It seems like some software can insert an additional ID3v2 tag instead of appending a frame to an existing one. One could argue that that is somewhat broken, but I agree it's better to handle it instead of returning an error. The changes in aacdec.c look ok to me.
The fate sample you provided has two tags at the beginning of the file. These will actually get parsed by avformat_open_input calling ff_id3v2_read_dict and not by the adts demuxer. The ff_id3v2_read_dict code path will already parse multiple tags, which is why the test passes. In order to exercise your code change, the ID3v2 tags would have to be between ADTS frames. -Richard On Mon, Apr 2, 2018 at 4:35 AM, Mattias Amnefelt <matti...@avm.se> wrote: > This is a follow-up to https://patchwork.ffmpeg.org/patch/7477/ and changes > so all ID3 tags between ADTS frames gets parsed, not just the first one. > > Sample media for the included fate test is available at > http://mattias.amnefe.lt/tmp/id3v2_two_tags.aac > > > > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-devel > _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
