From: Zebediah Figura <zfigura@codeweavers.com>
Subject: Re: [PATCH 3/5] d3dcompiler: Check for missing return value semantics on the entry point.
Message-Id: <2e05eca6-781d-dbfd-f947-72faa2426cab@codeweavers.com>
Date: Mon, 8 Jun 2020 10:07:58 -0500
In-Reply-To: <CABvNrtPs_jHOEuK1_0S_Dyd8oRJCw5rzzJ0=Gtw+WMZy_JLv_g@mail.gmail.com>
References: <20200605221933.1861389-1-zfigura@codeweavers.com> <20200605221933.1861389-3-zfigura@codeweavers.com> <CABvNrtPs_jHOEuK1_0S_Dyd8oRJCw5rzzJ0=Gtw+WMZy_JLv_g@mail.gmail.com>

On 6/8/20 9:14 AM, Matteo Bruni wrote:
> On Sat, Jun 6, 2020 at 12:26 AM Zebediah Figura <z.figura12@gmail.com> wrote:
>>
>> diff --git a/dlls/d3dcompiler_43/tests/hlsl_d3d9.c b/dlls/d3dcompiler_43/tests/hlsl_d3d9.c
>> index 059f39e4c3e..16d1e6dda05 100644
>> --- a/dlls/d3dcompiler_43/tests/hlsl_d3d9.c
>> +++ b/dlls/d3dcompiler_43/tests/hlsl_d3d9.c
>> @@ -1092,6 +1092,12 @@ static void test_fail(void)
>>           "{\n"
>>           "    return float4(0, 0, 0, 0);\n"
>>           "}",
>> +
>> +        /* 15 */
>> +        "float4 test()\n"
>> +        "{\n"
>> +        "    return float4(0, 0, 0, 0);\n"
>> +        "}",
> 
> Seems sensible, but, playing devil's advocate: does it fail if the
> function also has an out argument?
> 

Yes. I can add that as a test.