[comp.windows.ms] Windows 3.0 Protected mode problem

bg11+@andrew.cmu.edu (Brian E. Gallew) (06/06/90)

I seem to have a problem.  I have a 25MHz clone with an AMI BIOS, 2MG of
RAM, an 84MB SCSI drive, a Paelit VGA video card, a Basic Time modem,
and an IO card.  When I try to run windows in 386 enhanced mode, The
system always dies soon after the program manager loads.  Sometimes it
dies immediately, sometimes I can get other apps open and start using
them before they die.  Once I managed to exit before it quit.  In all
cases, it gave me an Invalid COMMAND.COM message.  This happens under
DOS 3.3.

Any help would be _greatly_ appreciated!

gpsteffl@sunee.waterloo.edu (Glenn Patrick Steffler (Co-Op Scum)) (06/07/90)

In article <YaP1NHK00Uh701wWxf@andrew.cmu.edu> bg11+@andrew.cmu.edu (Brian E. Gallew) writes:
>system always dies soon after the program manager loads.  Sometimes it
>dies immediately, sometimes I can get other apps open and start using
>them before they die.  Once I managed to exit before it quit.  In all
>cases, it gave me an Invalid COMMAND.COM message.  This happens under
>DOS 3.3.
>
>Any help would be _greatly_ appreciated!

That message does in fact mean that you have an invalid command.com (no kidding).

Search your hard disk for multiple copies of command.com
If there are multiple copies, deleted all but the one refered to
in the COMSPEC, and SHELL variables.  Make sure you are running
DOS 3.x because thats what the windows developer uses.

I hope this helps, if not, I'm not liable for damages :-)


-- 
  ########--########	Co-Op Scum	Red Heads Do It In The Shade
   #######/\#######
   #######  #######	Glenn Patrick Steffler		Comp. Eng.  U of 'Loo
    #####    #####	gpsteffl@sunee.waterloo.edu	gpsteffl@sunee.uwaterloo.ca

davidr@hplsla.HP.COM (David M. Reed) (06/12/90)

I have seen this error (Invalid COMMAND.COM; Halted) many times.  In fact,
I am trying to trace down why it is happening periodically on 3 different
systems here (one an 8 MHz '286, one a 20MHz '386, and a 25MHz '386), each
running different main applications (GRAPHICS GALLERY on one, VENTURA
PUBLISHER on another, and a custom application on the 3rd).  

COMMAND.COM has a transient portion that occasionally gets overwritten
(usually by a large application), and that is why there is a COMSPEC
variable, so that COMMAND.COM, upon determination that its transient
portion is corrupt (= Invalid COMMAND.COM) can reload itself.  However,
if the COMSPEC variable disappears, for is in any way unable to reference
the real COMMAND.COM, then the system HALTS (it is unable to continue).

I have been unable to determine exactly what is happening to the COMSPEC
variable (or something related to it, such as the code of COMMAND.COM that
stays resident to reload itself) so that it can not find and reload its
transient code.  In other words, I generally know what is probably happening,
and why it is probably happening, but am unable to determin the other
details in order to fix it.  Does anyone else have a method for trapping
this?

markg@cbnewsk.att.com (mark.r.gibaldi) (06/12/90)

In article <3130004@hplsla.HP.COM> davidr@hplsla.HP.COM (David M. Reed) writes:
>I have seen this error (Invalid COMMAND.COM; Halted) many times.  In fact,
>I am trying to trace down why it is happening periodically on 3 different
>systems here (one an 8 MHz '286, one a 20MHz '386, and a 25MHz '386), each
>running different main applications (GRAPHICS GALLERY on one, VENTURA
>PUBLISHER on another, and a custom application on the 3rd).  
>

I saw this problem when I accidentally erased autoexec.bat and config.sys on
my Win 3.0 equipped machine.  The problem came up when I had replaced my
system files with slightly different versions.  I made three changes and
the problem has not arisen since.  I may have just been lucky so far, but
in case this might help, hear are the three changes.

1) made command.com permanent in my Config.sys "SHELL" with "/p".
2) placed my command.com in the root directory instead of "c:\dos".
3) Set my COMPSPEC environment variable.

Hope this helps.

Mark R. Gibaldi
AT&T Bell Laboratories
mrg@cblph.att.com