| Age | Commit message (Collapse) | Author |
|
With this timeline:
-------> t
SUB1 SUB2 X
We might seek to X, then pass(). Before this change, SUB1 would have been emitted by
the call to the subtitle decoder, then we'd have emitted some black for X. This would
lose SUB2. Now we keep pass()ing and don't emit X until it's the earliest thing
(in the same way that the main decision of what to pass() works).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
seek position, and make Player cope with the FFmpegDecoder emitting
pre-roll data. This seems nicer than making FFmpegDecoder do it.
|
|
|
|
doing anything so that the exception can be reported.
|
|
|
|
|
|
|
|
|
|
efficient by adjusting decoders.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
sound/subtitle needs a KDM.
|
|
|