dvbsubdec: fix division by zero in compute_default_clut

This problem was introduced in commit
4b90dcb8.
Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
(cherry picked from commit c82b8ef0)
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

libavcodec/dvbsubdec.c

@@ -810,7 +810,7 @@ static void compute_default_clut(AVPicture *frame, int w, int h)
         list_inv[     i ] = bestv;
     }
 
-    count = i - 1;
+    count = FFMAX(i - 1, 1);
     for (i--; i>=0; i--) {
         int v = i*255/count;
         AV_WN32(frame->data[1] + 4*list_inv[i], RGBA(v/2,v,v/2,v));