[PATCH vkd3d 6/6] vkd3d-shader/hlsl: Write SM4 addition instructions.

[Zebediah Figura (she/her)]

On 8/26/21 10:48 AM, Matteo Bruni wrote:
> On Tue, Aug 24, 2021 at 8:19 AM Zebediah Figura <zfigura at codeweavers.com> wrote:
>>
>> Signed-off-by: Zebediah Figura <zfigura at codeweavers.com>
>> ---
>>   Makefile.am                  |  1 -
>>   libs/vkd3d-shader/hlsl_sm4.c | 62 ++++++++++++++++++++++++++++++++++++
>>   2 files changed, 62 insertions(+), 1 deletion(-)
> 
>> diff --git a/libs/vkd3d-shader/hlsl_sm4.c b/libs/vkd3d-shader/hlsl_sm4.c
>> index 27058fef..32a4eda1 100644
>> --- a/libs/vkd3d-shader/hlsl_sm4.c
>> +++ b/libs/vkd3d-shader/hlsl_sm4.c
> 
>> +        default:
>> +        {
>> +            struct vkd3d_string_buffer *string;
>> +
>> +            if ((string = hlsl_type_to_string(ctx, expr->node.data_type)))
>> +                hlsl_fixme(ctx, expr->node.loc, "SM4 %s expression.", string->buffer);
>> +            hlsl_release_string_buffer(ctx, string);
>> +            break;
>> +        }
>> +    }
>> +}
> 
> This means that the hlsl_fixme(), and thus the compilation failure,
> only happens when hlsl_type_to_string() succeeds. It's a pretty
> theoretical issue but still a bit surprising.
> 

That's true in general. On the other hand, if hlsl_type_to_string fails, 
it'll set ctx->result to VKD3D_ERROR_OUT_OF_MEMORY, which 
vkd3d_shader_compile() will return. Since we almost certainly can't 
print an error at that point anyway, I think that qualifies as the best 
we can do.