div.c and other questions

Michel Bardiaux mbardiaux at peaktime.be
Mon Sep 27 17:17:50 CEST 2004

Brian Hurt wrote:

> On Mon, 27 Sep 2004, Patrick Pelissier wrote:
>>>The problem is that MSVC implemented the 'near' keyword as a modifier to 
>>>pointers, like 'const'.  So you could have:
>>>    mp_limb_t near * foo_p;
>>>(foo_p is a near pointer to an mp_limb_t).  
>>>I agree that this is a non-conformance to the standard.  But this is also 
>>>the company that gave us WinMain().
>> Aren't there any flags/options to remove this checking?
> I don't think so, but it's been quite some time since I've played with 
> MSVC++.
> Hopefully, someone who has access to the documentation will be more
> specific.

I've looked in the doc for MV-VisualStudio.NET aka MSVC++7.0. The 
*compiler* has no knowledge of "near" and "far" (concept gone since 
MSVC++1.52 actually!) They are hash-defined to empty string by windef.h, 
and no header file shipped by MS uses near or far - *except* windef.h 
itself! Beurk.

Anyway, I see only 2 possibilities:

(1) Consider "near" and "far" as de facto reserved words.

(2) Replace the windef.h by one that makes aboslutely no mention of near 
of far.

Michel Bardiaux
Peaktime Belgium S.A.  Bd. du Souverain, 191  B-1160 Bruxelles
Tel : +32 2 790.29.41

More information about the gmp-discuss mailing list