[geos-devel] Building latest CVS HEAD with MSVC++ 8.0

[Stefan Zschocke]

Hi, 
seems I am missing something.
I have been experimenting with the geos-2.2.1.tar and I don't find round
anywhere in the sources.
Can you point me to the code where this round problem arises?
I have checked another c++ project of mine and there I do have a custom
round function, I don't remember why, but certainly it is because of the
same issue mentioned in this thread.
I use the following code:

long double roundl(long double d,short dec);  
double round(double d,short dec);
#include <math.h>
#include <float.h>
inline long double roundl(long double d,short dec){ 
  long double div=powl(10,dec);
  d*=div;
  d*=(1.+LDBL_EPSILON); 
  __asm{ 
    fld d 
    frndint
    fstp d
  }
  return d/div; 
}  

inline double round(double d,short dec){ 
  double div=pow(10.,dec);
  d*=div;
  d*=(1.+DBL_EPSILON); 
  __asm{ 
    fld d 
    frndint
    fstp d
  }
  return d/div; 
}  


Stefan

-----Original Message-----
From: geos-devel-bounces at geos.refractions.net
[mailto:geos-devel-bounces at geos.refractions.net] On Behalf Of Charlie
Savage
Sent: Thursday, February 23, 2006 1:02 AM
To: GEOS Development List
Subject: Re: [geos-devel] Building latest CVS HEAD with MSVC++ 8.0


Hi Matuesz,

I was just offering to donate the Visual Studio C++ Solution files 
(i.e., the VC++ project files) that I have to GEOS.

That is a separate thing from fixing the round function, which stops 
VC++ from compiling GEOS whether you use Visual Studio, the Visual
Studio command line or nmake.

However, I gather the VC++ community doesn't use GEOS much though, since

there aren't any Visual Studio files or build scripts that GEOS provides

that work with VC++.

Charlie

Mateusz Łoskot wrote:
> Charlie Savage wrote:
>> One thing to follow up on.  I realize there used to be VC++ solution

>> in the GEOS source tree.  However, it wasn't maintained and thus was 
>> removed a few months ago.
>>
>> Is there any interest in having a working VC++ solution again?  I 
>> don't mind keeping it up to date since I do that locally anyways.
>
> I don't understand. Are you talking about round function issue?
> Does it mean you are asking if it's worth to keep GEOS compilable
> with VC++ ?
> If I'm following your point, should I understand that GEOS is 
> (usually) not compiled by the GEOS Community with VC++ at all?
>
> Cheers