Mail Archives: djgpp/2001/06/28/15:43:41
> From: invalid AT erehwon DOT invalid (Graaagh the Mighty)
> Newsgroups: comp.os.msdos.djgpp
> Date: Thu, 28 Jun 2001 15:14:43 GMT
>
> >No. The DPMI host always gets the first chance to look at any
> >exceptional condition.
>
> That makes it responsible for giving the best data it can in every
> situation then.
Yes; and CWSDPMI did just that: you've got the EIP where your program
crashed. That's a lot.
> >Any number of things. It could simply load some value into SS in
> >inline assembly. Or call int86x with wrong arguments. Or installe a
> >real-mode callback with wrong parameters.
>
> My code does nothing of the sort as written.
I didn't say it did; I've just shown a few examples to demonstrate how
easy it is.
> All I want is the damned traceback, or
> other useful debugging information. All I want is that if I compile
> and run protected-mode code, it might produce bogus output, and it
> might even bomb with an application fault, but it won't reboot the
> machine, hang the OS, or do anything else outside of its PM sandbox.
> Is that really too much to ask? Sheesh.
It turns out that in your case, it _is_ too much to ask.
- Raw text -