🚀 go-pugleaf

jwbirdsa@amc.com (James Birdsall) writes:

>In article <1991Jun13.231255.7781@proto.com>, doug@proto.com (Doug Huffman) writes:
>>Has anyone used INTEL386 C?  It is supposed to make 386 executables
>>that don't need a DOS extender!  It apparently wraps its own

James, you make it look like you were quoteing me directly... I was quoteing
someone else...

>>The DOS extender is DPMI
>>compliant which means that it runs just fine under a Windows DOS box
>>when you're in 386 enhanced mode.

>   Unless you're trying to do floating point. Windows presently blocks
>the interrupt their floating-point emulator needs to see in order to
>function. If you have a coprocessor, it should work OK though.

Zortech's DOS extender (which I wrote) doesn't have this problem.  In fact
most of the development was done on a machine without a coprocessor and 
all the floating point code (as well as nearly all the Zortech C library, 
extensions and all) work just fine with all combinations of Windows modes,
coprocessor, VCPI, etc, etc.  The design philosophy was: "it should just 
Work".  By the way... the floating-point interrupt isn't the only bug-a-boo 
in Windows, in my more candor moments I express words (not fit for polite 
company) to the effect that Windows 3.0 isn't compliant with the DPMI spec.

>   Note that it isn't compatible with QEMM, 386^Max, and other such
>VCPI compliant software. However, it is supposed to work with QEMM/386^Max
>if you're also running Windows at the same time.

I believe Zortech will be the first to ship a DOS extender that is DOS, VCPI,
and DMPI compliant (correct me if I'm wrong).

doug@proto.com
uunet!proto!doug