Re: [PATCH 1/3] user32/tests: Added client-side raw input function tests (try 5 resend)

[Vincas Miliūnas]

On 06/22/2011 06:26 PM, André Hentschel wrote:
> Am 22/06/2011 06:26, schrieb Vincas Miliūnas:
> > - if (pDefRawInputProc && pGetRawInputBuffer && pGetRawInputData

&& pGetRawInputDeviceList &&
> > -
> >    pGetRawInputDeviceInfoA && pGetRawInputDeviceInfoW &&

pGetRegisteredRawInputDevices && pRegisterRawInputDevices)
> You should avoid such a construct, move the check to test_functions

that actually need these functions.
> e.g. check for pGetRawInputBuffer only in test_get_raw_input_buffer
> i didn't looked very close, but if the tests depend on each other

there should be a comment.
> the rest of that patch looks fine to me

Done - http://source.winehq.org/patches/data/75756

[PATCH 1/3] user32/tests: Added client-side raw input function tests (try 5 resend)

[André Hentschel]

Am 26.06.2011 18:41, schrieb Vincas Miliūnas:
> Done - http://source.winehq.org/patches/data/75756
>
> +static void test_raw_input_functions(void)
> +{
> - if (pGetRawInputData)
> -
>    test_get_raw_input_data();
> - else
> -
>    win_skip("GetRawInputData is not available\n");
> -
> - if (pRegisterRawInputDevices)
> -
>    test_register_raw_input_devices();
> - else
> -
>    win_skip("RegisterRawInputDevices is not available\n");
> -
> - if (pGetRawInputBuffer)
> -
>    test_get_raw_input_buffer();
> - else
> -
>    win_skip("GetRawInputBuffer is not available\n");
> -
> - if (pDefRawInputProc)
> -
>    test_def_raw_input_proc();
> - else
> -
>    win_skip("DefRawInputProc is not available\n");
> -
> - if (pGetRawInputDeviceList)
> -
>    test_get_raw_input_device_list();
> - else
> -
>    win_skip("GetRawInputDeviceList is not available\n");
> -
> - if (pGetRawInputDeviceInfoA && pGetRawInputDeviceList)
> -
>    test_get_raw_input_device_info_a();
> - else
> -
>    win_skip("GetRawInputDeviceInfoA and pGetRawInputDeviceList are not available\n");
> -
> - if (pGetRawInputDeviceInfoW && pGetRawInputDeviceList)
> -
>    test_get_raw_input_device_info_w();
> - else
> -
>    win_skip("GetRawInputDeviceInfoW and pGetRawInputDeviceList are not available\n");
> -
> - if (pGetRegisteredRawInputDevices && pRegisterRawInputDevices)
> -
>    test_get_registered_raw_input_devices();
> - else
> -
>    win_skip("GetRegisteredRawInputDevices and pRegisterRawInputDevices are not available\n");
> -
> - if (pRegisterRawInputDevices && pGetRawInputData && pDefRawInputProc && pGetRawInputBuffer && pSendInput)
> -
>    test_get_raw_input_data_simulation();
> - else
> -
>    win_skip("Functions required to perform raw input simulation are not available\n");
>
> +}

1.)
No, i thought about something like:

static void test_get_raw_input_data_simulation(void)
{
    HWND hWnd;
    MSG msg;
    WNDCLASSA wclass;
    HANDLE hInstance = GetModuleHandleA( NULL );
    BOOL ret;
    INPUT inputs[2];
    UINT timer, size, ret2;
    RAWINPUT *raw;

if(!pGetRawInputData)
    {
        win_skip( "some reason\n" );
        return;
    }

wclass.lpszClassName = "GetRawInputDataTestClass";
....

2.)
Why does patch 3/3 not remove todo_wines from the test?

3.)
I'm not sure about the way you splitted the patches. Splitting is good and sending tests first is often a good idea, but in 2/3 you introduce dead code which only later get's used in 3/3.
Maybe you should split per function if possible.