🚀 go-pugleaf

In article <2078@tellab5.TELLABS.COM> fayne@tellab5.TELLABS.COM (Jeffrey Fayne) writes:
>I seriously considering purchasing NuMega's 386 debugger, however
>I have a few questions for anyone who has used this program:
>
>	1) Is it as powerful as the ads make it out to be ?
>	   I'm particularly interested in it's 'back trace'
>	   capability

Yes on both counts.  It's not just a back trace, it gives you back trace
simulation, meaning you can see what the registers were at a given
instruction.  The back trace makes the program run slower, which may
preclude its use with some real-time systems.

>	2) I understand the program is codeview compatable, 
>	   therefore it should be TurboDebugger compatible
>	   as well. True or false ?

You can probably use it with just about any debugger that is out there.
I prefer to use it standalone.  It can set breakpoints that ordinary
debuggers can't, like break if code in this area writes to memory in
that area.  You can tell it what you want it to do when it gets one of
these breakpoints.  If you tell it to do an int 3, which most debuggers
use for breakpoints, you can probably get it to work with just about
anything.  But you can do symbolic debugging without using anything
except Soft-Ice.

>	3) Talking to a sales type, he claims
>          [salesman's claims deleted @;-) ]
>	   any recommendations as
>	   far as switching linkers?  I'm very happy with PLINK, but
>	   if I find SoftIce 'invaluable', I may switch.

I've never had any reason to switch to a linker that I had to pay
extra for.  But whatever reason you have, I don't see how it could
convince you to live without Soft-Ice.  I won't.

>	4) Ideally, I would eventually like to run SoftIce under
>	   VM/386 (multitasker), is this possible? If not why not
>	   and will it run under any other multitasker?

I seriously doubt this is possible.  With any other multitasker, I'm
quite sure of it.  Soft-Ice gets its power from using the privileged
modes of the 386, taking advantage of all kinds of processor exceptions
caused by certain kinds of instructions and the memory protection
scheme.  This would conflict with any other program (i.e. OS) that
does the same.  Theoretically, VM is an operating system that can run
other operating systems by providing them with the illusion of a
virtual machine.  As implemented on the old IBM mainframes, any OS that
could run on the bare machine could be fooled into running under VM.
As to whether this principle has been successfully transferred to
the VM/386 implementation, I don't know, but I'd like to hear from
anyone who knows for sure.  I went to an ACM convention once where
a couple of guys gave a talk about why it was impossible to do on
the 386, and this may be one of the main reasons for the 486.

But hell, I've already told you more than I know.  @;-)

I'm not associated with Nu-Mega.  I think their name is silly, but I
wouldn't be without Soft-Ice.  I have used it in debugging a multitasking
kernel (real mode only) and I swear by it.  But standard disclaimers
still apply...