On 05-01-2020 02:14, Zhiyi Zhang wrote:
>
> On 1/4/20 7:59 PM, Sven Baars wrote:
>> From: Alistair Leslie-Hughes <leslie_alistair@hotmail.com>
>>
>> Signed-off-by: Sven Baars <sbaars@codeweavers.com>
>> ---
>> I'm sending this series mostly so it can be reviewed, but I sent the
>> tests first so they may still be committed during code freeze, if
>> they are acceptable.
>>
>> dlls/d3dx9_36/tests/core.c | 63 ++++++++++++++++++++++++++++++++++++++
>> 1 file changed, 63 insertions(+)
>>
>> diff --git a/dlls/d3dx9_36/tests/core.c b/dlls/d3dx9_36/tests/core.c
>> index fa874a5ef5..8b90f0ef5f 100644
>> --- a/dlls/d3dx9_36/tests/core.c
>> +++ b/dlls/d3dx9_36/tests/core.c
>> @@ -637,6 +637,69 @@ static void test_ID3DXFont(IDirect3DDevice9 *device)
>> ID3DXFont_Release(font);
>> }
>> +
>> + /* ID3DXFont_DrawTextA, ID3DXFont_DrawTextW */
>> + hr = D3DXCreateFontA(device, 12, 0, FW_DONTCARE, 0, FALSE,
>> DEFAULT_CHARSET, OUT_DEFAULT_PRECIS, DEFAULT_QUALITY, DEFAULT_PITCH,
>> "Arial", &font);
> Arial may be unavailable. Please use Tahoma instead.
All tests in that file use Arial though, as well as the DrawText tests
in user32. And if I grep all tests in wine, I get about the same amount
of matches for Arial as for Tahoma. Should all other tests be changed as
well?
For purposes of d3dx9 tests I think it's enough to test that sizes match what gdi (or user32:DrawText) calls return, and to make it more predictable, using Tahoma, since wine provides it.