- 11 Apr, 2012 13 commits
-
-
Felix Paul Kühne authored
-
Rémi Denis-Courmont authored
Also fix huge memory leak in v4l2 access (buffers never released).
-
Rémi Denis-Courmont authored
-
Rémi Denis-Courmont authored
-
Jean-Baptiste Kempf authored
Close #6598
-
Jean-Baptiste Kempf authored
However, we still cannot select the full scale. Close #6603 and #6602
-
Rémi Denis-Courmont authored
-
Rémi Denis-Courmont authored
In case of pure access: - use correct access_sys_t type, - remove unused parameters.
-
Rémi Denis-Courmont authored
-
Rémi Denis-Courmont authored
Especially fix memory leaks on error.
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
- 10 Apr, 2012 23 commits
-
-
Jean-Baptiste Kempf authored
Those that use METADATA_PICTURE_BLOCK
-
Jean-Baptiste Kempf authored
-
Rémi Denis-Courmont authored
-
Rémi Denis-Courmont authored
-
Rémi Denis-Courmont authored
-
Rémi Denis-Courmont authored
The streaming parameters depend on the format, so the format must be set explicitly, then the streaming parameters queried every time. This essentially reverts 1bb938e4.
-
Rémi Denis-Courmont authored
This brings --v4l2-width and --v4l2-height back into use. A value range is now added as well.
-
Rémi Denis-Courmont authored
Cropping depends on the image resolution, which is selected while setting the format.
-
Erwan Tulou authored
This bug was unnoticed because this constructor is only called in a situation where really copying from origin or leaving it to the compiler to set it via the default constructor leads to the same result (an empty list). Yet, this was a bug. It may solve trac #6599 though, (a Linux port where there is a real initialization issue)
-
Felix Paul Kühne authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
Ref #2882
-
Jean-Baptiste Kempf authored
-
Felix Paul Kühne authored
additionally, clean up misc.m and move code shared by VLCVoutView, VLBrushedMetalImageView and VLCThreePartDropView to VLCCoreInteraction
-
Hugo Beauzée-Luyssen authored
-
Hugo Beauzée-Luyssen authored
("0 = no subtitles autodetected\n")
-
Hugo Beauzée-Luyssen authored
-
Hugo Beauzée-Luyssen authored
-
Hugo Beauzée-Luyssen authored
-
Hugo Beauzée-Luyssen authored
-
Martin Storsjö authored
Make sure we always remove all output frames from the output frame queue, even if there's no picture. This fixes (for most practical cases at least) the issue where the app could hang repeating "can't get output picture" after stopping playback with omxil video decoding. In that case, the vout didn't return any picture to write the decoded frames into, causing the omxil codec to run out of output buffers, blocking, and thus running out of input buffers, blocking the decoder thread while waiting for a free input buffer. Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
-
Georgi Chorbadzhiyski authored
Without this fix the second CSA key can not be set, which results in a stream that is impossible to decrypt. Tested with: cvlc \ -I dummy -vv \ --sout '#standard{access=udp,mux=ts,dst=239.78.78.78:5000}' \ --sout-ts-crypt-audio --no-sout-ts-crypt-video \ --sout-ts-csa-ck=1111111111111111 \ --sout-ts-csa2-ck=0000000000000000 \ http://example.com/source.ts Before the patch: mux_ts mux debug: using CSA (de)scrambling with odd key=11:11:11:11:11:11:11:11 mux_ts mux debug: no CSA found mux_ts mux debug: no CSA found After the patch: mux_ts mux debug: using CSA (de)scrambling with odd key=11:11:11:11:11:11:11:11 mux_ts mux debug: using CSA (de)scrambling with even key=0:0:0:0:0:0:0:0 Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
-
Jean-Baptiste Kempf authored
-
- 09 Apr, 2012 4 commits
-
-
Jean-Baptiste Kempf authored
Maybe it could work with VC-1 Main, but no samples so far
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-