video_screen_get_vblank() was also incorrect for games with MDRV_SCREEN_VBLANK_TIME, this could easily explain a bunch of changed screenshots.

This commit is contained in:
Zsolt Vasvari 2008-03-12 12:19:43 +00:00
parent 4bd4eceb06
commit 9626c879db

View File

@ -970,8 +970,11 @@ int video_screen_get_vblank(const device_config *screen)
{
screen_state *state = get_safe_token(screen);
internal_screen_state *internal_state = (internal_screen_state *)state->private_data;
int vpos = video_screen_get_vpos(screen);
return (vpos < internal_state->visarea.min_y || vpos > internal_state->visarea.max_y);
/* we should never be called with no VBLANK period - indication of a buggy driver */
assert(internal_state->vblank_period != 0);
return (attotime_compare(timer_get_time(), internal_state->vblank_end_time) < 0);
}
@ -1085,7 +1088,7 @@ attotime video_screen_get_time_until_vblank_end(const device_config *screen)
attotime current_time = timer_get_time();
/* we are in the VBLANK region, compute the time until the end of the current VBLANK period */
if (attotime_compare(current_time, internal_state->vblank_end_time) < 0)
if (video_screen_get_vblank(screen))
ret = attotime_sub(internal_state->vblank_end_time, current_time);
/* otherwise compute the time until the end of the next frame VBLANK period */