIMTEL20 use this format
ARC PKUNPAK PD1:<MSDOS.ARC-LBR>PK361.EXE SIMTEL20
ZIP PKUNZIP PD1:<MSDOS.ZIP>PKZ102.EXE grape
ZOO ZOO PD1:<MSDOS.ZOO>ZOO201.EXE comp.binaries.ibm.pc
PAK PAK210 PD1:<MSDOS.ARC-LBR>PAK210.EXE chyde.uwasa.fi
^^^ note: this is *different* from PKPAK.
LZH LHARC PD1:<MSDOS.ARC-LBR>LH113c.EXE chyde.uwasa.fi
Z COMPRESS PD1:<MSDOS.SQ-USQ>COMPRS16.ARC unix archive sites
The proper unpacking program should be available on the system
from which you obtained the packed file. Consult the "read me"
file for that particular archive site for information.
1.4) The unpacking program said that the archive was corrupted.
The two most common reasons for this are
(1) failing to use the magic word "tenex" (when connected to
SIMTEL20 and other TOPS20 systems) or "binary" (when
connected to UNIX systems) when transferring the file
from an ftp site to your host machine. The reasons for
this are technical and boring. A synonym for "tenex" is
"type L 8", in case your ftp doesn't know what "tenex" means.
(2) failing to use an eight-bit binary transfer protocol when
transferring the file from the host to your PC. Make sure
to set the transfer type to "binary" on both your host
machine and your PC.
Here is a way to check if you are doing binary transfers correctly.
Compile the following one-line C program: main(){puts("\250\n\250");}
Run it by typing "a.out >binary.fil". Transfer the file
"binary.fil" to your IBM PC, and TYPE it to the screen from the
DOS prompt. If you are doing things correctly, you should see
the following:
?
?
except that the question marks will be upside-down.
1.5) What public archives exist, and what are their IP numbers?
The most famous such is WSMR-SIMTEL20 [26.2.0.74], the White
Sands Missile Range in New Mexico USA.
This is a very busy site, so try connecting at really obscure
times of day. (SIMTEL20 is in the Mountain time zone, -7 hours
relative to GMT.)
In the PD1:<MSDOS.FILEDOCS> directory, the file AAAREAD.ME
has details on file directories and descriptions, and SIMIBM.ARC
is an index of the MSDOS archives.
Other popular archives include
Site Intro File Direct questions to
=================== ========== ===================
wuarchive.wustl.edu README archives@wugate.wustl.edu
128.252.135.4 (Washington University, St. Louis, MO USA)
A duplicate of the SIMTEL20 archives is kept in
/mirrors/msdos.
grape.ecs.clarkson.edu 00readme root@grape.ecs.clarkson.edu
128.153.13.196 (Clarkson University, Potsdam, NY USA)
chyde.uwasa.fi 00readme Timo Salmi (ts@chyde.uwasa.fi)
128.214.12.3 (University of Vaasa, Finland)
My email server has the index files for SIMTEL20, grape, and chyde.
A much larger list of ftp sites is posted to comp.archives
sporadically. You can also send the message "send help" to
comp-archives-server@twwells.com (or to comparc@twwells.com
if your mailer doesn't like long userids) to obtain information
on how to obtain information about archive sites. (You might
also look at the file "ftp.list" on my mail server.)
1.6) I do not have ftp access. Can somebody email me the files?
To obtain a file from SIMTEL20 via email, send mail to an email
server (e.g., listserv@vm1.nodak.edu or listserv@vm.ecs.rpi.edu) with
no subject and containing the single line
/PDGET MAIL PD1:<MSDOS.GRAPHICS>VGAIMG12.ARC UUENCODE
for example. Do not include a signature because it confuses
the server. No more than 100k or 3 files per day will be sent.
For general instructions, send the message "GET PDGET HELP".
BITNET users can also use send LISTSERV commands to LISTSERV@NDSUVM1
or LISTSERV@RPIECS to accomplish the same thing.
European users can use EARN TRICKLE servers. Send commands to
TRICKLE@site, where "site" is one of the following:
AWIWUW11 (Austria) DB0FUB11 (Germany) EB0UB011 (Spain)
BANUFS11 (Belgium) DTUZDV1 (Germany) TAUNIVM (Israel)
DKTC11 (Denmark) IMIPOLI (Italy) TREARN (Turkey)
BITNET users can send email to BITFTP@PUCC to perform ftp.
To receive instructions, send email consisting of the single
word HELP to BITFTP@PUCC. Using this service to ftp to SIMTEL20
is discouraged, however.
If you don't even have email, most MSDOS SIMTEL20 files are available
for downloading on the Detroit Download Central network at 313-885-3956.
It is not a free system but the hourly cost is only 17 cents.
It's also accessible on Telenet via PC Pursuit and on Tymnet via
StarLink outdial.
CHAPTER 2: PROGRAMMING
2.1) How do I program the mouse/joystick? What does interrupt X do?
How do I program my EGA/VGA/whatever graphics card?
A comprehensive listing of functions available via interrupt
calls may be FTP'd from CS.CMU.EDU [128.2.222.173] as file
interrup.zip in directory /afs/cs/user/ralf/pub or from
SIMTEL20 as PD1:<MSDOS.INFO>INTER290.ZIP (as of 1 MAy 1990).
This listing is updated four or five times per year.
A list of books on programming the IBM PC can be found in the file
"books" on my mail server.
2.2) How do I do direct video access from C? Read/write an arbitrary byte?
The hard-core (hard-code) method is to say
#define screen ((char far*)0xB8000000) /* B800:0000 */
and treat "screen" as if it were an array. A more flexible
approach is to say
char far *screen = (char far*)0xB8000000;
so your program can assign a value to "screen" depending on
whether the program is being run with a color or monochrome card.
(Use 0xB0000000 for monochrome cards.) See also the file
"flickerfree" on my mail server for tips on doing flicker-free
video access.
To read and write arbitrary bytes, you can use essentially
the same method, substituting whatever address you want in
place of 0xB8000000. For those fainter of heart, you can use
the peek() and poke() macros in <dos.h>. A similar trick can
be used for accessing/setting the segment or offset portion
of a far pointer.
2.3) How do I reboot my computer via software?
Although some people will suggest using interrupt 19h, that method has
its problems since it doesn't reset lots of things (e.g., the
interrupt vectors). A more effective method is to store the magic
number 0x1234 into 0x40:0x0072 and then perform a long jump to
0xffff:0x0000. The magic number suppresses the memory test.
2.4) What is the difference between extended/expanded memory, and
how can I access it from my C/Pascal program?
Expanded memory is a way of allowing a large quantity of memory all to
share the same memory address. The only way to access it is via
EMS function calls. (See the interrupt list.)
Extended memory is memory that lives above the 1Meg boundary.
You can access it via XMS function calls (see the interrupt list),
or you can switch into protected mode by yourself and access it
directly (risky, since you have to make sure you don't conflict
with any other program that uses extended memory). Much safer
is to use a DOS "extender" program that does this dirty work
for you.
A useful mnemonic is that exPanded memory uses Paging.
Allegedly, Zortech C has compiler support for expanded memory, and the
shareware CXL library (In PD1:<MSDOS.C> on SIMTEL20) provide EMS
handling for MSC and other compilers.
More details can be found in the file "memory" on my email server.
Books which describe the use of expanded and extended memory are listed
in the file "books" on my email server.
2.5) I get the message "floating point formats not linked" when I
try to run my C program. Why?
This is a painfully well-known bug in Turbo C. Just #include <math.h>
and stick the line "(void)exp(1.0);" at the start of your main() and
that should convince the linker to include the floating point formats.
A patch exists to remedy this bug.
Patches for Turbo C can be obtained from SIMTEL20, or from LIB 5 of
the Borland Forum on CompuServe. There is also a Turbo-C mailing list
on LISTSERV@UCF1VM.BITNET.
CHAPTER 3: SOFTWARE AND VENDORS
3.1) Has anyone had any experience with {company X, product Y}?
Here's a checklist:
[ ] Consult the files "products" and "old.questions" on my mail
server to see if somebody else has asked the same question and
to see who responded. Send them email asking them for their
opinions.
[ ] Write a letter to the Better Business Bureau.
[ ] If you're still not satisfied, then post a request. But it
is now your duty to collect the replies and edit them into a
GOOD summary. Concatenating all your responses does not
qualify as a good summary. Depending on the level of
interest you receive, you may wish to post your summary.
In any case, it is now your duty to keep the summary and
offer it (via email) to anyone who asks the same question later.
If someone else posts a request for comments and you are also
interested, do *NOT* post an article saying "Me, too!" For one thing,
people will probably ignore you. Instead, send email to the original
poster requesting that he send you a copy of his summary.
3.2) I'm having trouble with product X. Can anyone help me?
Contact the customer support department of the company who
produced the product. Many companies will have a toll-free
number for that purpose. Use it. After all, you paid for it.
3.3) What is the file format that application X uses? What are the GIF,
TIFF, BGI, OBJ, EXE file formats? How can I write my own BGI driver?
Does anybody have a BGI driver for X?
There are at least two books that cover this information:
File Formats for Popular PC Software - ISBN 0 471-83671-0 $25
More File Formats for Popular PC Software - ISBN 0 471-85077-2 $25
both by Jeff Walden from Wiley Books. Between the two, they cover
Lotus 1-2-3, MultiMate, Symphony, dBASE II & III, IBM Plans+,
Supercalc 3, Multiplan, DIF, ABILITY, Volkswriter, Supercalc 4,
Microsoft Rich Text Format, Framework, Reflex, WordPerfect, and
Superproject.
If you still can't find what you want, you can always contact the
company that produced the application.
GIF, TIFF and BGI file formats are available on SIMTEL20, and the
formats for .OBJ and .EXE files can be found in the MS-DOS Encyclopedia.
On SIMTEL20 is also the Borland BGI toolkit, which can also be obtained
directly from Borland. Send the message
old /BGI/
to my email server for information about obtaining BGI files.
3.4) Rumor has it that company X is working on a new version of product Y.
Does anybody have any information about it?
If it hasn't been announced, then the people who could possibly answer
your question are not allowed to tell you.
3.5) I don't like my keyboard's layout. How can I swap the X and Y keys?
Look in SIMTEL20's PD1:<MSDOS.KEYBOARD> directory. You'll find
a whole slew of programs that let you swap keys around.
CHAPTER 4: HARDWARE
4.1) What are the differences between chip X and chip Y?
These are thumbnail sketches. In general, as the chip number
goes up, the speed goes up, so such comments will be omitted.
This list is far from comprehensive.
8086: The original. 16-bit bus, addresses 1Meg.
8088: An 8086 with only an 8-bit bus.
80186: An 8086 with a few more instructions.
80286: An 80186 with protected mode. Can access 16Meg of memory
in protected mode. Segments still limited to 64K.
In real mode, it looks like an 80186.
80386: A slightly buggy version of the...
80386DX: An honest 32-bit CPU. An 80286 with better protected
mode and a "Virtual 8086" mode. Can access 4Gig of memory.
No 64K segment limitation. In real mode, it looks
pretty much like an 80186.
80386SX: Identical to the 80386, except it has only a 16-bit bus.
80486: An 80386 with a built-in floating point unit.
Officially called the i486.
The 80x87 family are floating point units (FPUs) for the 80x86 family.
8087: FPU for 8086 8088 80186
80287: FPU for 80286. Early 80386s can use it, too.
80387: FPU for 80386. Faster than the 80287. Better trig.
80387SX: FPU for 80386SX. Otherwise the same as an 80387.
The NEC Vxx series are CMOS replacements for parts of the 80x86 family.
V30: 8086 replacement, has performance and instructions of
80186, plus a few more instructions, and an 8080
emulation mode. Expect speedups of roughly 15%--30%.
V20: An 8-bit version of the V30. 5%--10% faster than the 8088.
Also, note that many benchmarks exaggerate the relative performance
of the chips since they tend to emphasize intructions which have
been significantly sped up in the newer chips.
4.2) Can I move floppies around freely between 1.2M and 360K drives?
How about formatting a 360K floppy as 1.2M? 720K as 1.44M?
After formatting a 360K disk in a 360K drive, that disk should be
written to ONLY by another 360K drive, although both 360K and 1.2M
drives can read it.
Before formatting a 360K disk in a 1.2M drive, first BULK ERASE the
diskette, then format it and write to it using ONLY a 1.2M drive.
This floppy can be read reliably on a 360K drive. This disk should
be considered READ-ONLY in a 360K drive.
As for formatting floppy disks at densities they weren't designed
for: You might be able to get away with it. But it's not
guaranteed to work. Proceed at your own risk.
4.3) How do I do a low-level reformat of my hard drive?
Depends. Some diagnostics disks have this ability. Compaq Advanced
Diags does, Dell Systems Support disk does. Many controllers can do
it using DEBUG. Wester Digital controllers use G=C800:5, or sometimes
G=CA00:5 or G=CC00:5. PS/2 hard drives can be formatted by booting
the Reference disk and typing Ctrl-A to call up a hidden menu.
4.4) What does interleave mean? What interleave should I use?
If a disk has n:1 interleaving, that means that if you have just read
a sector, then the next higher-numbered sector is n sectors away. For
example, if a track has 11 sectors, then 2:1 interleaving would put
the sectors on a track in the following order 1 7 2 8 3 9 4 10 5 11 6.
The following numbers are rough guides; your actual mileage may vary.
XTs may need 5:1 or higher; ATs typically 3:1; fast ATs and 386s 2:1.
Double buffered controllers allow 1:1 for any processor.
Smaller interleaves are faster, but if you pick an interleave smaller
than your computer can handle, your disk will run pathetically slow.
So if you're going to be wrong, err on the high side. There are
programs on SIMTEL20 that claim to test your interleave factor. The
commercial program SPINRITE claims to calculate the proper interleave
factor and reformat your drive in place (= doesn't destroy the data).
4.5) What do ESDI, SCSI, RLL, MFM mean?
ESDI = Enhanced Small Device Interface.
SCSI = Small Computer Systems Interface.
These are two interface standards which describe how the drive
and the controller communicate.
RLL = Run Length Limited.
MFM = Modified Frequency Modulation.
These are two recording methods which describe how the information
is physically stored on the disk.
4.6) Can I use an RLL controller on my MFM drive?
As is often the case, the answer is "maybe". The likelihood that it
will work is good, but there are some older Seagate ST251s that just
won't. Proceed at your own risk.
4.7) Should I leave my PC turned on all the time? What about
mounting it sideways?
A record of these debates the last time they went around can be
obtained from my email server. Request the file "24hrs".
CHAPTER 5: MISCELLANEOUS
5.1) How do I undelete a file? Recover from an accidental reformat?
The more famous utilities for undeleting files are Paul Mace's
Utilities, Peter Norton's Utilities, and PC-Tools. They also provide
some way of recovering from reformatting your hard disk, although you
have to have run some special program beforehand. (That is, you have
to know in advance that you might accidentally reformat your hard disk
in the future.)
5.2) Where can I get curses? vi? emacs? How do I remove a TSR from memory?
How do I write my own TSR? Is there a program that does X?
These programs and others are available from SIMTEL20:
PD1:<MSDOS.SCREEN>PCCURS14.ARC Unix-compatible version of Curses screen mgmt
PD1:<MSDOS.EDITOR>STEVIE.ARC Clone of Unix 'vi' text editor
PD1:<MSDOS.EDITOR>VI-19A.ARC Unix-like vi screen text editor for IBM PC
PD1:<MSDOS.SYSUTL>TSRCOM29.ARC TSR memory management utils (MARK/RELEASE)
PD1:<MSDOS.TURBO-C>RESLB201.ARC Function library for Turbo-C to make TSRs
Books on TSR programming are listed in the file "books" on my email server.
A more complete vi comes as part of the MKS Toolkit, available
from Mortice Kern Systems. Send email to inquiry@mks.com.
(This is a commercial product.)
There are several different emacs implementations for the PC. You can
get public versions like MicroEmacs and Freemacs or commercial
programs like Epsilon or Unipress Emacs. The similarity to GNU Emacs
varies.
In general, if you are looking for a program, look first in the file
PD1:<MSDOS.FILEDOCS>SIMIBM.ARC. This is a directory of all the files
in SIMTEL20's MS-DOS archives.
You can also send requests to my email server to search through the
index for a word or phrase.
5.3) What about GIF images? How do I view a GIF image?
GIF images are kept on wuarchive.wustl.edu (/graphics/gif),
funic.funet.fi (/disk1/pub/misc/gif) and elsewhere. GIF viewers can
be found at those sites, as well as on SIMTEL20 (PD1:<MSDOS.GIF>).
See the file "gif.sites" available from my mail server for a larger list.
5.4) How do I turn the ECHO off from a batch file?
In DOS 3.3 and higher, you can put the line "@echo off" at the
start of your batch file. For other DOS versions, you'll have
to apply a patch. The file PD1:<MSDOS.BATUTL>ECHOOFF.PAT contains
a list of the patches to apply depending on which version of DOS
you are using.
5.5) How can I set a long path? How can I enlarge my environment?
How do I get/set an environment variable from a C/pascal program?
There are a few ways of setting a long path. One solution is to set
things up so you don't need a long path in the first place. For
example, create a \bat directory with lots of batch files which save
the current path, set the path to something else, run a program, then
reset the path to what it originally was. Or, use the SUBST program
to make drive-letter aliases for long pathnames. If you really want a
big path, fetch the program BIGPATH.ARC or SETPTH13.ARC from SIMTEL20.
The size of the environment is controlled by the "/E" option to
COMMAND.COM. Add the line "SHELL=COMMAND.COM /P /E:xxx" to your
CONFIG.SYS, where xxx is the desired environment size (in either
bytes or paragraphs, depending on your DOS version).
Reading environment variables is easy. Look for a function called
getenv(). Setting environment variables is trickier. If you only
want the variable to be seen in spawned processes, use putenv().
Otherwise, you're going to have to dig into the depths of undocumented
features. It's definitely not easy. Look at the program
PD1:<MSDOS.SYSUTL>SETENVJW.ARC for an example. (There are other
programs on SIMTEL20 that claim to do roughly the same thing as this.)
5.6) How do I install a new version of DOS on my hard drive?
In principle, you should be able to boot off a floppy containing the
new version and type "SYS C:". Then copy COMMAND.COM and all the
other DOS programs (like CHKDSK.COM) to your hard disk.
Unfortunately, this doesn't always work.
Although you could poke around and try to install the new
version manually, it's too complicated to go into. Most people
recommend a procedure similar to the following:
[0] Uninstall any copy-protected software that's on your hard drive.
[1] Back up your hard drive (twice, to be extra sure), making sure to
use a file-by-file backup and not an image backup. Verify your
backup.
[2] Optionally, perform a low-level reformat of the hard disk.
[3] Boot off a floppy containing the new version and install DOS just
like it was a brand new machine.
[4] Restore all the files from your backup and verify the restoration.
[5] If you have a nonstandard SHELL= line in your CONFIG.SYS, make sure
it's set up properly for the new DOS version. (For example, different
versions interpret the /E option differently.)
APPENDIX A: ANSWERS TO MORE QUESTIONS and AVAILABILITY OF THIS FILE
I operate a mail server which not only dispenses files such as this
one, but also looks up programs in the SIMTEL20 index, the grape
index, and the chyde index. (So you have no excuse for not looking
in these indexes before posting a request.) If you send email to
raymond@math.berkeley.edu with no subject and containing
the single line "help", you will receive instructions; sending the
message "directory" gives you an index of all the files
available on the server.
One file in particular you might be interested in is "faq.more"
which contains answers to MORE questions, but which weren't included
here for reasons of space. Send the message "send faq.more" to
obtain it. Its table of contents follows:
1) I found a bug in program X!
2) Why do I get disk full errors when I only have a bunch of tiny files
in my root directory?
3) I found that program X contains the string "Copyright Microsoft", so
it must be a pirated copy of a commercial program, right?
4) Can you summarize the history of the 80x86 family of processors?
5) How about other chips that didn't make on the main faq list?
6) Why do people say that the 80286 is "brain-dead"?
7) How do I determine whether I am running on a monochrome or color
machine?
8) I have written a program that uses the mouse, and when I use
CGA/EGA/VGA it is fine, but the mouse does not appear on
a Hercules card. Why?
9) Why does DOS use the \ instead of a / as a directory name separator?
How can I tell DOS that I want to use slashes instead?
10) What is the SysReq key for?
11) How do I switch my screen into 43 line mode? How can I detect the
number of lines from a program?
12) How do I echo a blank line from a batch file?
13) Why does my PC lose a day if I leave it unattended for more than 24 hours?
14) How can I toggle NumLock/CapsLock/ScrollLock from a batch file or program?
15) I just installed a high density (1.2M) floppy drive, and weird things
are happening. How do I fix it?
16) What is 4DOS? Where can I get it? How can I disable the annoying beeps
that it makes when I run it?