- 27 Nov, 2010 1 commit
-
-
Pierre Ynard authored
Based on keep-alive RTSP requests
-
- 26 Nov, 2010 19 commits
-
-
Laurent Aimar authored
CommonDisplay() must be called at least once to ensure that the window content is shown.
-
Laurent Aimar authored
-
Laurent Aimar authored
-
Laurent Aimar authored
-
Laurent Aimar authored
It computes correct crop/ar settings when scaling a video format.
-
Laurent Aimar authored
-
Hugo Beauzée-Luyssen authored
Signed-off-by: Rémi Duraffort <ivoire@videolan.org>
-
Hugo Beauzée-Luyssen authored
Signed-off-by: Rémi Duraffort <ivoire@videolan.org>
-
Jean-Paul Saman authored
Move logging of current segment and stream. The previous place gave confusing information.
-
Jean-Paul Saman authored
After the adaptive bandwidth changes seeking could end up in a segment that is not available. Each segment is downloaded once and after switching to higher or lower bandwidth segments downloaded for previous bandwidths will not be deleted or downloaded. The later is to be fixed.
-
Jean-Paul Saman authored
Start playback with the lowest bandwidth stream.
-
Jean-Paul Saman authored
NextSegment() assumed that the download thread its hls_stream_t was leading. However this was too simplistic. NextSegment() seeks the segment in *any* of the hls_stream_t, assuming that each segment will be downloaded only once.
-
Jean-Paul Saman authored
Test script for traffic shaping with tc (iproute).
-
Jean-Paul Saman authored
Download the next segment from a playback standpoint.
-
Jean-Paul Saman authored
Cleanup NextSegment().
-
Jean-Paul Saman authored
The bandwidth calculation needed corrections and is less agressive in switching to higher bandwidths.
-
Rafaël Carré authored
(It was broken by a8e4b30b) Move the definition to qt4.hpp and add a comment
-
Pierre Ynard authored
-
Pierre Ynard authored
-
- 25 Nov, 2010 1 commit
-
-
Pierre Ynard authored
If you want to use #description, you need to set up the appropriate struct on pass it down as a variable
-
- 24 Nov, 2010 12 commits
-
-
Rémi Denis-Courmont authored
-
Rémi Denis-Courmont authored
-
Rémi Duraffort authored
-
Rémi Denis-Courmont authored
-
Rémi Denis-Courmont authored
-
Rémi Denis-Courmont authored
This also fixes error handling: if the thread cannot be created, we do fail cleanly.
-
Rémi Duraffort authored
-
Rémi Duraffort authored
-
Pierre Ynard authored
This allows using it at another time than destroying the object
-
Rémi Denis-Courmont authored
-
Rémi Denis-Courmont authored
-
Pierre Ynard authored
-
- 23 Nov, 2010 7 commits
-
-
Rémi Denis-Courmont authored
...and less prone to deadlocks. This restarts until all keys are NULL, in case cleaning up reset one key back to a non-NULL value. There is still a slight difference with POSIX still. POSIX will restarts only after all destructors have been invoked, not after the first one.
-
Rémi Denis-Courmont authored
-
Rémi Denis-Courmont authored
-
Rémi Denis-Courmont authored
In practice, this bug was hidden by another bug in the thread variables implementation: the destruction callback is never invoked.
-
Rafaël Carré authored
This reverts commit eea468e6. When the media is played, the formats represent decoder output, and the ids are not preserved (video ES can be replaced by audio ES)
-
Rafaël Carré authored
This reverts commit 708a15ff.
-
Rémi Duraffort authored
-