Jump to content

  • Log In with Google      Sign In   
  • Create Account


Getting a weird shader error: error X3018: invalid subscript 'xyz'


Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.

  • You cannot reply to this topic
4 replies to this topic

#1 mrheisenberg   Members   -  Reputation: 356

Like
0Likes
Like

Posted 11 September 2012 - 11:23 AM

I get error X3018: invalid subscript 'xyz' when I try to compile my vertex shader.Here is my shader code:

cbuffer PerFrameConstants : register(b0)
{
	float4x4 worldViewProjectionMatrix;
	float4x4 worldViewMatrix;
	float4x4 viewProjectionMatrix;
	float4x4 projectionMatrix;
	float4 cameraNearFar;
	uint4 framebufferDimensions;
	float tessellationFactor;
	float3 padding;
};

struct GeometryVSIn
{
	float3 position : position;
	float3 normal   : normal;
	float2 texCoord : texCoord;
	float3 instancePosition : instancePosition;
	float3 instanceRotation : instanceRotation;
};
struct GeometryVSOut
{
	float4 position	 : SV_Position;
	float3 positionView : positionView;
	float3 normal	   : normal;
	float2 texCoord	 : texCoord;
};


float4x4 RotateV2M(float3 rotation)
{
float4x4 rotateX = { 1.0f, 0.0f, 0.0f, 0.0f,
							   0.0f, cos(rotation.x), -sin(rotation.x), 0.0f,
							   0.0f, sin(rotation.x), cos(rotation.x), 0.0f,
							   0.0f, 0.0f, 0.0f, 1.0f  };
float4x4 rotateY = { cos(rotation.y), 0.0, sin(rotation.y), 0.0f,
							   0.0f, 1.0f, 0.0f, 0.0f,
							   -sin(rotation.y), 0.0, cos(rotation.y), 0.0f,
							   0.0f, 0.0f, 0.0f, 1.0f   };
float4x4 rotateZ = { cos(rotation.z), -sin(rotation.z), 0.0f, 0.0f,
							   sin(rotation.z), cos(rotation.z), 0.0f, 0.0f,
							   0.0f, 0.0f, 1.0f, 0.0f,
							   0.0f, 0.0f, 0.0f, 1.0f   };
return mul(rotateY, mul(rotateX, rotateZ));
}

GeometryVSOut GeometryVS(GeometryVSIn input)
{
	GeometryVSOut output;
	output.position.x += input.instancePosition.x;
	output.position.y += input.instancePosition.y;
	output.position.z += input.instancePosition.z;
float4x4 rotationMatrix = RotateV2M(input.instanceRotation);
	output.position	 = mul(float4(input.position, 1.0f), mul(worldViewProjectionMatrix, rotationMatrix));
	output.positionView = mul(float4(input.position, 1.0f), mul(worldViewMatrix, rotationMatrix).xyz;
	output.normal	   = mul(float4(input.normal, 0.0f), mul(worldViewMatrix, rotationMatrix)).xyz;
	output.texCoord	 = input.texCoord;
  
	return output;
}





So the problem is in the last function(the main vertex shader one),it somehow has trouble converting from float3 to float4 I think,I'm actualy new to the .xyz functionality,I think it's called swizzling.Maybe I'm not implementing the swizzling properly?

Sponsor:

#2 Nik02   Crossbones+   -  Reputation: 2629

Like
0Likes
Like

Posted 11 September 2012 - 11:29 AM

There is a syntax error in this row:

output.positionView = mul(float4(input.position, 1.0f), mul(worldViewMatrix, rotationMatrix).xyz;

As a hint, the error is not related to the swizzle itself.

Edited by Nik02, 11 September 2012 - 11:31 AM.

Niko Suni


#3 mrheisenberg   Members   -  Reputation: 356

Like
0Likes
Like

Posted 11 September 2012 - 12:16 PM

There is a syntax error in this row:

output.positionView = mul(float4(input.position, 1.0f), mul(worldViewMatrix, rotationMatrix).xyz;

As a hint, the error is not related to the swizzle itself.


Thanks,I can't believe I missed that one,however I think there is still some problem with the actual vertex declaration,or I'm not doing the C++ side right:

UINT shaderFlags = D3D10_SHADER_ENABLE_STRICTNESS | D3D10_SHADER_PACK_MATRIX_ROW_MAJOR;
	    ID3D10Blob *bytecode = 0;
	    HR(D3DX11CompileFromFile(L"GeometryVS.hlsl", defines, 0, "GeometryVS", "vs_5_0", shaderFlags, 0, 0, &bytecode, 0, 0));
	    const D3D11_INPUT_ELEMENT_DESC polygonLayout[] =
	    {
		    {"position",  0, DXGI_FORMAT_R32G32B32_FLOAT, 0, 0,  D3D11_INPUT_PER_VERTEX_DATA, 0},
		    {"normal",    0, DXGI_FORMAT_R32G32B32_FLOAT, 0, 12, D3D11_INPUT_PER_VERTEX_DATA, 0},
		    {"texCoord",  0, DXGI_FORMAT_R32G32_FLOAT,    0, 24, D3D11_INPUT_PER_VERTEX_DATA, 0},
		    {"instancePosition",  0, DXGI_FORMAT_R32G32B32_FLOAT, 0, 0,  D3D11_INPUT_PER_INSTANCE_DATA, 0},
		    {"instanceRotation",    0, DXGI_FORMAT_R32G32B32_FLOAT, 0, 12, D3D11_INPUT_PER_INSTANCE_DATA, 0},
	    };
	   
	    HR(gD3DDevice->CreateInputLayout(polygonLayout, ARRAYSIZE(polygonLayout), bytecode->GetBufferPointer(), bytecode->GetBufferSize(), &vertexLayout));
  gD3DImmediateContext->IASetInputLayout(vertexLayout);
	    SafeRelease(bytecode);

I saw in a tutorial that you define the first 3 ones like this and the ''AlignedByteOffset'' was '0' for the first one, '12' for the second and '24' for the third,even tho the third(texcoord-2 elements) is smaller than the second(position-3 elements) and yet theyre both 12 bytes apart,so I'm confused-should I put the instance ones 12 bytes apart too for it to work with the above shader?

#4 Nik02   Crossbones+   -  Reputation: 2629

Like
0Likes
Like

Posted 11 September 2012 - 12:44 PM

The instance data is (likely) in a separate vertex stream. In the input element descriptions, you still say that they are in the first stream (the InputSlot field of the structure).

The offsets are correct.

Edited by Nik02, 11 September 2012 - 12:49 PM.

Niko Suni


#5 BCullis   Crossbones+   -  Reputation: 1813

Like
0Likes
Like

Posted 11 September 2012 - 02:09 PM

I saw in a tutorial that you define the first 3 ones like this and the ''AlignedByteOffset'' was '0' for the first one, '12' for the second and '24' for the third,even tho the third(texcoord-2 elements) is smaller than the second(position-3 elements) and yet theyre both 12 bytes apart,so I'm confused


Offset is calculated from the end of the previous element, it's not related to the current element you're declaring. If I have a 12 byte element, a 12 byte element, and an 8 byte element in that order:
- the first offset is 0 (since it's first in the list),
- the second is (previous offset ((0)) + length of previous element((12))), so 12,
- the third is (previous offset ((12)) + length of previous element ((12))), so 24.
if there had been a fourth element in the same stream, THEN you would have seen the 8-byte alignment make a difference in an offset value, as the next up would be 32.
Hazard Pay :: FPS/RTS in SharpDX
DeviantArt :: Because right-brain needs love too




Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.



PARTNERS