stride3d/stride · error · NotSupportedException

Unsupported float vector size

Error message

Unsupported float vector size: {v.Length}

What it means

ConvertDefaultValue maps a shader constant's default float vector into a managed Stride vector type, but only sizes 2, 3, and 4 have mappings. Any other component count (e.g. 1, 5, or non-square matrices flattened) hits the switch default and throws NotSupportedException.

Solutions

  1. Change the default value/member to float2, float3, or float4
  2. Use a plain scalar float instead of a 1-component vector
  3. Split larger composites into supported vector types

Example fix

// before
float1 Scale = float1(2.0);
// after
float Scale = 2.0;
Defensive patterns

Strategy: validation

Validate before calling

if (elemType == "float" && composite.Length is < 2 or > 4)
    throw new ArgumentException("float default vectors must have 2-4 components");

Try / catch

try { result = mixer.MergeSDSL(tree); }
catch (NotSupportedException ex) when (ex.Message.StartsWith("Unsupported float vector size")) {
    fixDefaultValueType(memberName);
}

Prevention

When it happens

Trigger: A cbuffer member with a default value declared as float1 (or floatN with N outside 2..4) reaching ConvertDefaultValue via ComputeCBufferReflection; matrix defaults encoded as vectors of unexpected length.

Common situations: float scalar defaults declared with vector initializer syntax; shader code generators emitting float5+ composites; GLSL/HLSL ported code using vec1-like constructs.

Related errors

AI-assisted analysis of stride3d/stride@96fad776d2 (2026-09-14). Data as JSON: /api/errors/e7bc0a0a58769043. Report an issue: GitHub.

Appendix: source

Thrown at sources/shaders/Stride.Shaders.Compilers/SDSL/ShaderMixer.CBuffers.cs:641

        /// <summary>
        /// Converts compiler-internal <see cref="ConstantVector"/> to serializable CLR types
        /// that the engine expects (Vector2/3/4, Int2/3/4, etc.).
        /// Scalars (float, int, uint, bool) pass through unchanged.
        /// </summary>
        private static object? ConvertDefaultValue(object? value)
        {
            if (value is not ConstantVector cv)
                return value;

            var v = cv.Values;
            return v[0] switch
            {
                float => v.Length switch
                {
                    2 => new Vector2((float)v[0], (float)v[1]),
                    3 => new Vector3((float)v[0], (float)v[1], (float)v[2]),
                    4 => (object)new Vector4((float)v[0], (float)v[1], (float)v[2], (float)v[3]),
                    _ => throw new NotSupportedException($"Unsupported float vector size: {v.Length}"),
                },
                int => v.Length switch
                {
                    2 => new Int2((int)v[0], (int)v[1]),
                    3 => new Int3((int)v[0], (int)v[1], (int)v[2]),
                    4 => (object)new Int4((int)v[0], (int)v[1], (int)v[2], (int)v[3]),
                    _ => throw new NotSupportedException($"Unsupported int vector size: {v.Length}"),
                },
                uint => v.Length switch
                {
                    4 => (object)new UInt4((uint)v[0], (uint)v[1], (uint)v[2], (uint)v[3]),
                    _ => throw new NotSupportedException($"Unsupported uint vector size: {v.Length}"),
                },
                _ => throw new NotSupportedException($"Unsupported constant composite element type: {v[0]?.GetType()}"),
            };
        }
    }
}

View on GitHub (pinned to 96fad776d2)