No Float16 emitted with use of input 16-bit float

to sum up the issue, I think `Float16` is missing with use of `GL_EXT_shader_explicit_arithmetic_types_float16 `. The following example shows `Int16` is set, but not `Float16`

---

So I have 2 shaders

```glsl
#version 450
#extension GL_EXT_shader_explicit_arithmetic_types_float16 : enable
layout(location = 0) in f16vec2 inPosition;
```

and

```glsl
#version 450
#extension GL_EXT_shader_explicit_arithmetic_types_int16 : enable
layout(location = 0) in i16vec2 inPosition;
```

The first produces

```
Capability Shader
Capability StorageInputOutput16
Extension "SPV_KHR_16bit_storage"
```

While the second produces

```
Capability Shader
Capability Int16
Capability StorageInputOutput16
Extension "SPV_KHR_16bit_storage"
```

Start with glslangvalidator and compile the two minimal GLSL examples from the issue, comparing their generated SPIR-V capabilities and extensions. Trace the compiler path handling GL_EXT_shader_explicit_arithmetic_types_float16 and verify completion when the float16 shader emits the expected Float16 capability alongside its existing output.

Written by the indexing model from the issue text.