Re: [PATCH 0/3] MR6925: d3dcompiler: Enable double-as-float option for older modules.

[Matteo Bruni (＠Mystral)]

Matteo Bruni (@Mystral) commented about dlls/d3dcompiler_43/compiler.c:
> #endif
>
> -    if (size >= 4 && read_u32(&ptr) == TAG_DXBC)
> +    source.code = data;
> +    source.size = size;
> +
> +    source_type = VKD3D_SHADER_SOURCE_D3D_BYTECODE;
> +    if (vkd3d_shader_parse_dxbc(&source, 0, &desc, NULL) >= 0)
> +    {
>          source_type = VKD3D_SHADER_SOURCE_DXBC_TPF;
> -    else
> -        source_type = VKD3D_SHADER_SOURCE_D3D_BYTECODE;
> +
> +        for (i = 0; i < desc.section_count; ++i)
> +        {
> +            if (TAG_FX10 == desc.sections[i].tag)
> Very nitpicky, but I think we consistently put the constant on the right side in this kind of comparisons.

-- 
https://gitlab.winehq.org/wine/wine/-/merge_requests/6925#note_89507