| commit | 688a858834ec7cdb3f7abb652b3287cd05cd35db | |
| author | Thomas Guillem <thomas@gllm.fr> | |
| Wed, 5 Sep 2018 10:52:05 +0000 (5 12:52 +0200) | ||
| committer | Thomas Guillem <thomas@gllm.fr> | |
| Wed, 5 Sep 2018 10:52:05 +0000 (5 12:52 +0200) | ||
| tree | f9a034a621237a50014e59f84ee074518349d953 | treesnapshot (tar.gz zip) |
| parent | e56bb85d41cd020bbb69cabd8d429f0463dcfb31 | commitdiff |
es_out: fix track infos regression
Regression from 7e719ff464f48cc13977fd54980e28e783612816
Add a missing input_item_UpdateTracksInfo(). Indeed, tracks were updated only
if the format was changed from the decoders (and therefore killed preparsing).
Revert back tracks i_codec to the original demuxer i_codec. Indeed, it seems
that the following statement was a bad idea:
"The fmt of the ES events will now keep the decoder codec. The original codec
from the demuxer is still available via fmt.i_original_fourcc."
Regression from 7e719ff464f48cc13977fd54980e28e783612816
Add a missing input_item_UpdateTracksInfo(). Indeed, tracks were updated only
if the format was changed from the decoders (and therefore killed preparsing).
Revert back tracks i_codec to the original demuxer i_codec. Indeed, it seems
that the following statement was a bad idea:
"The fmt of the ES events will now keep the decoder codec. The original codec
from the demuxer is still available via fmt.i_original_fourcc."
| src/input/es_out.c | diffblobblamehistory |