Originally committed as revision 3883 to svn://svn.ffmpeg.org/ffmpeg/trunk
for(y= bp->height%tile_h; y< bp->height; y+=tile_h){
for(x= bp->width%tile_w; x< bp->width; x+=tile_w){
code = get_vlc2(&v->gb, vc9_norm6_vlc.table, VC9_NORM6_VLC_BITS, 2);
+ if(code<0){
+ av_log(v->avctx, AV_LOG_DEBUG, "inavlid NORM-6 VLC\n");
+ return -1;
+ }
//FIXME following is a pure guess and probably wrong
//FIXME A bitplane (0 | !0), so could the shifts be avoided ?
planep[x + 0*bp->stride]= (code>>0)&1;