Shader compile errror: "Fragment program 'frag_surf': non-4 sized matrix"?

I'm currently trying to find a bug in a shader. The error message is:

Code (csharp):

1.  
2. Shader error in 'Hidden/TestBug': Fragment program 'frag_surf': non-4 sized matrix M at line 10
3.  

As this error occurs in a larger shader, I tried to extract the problem and wrote a dummy shader which demonstrates the problem. Note that this shader does not do anything useful, it's just here to demonstrate the error.

The code below results in the error above. If I comment out the line producing the error and uncomment the other lines, the shader compiles. In both cases i'm using a float scalar - so the problem is definitively produced by the matrix operation.

Code (csharp):

1.  
2. Shader "Hidden/TestBug" {
3.     Properties {
4.         _MainTex ("Texture", 2D) = "white" {}
5.     }
6.  
7.     SubShader {
8.         Tags { "RenderType" = "Transparent"  "Queue"="Transparent"}
9.  
10.         CGPROGRAM
11.         #pragma surface surf Lambert
12.         #pragma target 3.0
13.  
14.         struct Input { float2 uv_MainTex; };
15.         sampler2D _MainTex;
16.         const float3x3 M = { 1,2,3,   4,5,6,   7,8,9  };
17.      
18.         void surf (Input IN, inout SurfaceOutput o)
19.         {
20.             float3x3 I;
21.             float r = dot(M[0], I[0]);
22.  
23.             // the scalar resulting from the matrix dot product throws an error... why?
24.             // "Fragment program 'frag_surf': non-4 sized matrix M"
25.             o.Albedo = float4(r,0,0,0);
26.  
27.             // using another scalar (which does not result from a matrix operation) works
28.                         // uncommenting the following two lines and commenting out the above one,
29.                         // makes the shader compile
30.             // float s = 0.4;
31.             // o.Albedo = float4(s,0,0,0);
32.          }
33.          ENDCG
34.     }
35.  
36.     Fallback "Diffuse"
37. }
38.  

Does anybody know what's wrong here?

 

Matrixes need to be 4x4, as far as I know.

If you need to multiply something as if it were a 3x3 you'd use

Code (csharp):

1. mul((float3x3)matrix, value).xyz;

So you can just fill the other entries in the matrix with 0 and it'll work the same.

 

Hi, when are you going to support float3x3 as a uniform variable? I use these each day in my work for image processing constants, it would keep my code much tidier and less work to maintain.