Re: [PATCH 02/10] d3d9: avoid using a wined3d_surface in get_desc calls inside d3d9_device_StretchRect

[Henri Verbeet]

On 19 October 2015 at 19:07, Riccardo Bortolato <rikyz619(a)gmail.com> wrote:
> +    if (!(sub_resource = wined3d_texture_get_sub_resource(dst->wined3d_texture, dst->sub_resource_idx)))
> +        goto done;
> This should never fail.

[PATCH 02/10] d3d9: avoid using a wined3d_surface in get_desc calls inside d3d9_device_StretchRect

[Riccardo Bortolato]

Alright, I will remove the checks even I thought we were checking this
everytime we call it regardless it can fail or not (plus it seems
somehow harmless to me).

Ciao,
Riccardo

2015-10-20 16:36 GMT+02:00 Henri Verbeet <hverbeet(a)gmail.com>:
> On 19 October 2015 at 19:07, Riccardo Bortolato <rikyz619(a)gmail.com> wrote:
> > +    if (!(sub_resource = wined3d_texture_get_sub_resource(dst->wined3d_texture, dst->sub_resource_idx)))
> > +        goto done;
> > This should never fail.

[PATCH 02/10] d3d9: avoid using a wined3d_surface in get_desc calls inside d3d9_device_StretchRect

[Henri Verbeet]

On 20 October 2015 at 16:48, Riccardo Bortolato <rikyz619(a)gmail.com> wrote:
> Alright, I will remove the checks even I thought we were checking this
> everytime we call it regardless it can fail or not (plus it seems
> somehow harmless to me).

It's mostly harmless, though if it ever did fail for some reason, it
would be better to crash than to return a failure to the application.
We check the return if the index comes from outside the dll, we don't
if it's known to be valid.