www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.learn - Invalid Floating Point Operation (DMD 2067)

reply "ref2401" <refactor24 gmail.com> writes:
Hi Everyone,

After I switched to DMD 2067 my code that previously worked began 
crashing.
If I change the return statement in the matrixOrtho function

from:
	return Matrix(...);
to:
	Matrix m = Matrix(...);
	return m;

then the error won't occur anymore. What is going on?

Thank you.

import std.math;
import std.stdio;

struct Matrix {
	float m00, m01, m02, m03;
	float m10, m11, m12, m13;
	float m20, m21, m22, m23;
	float m30, m31, m32, m33;

	this(float m00, float m01, float m02, float m03,
		 float m10, float m11, float m12, float m13,
		 float m20, float m21, float m22, float m23,
		 float m30, float m31, float m32, float m33) nothrow pure  nogc 
{

		 this.m00 = m00; this.m01 = m01; this.m02 = m02; this.m03 = m03;
		 this.m10 = m10; this.m11 = m11; this.m12 = m12; this.m13 = m13;
		 this.m20 = m20; this.m21 = m21; this.m22 = m22; this.m23 = m23;
		 this.m30 = m30; this.m31 = m31; this.m32 = m32; this.m33 = m33;
	 }
}

void main(string[] args) {

	float width = 800f;
	float height = 600f;
	float near = -5f;
	float far = 10f;

	float hw = width / 2f;
	float hh = height / 2f;

	Matrix ortho = matrixOrtho(-hw, hw, -hh, hh, near, far);
}

Matrix matrixOrtho(float left, float right, float bottom, float 
top, float near, float far) /* pure  nogc */ {
	debug {
		FloatingPointControl fpCtrl;
		fpCtrl.enableExceptions(FloatingPointControl.severeExceptions);
	}

	float doubledNear = near*2f;
	float farMinusNear = far - near;
	float rightMinusLeft = right - left;
	float topMinusBottom = top - bottom;
	

	return Matrix(2f / rightMinusLeft, 0f, 0f, -(right + 
left)/rightMinusLeft,
		0f, 2f / topMinusBottom, 0f, -(top + bottom)/topMinusBottom,
		0f, 0f, -2f / farMinusNear, -(far + near)/farMinusNear,
		0f, 0f, 0f, 1f);

	//Matrix m = Matrix(2f / rightMinusLeft, 0f, 0f, -(right + 
left)/rightMinusLeft,
	//	0f, 2f / topMinusBottom, 0f, -(top + bottom)/topMinusBottom,
	//	0f, 0f, -2f / farMinusNear, -(far + near)/farMinusNear,
	//	0f, 0f, 0f, 1f);

	//return m;
}


Error report:

object.Error (0): Invalid Floating Point Operation
----------------
0x00402340
0x004020E2
0x004029E2
0x004029B7
0x004028CF
0x004022D7
0x758D7C04 in BaseThreadInitThunk
0x77ADB54F in RtlInitializeExceptionChain
0x77ADB51A in RtlInitializeExceptionChain
object.Error (0): Invalid Floating Point Operation
----------------
0x00402340
0x004020E2
0x004029E2
0x004029B7
0x004028CF
0x004022D7
0x758D7C04 in BaseThreadInitThunk
0x77ADB54F in RtlInitializeExceptionChain
0x77ADB51A in RtlInitializeExceptionChain
object.Error (0): Invalid Floating Point Operation
----------------
0x00406E5B
0x004029C5
0x004028CF
0x004022D7
0x758D7C04 in BaseThreadInitThunk
0x77ADB54F in RtlInitializeExceptionChain
0x77ADB51A in RtlInitializeExceptionChain
Apr 16 2015
parent reply Steven Schveighoffer <schveiguy yahoo.com> writes:
On 4/16/15 10:23 AM, ref2401 wrote:

 Hi Everyone,

 After I switched to DMD 2067 my code that previously worked began crashing.
 If I change the return statement in the matrixOrtho function

 from:
      return Matrix(...);
 to:
      Matrix m = Matrix(...);
      return m;

 then the error won't occur anymore. What is going on?


Builds and runs fine for me. What is your OS and build command?

-Steve
Apr 16 2015
parent reply "ref2401" <refactor24 gmail.com> writes:
 Builds and runs fine for me. What is your OS and build command?

 -Steve


Win 8.1

dmd main.d -ofmain.exe -debug -unittest -wi
if %errorLevel% equ 0 (main.exe)
Apr 16 2015
parent reply "rumbu" <rumbu rumbu.ro> writes:
On Thursday, 16 April 2015 at 15:00:39 UTC, ref2401 wrote:

 Builds and runs fine for me. What is your OS and build command?

 -Steve


 Win 8.1

 dmd main.d -ofmain.exe -debug -unittest -wi
 if %errorLevel% equ 0 (main.exe)


Reduced case:

struct S
{
    float f;

}
Apr 16 2015
parent reply "rumbu" <rumbu rumbu.ro> writes:
On Thursday, 16 April 2015 at 20:27:46 UTC, rumbu wrote:
Hit send by accident :)


Reduced case:

struct S {
     float f; //or double or real
     this(float f) {
        this.f = f;
     }
}

S foo() {
    return S(0f);
}

void main() {
   auto s = foo();
}

Win 8.1 also, 32 and 64 bit, debug or release, same exception
Apr 16 2015
parent reply Steven Schveighoffer <schveiguy yahoo.com> writes:
On 4/16/15 4:32 PM, rumbu wrote:

 On Thursday, 16 April 2015 at 20:27:46 UTC, rumbu wrote:
 Hit send by accident :)


 Reduced case:

 struct S {
      float f; //or double or real
      this(float f) {
         this.f = f;
      }
 }

 S foo() {
     return S(0f);
 }

 void main() {
    auto s = foo();
 }

 Win 8.1 also, 32 and 64 bit, debug or release, same exception


Thanks, please file a bug, make it Windows specific, as I did not get 
the exception on OSX.

-Steve
Apr 16 2015
parent "rumbu" <rumbu rumbu.ro> writes:
 Thanks, please file a bug, make it Windows specific, as I did 
 not get the exception on OSX.

 -Steve



Done: https://issues.dlang.org/show_bug.cgi?id=14452
Apr 16 2015