X-Authentication-Warning:	delorie.com: mail set sender to djgpp-bounces using -f
NNTP-Posting-Date:	Tue, 13 May 2014 21:28:17 -0500
From:	"Charles Sandmann" <cwsdpmi AT earthlink DOT net>
Newsgroups:	comp.os.msdos.djgpp
References:	<op DOT xfpi3cic6zenlw AT localhost> <201405111842 DOT s4BIgrRx012234 AT delorie DOT com> <op DOT xfprrlvs6zenlw AT localhost>
In-Reply-To:	<op.xfprrlvs6zenlw@localhost>
Subject:	Re: secret DJGPP documents?
Date:	Tue, 13 May 2014 21:28:04 -0500
MIME-Version:	1.0
X-Priority:	3
X-MSMail-Priority:	Normal
Importance:	Normal
X-Newsreader:	Microsoft Windows Live Mail 16.4.3528.331
X-MimeOLE:	Produced By Microsoft MimeOLE V16.4.3528.331
Message-ID:	<PKedneUwMOJcSe_OnZ2dnUVZ_u2dnZ2d@earthlink.com>
Lines:	43
X-Usenet-Provider:	http://www.giganews.com
NNTP-Posting-Host:	64.91.136.27
X-Trace:	sv3-IwQ7PRMBOEdSONqs+KRpbKfDoBBcw0iknMOZhNJDst3lAJZfzH9TnWhEaRohbHumjNvQHzERIvDfXyu!EBUvm8eKl9Wz7aYsWvJpYfXJFViKhf5k+kTv5O4THLyvcbtvsuco26rqhGzzTYmNJCVrbh7IDLJG!El3ZPryU0Kw2TzSTO+xxRFWvVdLpQbs=
X-Abuse-and-DMCA-Info:	Please be sure to forward a copy of ALL headers
X-Abuse-and-DMCA-Info:	Otherwise we will be unable to process your complaint properly
X-Postfilter:	1.3.40
Bytes:	3913
X-Original-Bytes:	3852
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp
Reply-To:	djgpp AT delorie DOT com

>"Rod Pemberton"  wrote in message news:op DOT xfprrlvs6zenlw AT localhost... Are 
>you saying no one did any testing?

I did extensive testing with DJGPP and all known DPMI hosts in the mid-90s, 
first with DJGPP 1.x, then later DJGPP 2.x
Somewhere on an ancient CD copy of an ancient 486 drive's content, are all 
the results of what was supported, what versions were claimed, etc.
This included many versions of QEMM, 386MAX, OS/2 (2.0 to 4.x), Borland 
DPMI, Windows 3.0, 3.1, NT 3.1, 3.5, 3.51, 4.0 (several service packs), Win 
9x, etc.
I also did several releases of CWSDPMI and PMODE.  There were some private 
versions, like a CWSDPMI version with hooks to support the V1.x graphics 
linear memory (with demand/page).
But in the end, everyone wrote to the most common environments, which were 
Windows 9.x and CWSDPMI (and later Windows 2000/XP).
I don't think anyone cares how Windows 3.0 differed from NT 3.1 anymore, 
much less the bugs, features and memory footprints of various QEMM or 386MAX 
releases.
Little details like images running OK, but not nesting properly (especially 
when mixing DJGPP 1.x and 2.x images), or spawning images built with 
extenders or using different DPMI calls (like pkzip).   Getting DOOM to run 
under DPMI.  Figuring out why service packs broke DPMI stack usage.  Trying 
to make sure CTRL-C in DJGPP didn't reboot computers with bad DPMI 
implementations.  Ring 0 implementations for special users.

I remember some of this being in a Lotus/123 .wk1 file someplace. 
Interesting exceptions were shared with the various mailing lists, and 
details shared with people also working on it.
There were a set of test cases written and run to call and test each API, 
how systems behaved at the extremes (how many selectors could you allocate? 
What happened when you allocated the memory DPMI claimed was available? 
What happened when you hooked all the hardware interrupts (especially timer, 
disk, keyboard), then allocated all the memory and randomly touched all the 
pages while recording nested HW interrupts, with a book on the keyboard with 
30cps autorepeat?  For a solid weekend?  Morten Welinder did lots of test 
development, and was working on the replacement for CWSDPMI to be built with 
all FSF tools (he needed full test cases to verify functionality).

So no secret documents, but I don't think I have looked at (or for) that 
type of documentation for at least 17 years or so.   About 3 people had any 
interest beyond "does it work?"

- Raw text -