Re: [PATCH 3/7] ddraw: Fix IDirectDraw_CreateSurface's invalid descriptor error code.

[Henri Verbeet]

On 15 February 2017 at 07:53, Alex Henrie  wrote:
> @@ -3000,7 +3000,7 @@ static HRESULT WINAPI ddraw1_CreateSurface(IDirectDraw *iface,
>      {
>          WARN("Application supplied invalid surface descriptor\n");
>          wined3d_mutex_unlock();
> -
>    return DDERR_INVALIDPARAMS;
> -
>    return E_INVALIDARG;

Did you verify the test fails without this change? How is
DDERR_INVALIDPARAMS defined?

[PATCH 3/7] ddraw: Fix IDirectDraw_CreateSurface's invalid descriptor error code.

[Alex Henrie]

2017-02-15 6:49 GMT-07:00 Henri Verbeet :
> On 15 February 2017 at 07:53, Alex Henrie  wrote:
> > @@ -3000,7 +3000,7 @@ static HRESULT WINAPI ddraw1_CreateSurface(IDirectDraw *iface,
> >      {
> >          WARN("Application supplied invalid surface descriptor\n");
> >          wined3d_mutex_unlock();
> > -
> >    return DDERR_INVALIDPARAMS;
> > -
> >    return E_INVALIDARG;
>
> Did you verify the test fails without this change? How is
> DDERR_INVALIDPARAMS defined?

You're right; these patches are unnecessary. ddraw.h has:

#define DDERR_INVALIDPARAMS E_INVALIDARG

Which of the two is preferred for ddraw code?

-Aex

[PATCH 3/7] ddraw: Fix IDirectDraw_CreateSurface's invalid descriptor error code.

[Stefan Dösinger]

Am 2017-02-15 um 16:17 schrieb Alex Henrie:
> Which of the two is preferred for ddraw code?

We generally use DDERR_INVALIDPARAMS.

One could argue that an inherited method like QueryInterface should
return E_INVALIDARG instead, but this is all very academic. I'd say just
keep what the existing code has.