• rumbu (26/26) Dec 09 2017 float f = float.min_normal;
• Iain Buclaw (3/29) Dec 09 2017 Raise a bug if you are unsure.

rumbu <rumbu rumbu.ro> writes:

     float f = float.min_normal;
     bool fcast1 = cast(bool)f;
     bool fcast2 = cast(bool)float.min_normal;

     if (fcast1)
         writeln("variable casting to bool is true");
     else
         writeln("variable casting to bool is false");

     if (fcast2)
         writeln("constant casting to bool is true");
     else
         writeln("constant casting to bool is false");

     if (float.min_normal)
         writeln("implicit conversion is true");
     else
         writeln("implicit conversion is false");


variable casting to bool is true
constant casting to bool is false
implicit conversion is true

Now, casting any float constant in the interval (-1.0;1.0) to 
bool will return false, but the same value considered as an if 
condition or variable will always return true;

Looking at disassembly, fcast2 is directly initialized to false, 
the compiler does not emit any conversion code:

bool fcast2 = cast(bool)float.min_normal;
004174fa  mov byte [ebp-0x3], 0x0

Is this intended behavior or it's a bug?

Iain Buclaw <ibuclaw gdcproject.org> writes:

On 9 December 2017 at 12:51, rumbu via Digitalmars-d
<digitalmars-d puremagic.com> wrote:
     float f = float.min_normal;
     bool fcast1 = cast(bool)f;
     bool fcast2 = cast(bool)float.min_normal;

     if (fcast1)
         writeln("variable casting to bool is true");
     else
         writeln("variable casting to bool is false");

     if (fcast2)
         writeln("constant casting to bool is true");
     else
         writeln("constant casting to bool is false");

     if (float.min_normal)
         writeln("implicit conversion is true");
     else
         writeln("implicit conversion is false");


 variable casting to bool is true
 constant casting to bool is false
 implicit conversion is true

 Now, casting any float constant in the interval (-1.0;1.0) to bool will
 return false, but the same value considered as an if condition or variable
 will always return true;

 Looking at disassembly, fcast2 is directly initialized to false, the
 compiler does not emit any conversion code:

 bool fcast2 = cast(bool)float.min_normal;
 004174fa  mov byte [ebp-0x3], 0x0

 Is this intended behavior or it's a bug?

Raise a bug if you are unsure.