Re: [PATCH 3/5] wined3d: Introduce "context_apply_state()" to setup a context for a specific usage.

[Stefan Dösinger]

Am 27.12.2009 um 19:57 schrieb Henri Verbeet:
> +/* Context activation is done by the caller. */
> +static void context_apply_state(struct wined3d_context *context, IWineD3DDeviceImpl *device, enum ContextUsage usage)
> The patch looks ok to me, but out of curiosity: Do you intend to make this function available to other wined3d code like context_acquire is, or is the main motivation of the patch to tidy up the context_acquire() function?

[PATCH 3/5] wined3d: Introduce "context_apply_state()" to setup a context for a specific usage.

[Henri Verbeet]

2009/12/28 Stefan Dösinger <stefandoesinger(a)gmx.at>:
> The patch looks ok to me, but out of curiosity: Do you intend to make this function available to other wined3d code like context_acquire is, or is the main motivation of the patch to tidy up the context_acquire() function?

I'll probably kill the "usage" argument to context_acquire() at some
point, yeah.

[PATCH 3/5] wined3d: Introduce "context_apply_state()" to setup a context for a specific usage.

[Stefan Dösinger]

Am 28.12.2009 um 11:32 schrieb Henri Verbeet:
> 2009/12/28 Stefan Dösinger <stefandoesinger(a)gmx.at>:
> > The patch looks ok to me, but out of curiosity: Do you intend to make this function available to other wined3d code like context_acquire is, or is the main motivation of the patch to tidy up the context_acquire() function?
>
> I'll probably kill the "usage" argument to context_acquire() at some
> point, yeah.
> This might be useful in some places, although currently I can only think of the non-fbo-blit swapchain present blit(which could be dealt with in other ways maybe).

Keep in mind though that context_apply_state only works properly if the context is current, and that might make it somewhat awkward to use because callers have to acquire the context first.