🚀 go-pugleaf

In article <1992Oct30.025344.9820@netcom.com>, geo@netcom.com (George Pontis) writes:
> rwh@moose.cccs.umn.edu (RICHARD HOFFBECK) writes:
>  
>  
> >Do yourself a favor and drop 386MAX into the nearest trash can.  QEMM is a
> >bargin at twice its discounted price and of course himem/emm is included
> >with DOS and Windows.  Compiling in a Windows DOS box isn't even too bad
>  
> The bad experiences this individual has had are not typical. 

Both Microsoft and Qualitas acknowledge that 386MAX as shipped will not
work with the virtual memory manager included in the C 7.0 library. 
Qualitas as posted a series of patches on Compuserve intended to fix
the problem, but to date the problem persists.

>386Max can,
> and does work properly in plenty of systems, including several that I use
> with C/C++ 7.0. 

Really, try this quick code segment:

	_vmhnd_t memhnd ;
	for( i=0 ; i < 40 ; i++)
	   memhnd = vmalloc( 20000) ;

Now try and run it under 386MAX.  If the vm handler is initialized to use
XMS memory, you get a hung machine.  If you use EMS or disk based vm, it
will usually hang when you exit the program.  If you have the misfortune
of using EMS based virtual memory and you are using smartdrv, you'll find
that the disk cache has been corrupted.
 	
>( Did you know that with a monochrome card, 386Max can give
> you more than 700K free low memory, after installing disk caches, shells,
> and other options ? ) 

So does QEMM and I suspect will most reasonable third party memory managers.
Probably more importantly, QEMMs Stealth mode lets me get all of my network
drivers in high memory and still leave 626K free on a VGA system.  The best
386MAX could do was around 512K. Granted Pathworks uses a LOT of memory for
drivers.

> You may be interested to know that PC tested
> several managers, including 386Max and QEMM. The essence of their rating was
> that QEMM had some potential of recovering slightly more high DOS memory than
> did 386Max, but that 386Max was the most dependable of all the products.

I read the PC review and was looking forward to having a chance to look at
386MAX; but remember, the PC Magazine review was prior to the widespread use
of DMPI. I know it doesn't work well on my machine, and judging from the
traffic on the Windows/C++/etc group here and the questions being fielded on
Compuserve by the Microsoft support people, I am not the only one that had
a problem.  

In the mean time, I don't use it and our beta testers have been advised that
we will not support 386MAX until we have a version that actually works.  Its
hard enough to find the bugs that I introduce without having to constantly
questioning the supporting tools.

--rick

+-----------------------------------------------------------------+
| Richard Hoffbeck               INTERNET: rwh@moose.cccs.umn.edu |
| Colon Cancer Control Study     ATT     : (612) 627-4151         |
| University of Minnesota        CIS     : 72406,521              |
| Minneapolis, MN 55455                                           |
+-----------------------------------------------------------------+