| Age | Commit message (Collapse) | Author |
|
in one case where it should say "opening for read/write".
Also add some unit tests for ReelWriter.
|
|
Backported from f4f6a2504e5fedb67513689eaeac9dd88ee38ca9 in 2.15.x.
|
|
|
|
|
|
|
|
|
|
|
|
streams with no codec (#1184).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
an image subtitle is completed with a rect count of zero) which
subsequently cause hangs while the decoder looks for these
phantom text subtitles. Fixes #812.
|
|
|
|
|
|
Sadly an example shows that from times are not unique. This patch
uses a hash of stuff from the first AVSubtitle as the key.
|
|
|
|
|
|
|
|
in FFmpegDecoder. This offset was not being taken into account for
subtitles prior to this commit.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
There was existing code to work around the case when FFmpeg files
have all-zero stream IDs. Extend this to cope with duplicate
stream IDs, as these have been seen in the wild.
|
|
in a similar way to the V1 patch.
|
|
some video MXFs (#566).
|
|
|
|
|
|
|
|
iTunes).
|
|
|
|
unnecessary uses of stringstream.
|
|
|
|
|
|
|
|
|
|
|