Mail Archives: djgpp/2000/07/19/01:08:05
Okay, I found my floppy and played with it a bit. Error depends on what's
in the environment, such as the length of the path. Different versions of
CWSDPMI behave differently. Just rearranging the order of the environment
with exact same environment variable contents causes different failures.
An exact copy of the exact same files will fail differently when run from
the hard drive instead of the floppy.
Looks like an uninitialized variable in either the PMODE loader or in CWSDPMI.
It may take a while to find this one... Since most configurations either
end up in a hard hang or instant reboot ...
I may have to remember how to "debug" into CWSDPMI to understand what got
scrogged and maybe a hint of how it happened... Unfortunately this changes
the error behavior too :-(
> I think I sent you all the relevant information back in November, and
> you were eventually able to reproduce these problems on your machine.
> From: sandmann AT clio DOT rice DOT edu (Charles Sandmann)
> Subject: Re: Read Me Third: A Short FAQ List (fwd)
> To: eliz AT gnu DOT org (Eli Zaretskii)
> Date: Mon, 15 Nov 1999 19:29:39 -0600 (CST)
>
> I can reproduce your error. After the error it is very interesting.
> I see CWSDPMI loaded in memory, DPMI appears to be available,
> cwsdpmi -u executes fine, but any djgpp image indicates there is no
> DPMI memory.
>
> So, it still thinks it's nested inside CWSDPMI, and no DPMI memory is
> available. Fascinating.
- Raw text -