«

On Oct 18, 7:34*am, SG <s.gesem...@gmail.com> wrote:
> On 18 Okt., 01:55, Paul wrote:
>
> > Ian Collins wrote in message
> > > [...] ramtin wrote:
>
> > >> int main(array<System::String ^> *^args)
>
> > > This looks like a language almost, but not entirely, unlike C++.
>
> > That is the syntax use din MSVC++ when you create a windows form
> > projec(C++), probably a .net thing.
>
> I would guess that a large fraction of the competent regulars
> (including Ian) already know exactly what it is: C++/CLI. And yes,
> that's what you get if you ask Herb Sutter to make .NET easily
> accessible from C++ by use of language extensions.
>
> Off-topic: I makes me chuckle a bit to see almost the same syntax
> being used for C++/CX where the CX language extensions are used to
> deal with WinRT's ref-counting object system. I guess Microsoft lost a
> bit of interest in .NET and C++/CLI since this seems to make the CLI
> and CX extensions mutually exclusive. Btw, I can recommend Herb
> Sutter's "modern C++" talk. It almost sounds like he's trying to
> convert .NET developers. "Look, how easy it is to write modern and
> safe C++ code!".
>
> Cheers!
> SG

I never really understood what advantage there would be in using C++/
CLI. If I want CLI, why wouldn't I just use C# instead? Maybe so I can
have pointers? OT anyway I suppose.

Op 27-Oct-11 23:29, Christopher schreef:
> On Oct 18, 7:34 am, SG<s.gesem...@gmail.com> wrote:
>> On 18 Okt., 01:55, Paul wrote:
>>
>>> Ian Collins wrote in message
>>>> [...] ramtin wrote:
>>
>>>>> int main(array<System::String ^> ^args)
>>
>>>> This looks like a language almost, but not entirely, unlike C++.
>>
>>> That is the syntax use din MSVC++ when you create a windows form
>>> projec(C++), probably a .net thing.
>>
>> I would guess that a large fraction of the competent regulars
>> (including Ian) already know exactly what it is: C++/CLI. And yes,
>> that's what you get if you ask Herb Sutter to make .NET easily
>> accessible from C++ by use of language extensions.
>>
>> Off-topic: I makes me chuckle a bit to see almost the same syntax
>> being used for C++/CX where the CX language extensions are used to
>> deal with WinRT's ref-counting object system. I guess Microsoft lost a
>> bit of interest in .NET and C++/CLI since this seems to make the CLI
>> and CX extensions mutually exclusive. Btw, I can recommend Herb
>> Sutter's "modern C++" talk. It almost sounds like he's trying to
>> convert .NET developers. "Look, how easy it is to write modern and
>> safe C++ code!".
>>
>> Cheers!
>> SG
>
> I never really understood what advantage there would be in using C++/
> CLI. If I want CLI, why wouldn't I just use C# instead?

+1

float d1= 4/3 ; /* promote to inetger only in the right hand side result */

float d2= 4.0/3 ; /* promote to float */
float d3= 4/3.0 ;
float d4= 4.0/3.0;

On 10/28/2011 9:06 PM, 88888 Dihedral wrote:
> float d1= 4/3 ; /* promote to inetger only in the right hand side result */
No, that's integer arithmetic, result promoted to float.

>
> float d2= 4.0/3 ; /* promote to float */
> float d3= 4/3.0 ;
> float d4= 4.0/3.0;

Technically, the first two are promote to double, truncate to float.
The third is already double, truncate to float.

red floydæ–¼ 2011å¹´10月29日星期å…*UTC+8下å�ˆ1時14分07ç§ ’寫é�“：
> On 10/28/2011 9:06 PM, 88888 Dihedral wrote:
> > float d1= 4/3 ; /* promote to inetger only in the right hand side result */
> No, that's integer arithmetic, result promoted to float.
>
> >
> > float d2= 4.0/3 ; /* promote to float */
> > float d3= 4/3.0 ;
> > float d4= 4.0/3.0;
>
> Technically, the first two are promote to double, truncate to float.
> The third is already double, truncate to float.
------------------------------------------------------
int x=3;
int y=4;

float d = x/y ;
float d= ((float)x)/y ;

/* whether the compiler knows x,y are constants up to now ? */

On 10/28/2011 11:27 PM, 88888 Dihedral wrote:
>
>
> red floydæ–¼ 2011å¹´10月29日星期å…*UTC+8下å�ˆ1時14分07ç§ ’寫é�“：
>> On 10/28/2011 9:06 PM, 88888 Dihedral wrote:
>>> float d1= 4/3 ; /* promote to inetger only in the right hand side result */
>> No, that's integer arithmetic, result promoted to float.
>>
>>>
>>> float d2= 4.0/3 ; /* promote to float */
>>> float d3= 4/3.0 ;
>>> float d4= 4.0/3.0;
>>
>> Technically, the first two are promote to double, truncate to float.
>> The third is already double, truncate to float.
> ------------------------------------------------------
> int x=3;
> int y=4;
>
> float d = x/y ;
> float d= ((float)x)/y ;
>
> /* whether the compiler knows x,y are constants up to now ? */
>

What's your point? 4.0 and 3.0 are *DOUBLE* constants.

Is there any compelling reason to use float instead of double ? If
not, use double.

On 10/30/11 09:40 PM, ptyxs wrote:
> Is there any compelling reason to use float instead of double ? If
> not, use double.

Is there any compelling reason not to include any context in your post?

--
Ian Collins