On Mon Feb 27 14:19:23 2023 +0000, Henri Verbeet wrote:
```diff @@ -1934,10 +1967,15 @@ VKD3D_SHADER_API void
vkd3d_shader_free_dxbc(struct vkd3d_shader_dxbc_desc *dxbc
* vkd3d_shader_free_dxbc() when no longer needed. * * \param messages Optional output location for error or informational messages - * produced by the parser. + * produced by the compiler. * \n - * This parameter behaves identically to the \a messages parameter of - * vkd3d_shader_compile(). + * This string is null-terminated and UTF-8 encoded. + * \n + * The messages are allocated by vkd3d-shader and should be freed with + * vkd3d_shader_free_messages() when no longer needed. + * \n + * If no compilation messages are produced by the compiler, this parameter may + * receive NULL instead of a valid string pointer. * * \return A member of \ref vkd3d_result. * ``` This is the comment for vkd3d_shader_parse_dxbc(), though there are similar issues with the other functions. References like "produced by the compiler", and "compilation messages" seem somewhat out of place. I suppose "the compiler" could be interpreted as vkd3d-shader in general, but in that case we might as well replace it with just "vkd3d-shader". Right, I fixed them. I also fixed a few other comments that were sort of inconsistent in the same way. And also changed "serialis-" to "serializ-", mostly choosing the "z" version because it's what's being used in the function names. It's also the version closer to the original Greek suffix, so I hope it will make @zfigura happier.
-- https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/110#note_25594