video: create a separate context for video filter chain

This adds vf_chain, which unlike vf_instance refers to the filter chain
as a whole. This makes the filter API less awkward, and will allow
handling format negotiation better.

1 files changed, 2 insertions, 4 deletions

diff --git a/mpvcore/player/screenshot.c b/mpvcore/player/screenshot.c
index 6a3c258094..bf7a0ce9ee 100644
--- a/mpvcore/player/screenshot.c
+++ b/mpvcore/player/screenshot.c
@@ -314,10 +314,8 @@ static struct mp_image *screenshot_get(struct MPContext *mpctx, int mode)
         struct voctrl_screenshot_args args =
                             { .full_window = (mode == MODE_FULL_WINDOW) };
 
-        if (mpctx->d_video && mpctx->d_video->vfilter) {
-            struct vf_instance *vfilter = mpctx->d_video->vfilter;
-            vfilter->control(vfilter, VFCTRL_SCREENSHOT, &args);
-        }
+        if (mpctx->d_video && mpctx->d_video->vfilter)
+            vf_control_any(mpctx->d_video->vfilter, VFCTRL_SCREENSHOT, &args);
 
         if (!args.out_image)
             vo_control(mpctx->video_out, VOCTRL_SCREENSHOT, &args);