From: Charles Sandmann <sandmann AT praline DOT no DOT NeoSoft DOT com>
Subject: Re: Inhibit system() swapout??
To: babcock AT cfa DOT harvard DOT edu
Date: Thu, 9 Mar 1995 20:24:16 -0600 (CST)
Cc: djgpp AT sun DOT soe DOT clarkson DOT edu

> I think, when the question of avoiding GO32's desire to page out everything
> came up earlier, DJ said that GO32 would intercept attempts to sneak around
> it and do the paging anyway.  

There are ways around it.  They may not be obvious or easy.

> This was a few releases back, and I never
> actually tried it.  I'd love to have a GO32 that allowed a system() call
> under both DOS and DPMI without paging everything out.  (That hasn't changed,
> has it?  Last time I checked, DPMI always paged out, non-DPMI depended on
> stubedit settings.)

Well, fixing GO32 to not page out on DPMI is a simple fix, it just never
got patched into a release (I would rather work on V2 than code which will
be discarded).  I have a version of GO32 that avoids this - it is actually
just a matter of deleting a bunch of code.  It probably ought to be that
way by default, but live and learn.  I know a lot more about DPMI now than
when I DPMI'ed GO32 2 years ago.  V2 does not swap out.