- 11 Apr, 2015 10 commits
-
-
Mario Siegmann authored
Signed-off-by:Christoph Miebach <christoph.miebach@web.de>
-
VideoLAN authored
Signed-off-by: -
Tomáš Chvátal authored
Signed-off-by: -
VideoLAN authored
Signed-off-by: -
VideoLAN authored
Signed-off-by: -
VideoLAN authored
Signed-off-by: -
David Fuhrmann authored
Because the playlist is started before the interface is up, the vout window provider waits until basic initalization is done in the interface (initialization of VLCIntf, NSApp, static vars). This is still sort of a hack, ideally the playlist should be started only after all interfaces are fully initialized. close #11585 (cherry picked from commit 4a17358abea2b8fce47e8af3888aec638debfb42) Signed-off-by:David Fuhrmann <dfuhrmann@videolan.org>
-
David Fuhrmann authored
Because the playback might have been started before the interface registeres its callbacks, set the state manually. close #14046 (cherry picked from commit 5598f816905d5262928650c38d6b9a501b232deb) Signed-off-by: -
David Fuhrmann authored
This is still racy, but it should reduce the time between playlist and interface start. (cherry picked from commit b5c0fe454de96a1328751a896b42daf6ee732284) Signed-off-by: -
Christoph Miebach authored
-
- 10 Apr, 2015 12 commits
-
-
Jean-Baptiste Kempf authored
Close #14294 (cherry picked from commit ac3ff888e2db9dab0319ff0715d849572b7a7145) Signed-off-by:Jean-Baptiste Kempf <jb@videolan.org>
-
Francois Cartegnie authored
Close #14078 Signed-off-by: -
Jean-Baptiste Kempf authored
This is a HACK to work-around #14054 This is NOT a fix, but well. It will default the Windows to "no" acceleration, until it is fixed.
-
Jean-Baptiste Kempf authored
This reverts commit fc6d0fc4. This should now be fixed by mstack
-
Jean-Baptiste Kempf authored
This reverts commit 4c0b7973. Ref #14062
-
Jean-Baptiste Kempf authored
This reverts commit 7e94e20b. Ref #14062
-
Jean-Baptiste Kempf authored
This reverts commit 74c432a3. Ref #14062
-
Jean-Baptiste Kempf authored
This reverts commit 5b7b2a5e. Ref #14062
-
Jean-Baptiste Kempf authored
This reverts commit 2983a6e3. Ref #14062
-
Tristan Matthews authored
Fixes #14104 Signed-off-by:
-
Jean-Baptiste Kempf authored
-
Francois Cartegnie authored
(cherry picked from commit 624a2a4c1d53731126f4c1c41682d8d9649870d4) Signed-off-by:
-
- 09 Apr, 2015 2 commits
-
-
Jean-Baptiste Kempf authored
The changes on preferences association registration, seems to have attacked the taskbar part (cherry picked from commit 0dce302d4cb74f4f9a2ddebb4a5db315895d58ea) Signed-off-by: -
Jean-Baptiste Kempf authored
Close #10350 (cherry picked from commit cf1631d2387249e1e9355cfe9a7378ec7ac8df37) Signed-off-by:
-
- 08 Apr, 2015 9 commits
-
-
Tristan Matthews authored
It is now under "archive" instead of "official_releases"
-
Rémi Denis-Courmont authored
The model does not matter here as affected objects do not outlive the function call and are not used across threads. Still the references counter of the apartment must remain balanced. (cherry picked from commit aa9976b1f41fe7eb44b640163c5ff5f00a4e2953) Signed-off-by: -
Francois Cartegnie authored
output pts/length is computed from input block pts. If sample rate is wrong, all timings are borked. Try to use the declared sample rate if possible. If source has wrong sample rate (cherry picked from commit 06682f220832758726789cb735bfb89f80cabe56) Signed-off-by: -
Sean McGovern authored
Signed-off-by:
-
Sean McGovern authored
Signed-off-by:
-
Francois Cartegnie authored
(cherry picked from commit 751ecf1c78191151e27f898bd58345744f35bead) Signed-off-by: -
Rémi Denis-Courmont authored
(cherry picked from commit 2c785718280fd69bdfe4c033c454a0834630bf19) Signed-off-by: -
Hugo Beauzée-Luyssen authored
SetThreadErrorMode returns 0 on failure (cherry picked from commit 72b2f4a51d9cfe103bc9ca5644d87c0e5abdab4d) Signed-off-by: -
Steve Lhomme authored
Signed-off-by:
-
- 01 Apr, 2015 1 commit
-
-
Rémi Denis-Courmont authored
(cherry picked from commit 0e8b7471fe0249c59a9e02e556a5df8dad0b224f) Signed-off-by:
-
- 29 Mar, 2015 2 commits
-
-
Rémi Denis-Courmont authored
(cherry picked from commit e7ef356de5e6d6f931478e92cc000480d8284974) Signed-off-by: -
Rémi Denis-Courmont authored
(cherry picked from commit 39e4fad1bc929bb35a5353380b456e36f23ea1c4) Signed-off-by:
-
- 26 Mar, 2015 1 commit
-
-
Fabian Yamaguchi authored
The variable len is a raw 32 bit value read using GetDWBE. If this value is larger than UINT32_MAX - sizeof(eos), this will cause an integer overflow in the subsequent call to malloc, and finally a buffer overflow when calling memcpy. We fix this by checking len accordingly. Signed-off-by:
-
- 25 Mar, 2015 3 commits
-
-
Hannes Domani authored
Close #14140 Signed-off-by:
-
Jonathan McGowan authored
A bad parameter definition in function RegisterExtension resulted in the same value entry being written every time ("${EXT}"=""). Signed-off-by: -
Jean-Baptiste Kempf authored
Of course, it's technically a GCC bug, because selectany is broken, but there is no simple fix for that. (cherry picked from commit 48d665494c2d23e474081a8b222942068b74b9b3) Signed-off-by:
-
