On Mon, Nov 07, 2016 at 16:30:02 -0800, Zhenni Huang wrote: > - avio_rb32(pb); /* vendor */ > + video_vendor_id = avio_rb32(pb); /* vendor */ > avio_rb32(pb); /* temporal quality */ > avio_rb32(pb); /* spatial quality */ > > + /* set video_vendor_id */ > + video_vendor_id_buffer[0] = (video_vendor_id >> 24) & 0xff; > + video_vendor_id_buffer[1] = (video_vendor_id >> 16) & 0xff; > + video_vendor_id_buffer[2] = (video_vendor_id >> 8) & 0xff; > + video_vendor_id_buffer[3] = (video_vendor_id >> 0) & 0xff; > + video_vendor_id_buffer[4] = 0; > + av_dict_set(&st->metadata, "vendor_id", video_vendor_id_buffer, 0);
Wouldn't you avoid shifting the bytes around if you did an avio_rl32() in the first place, i.e. video_vendor_id = avio_rl32(pb); Then you could just initialize the buffer to 0 and memcopy the id over. Or am I missing something? (If not: Others tend to complain about useless masks and useless shifts of 0. I personally think they're okay for readability.) Moritz _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
