- 06 Sep, 2012 15 commits
-
-
Rafaël Carré authored
Signed-off-by: Rafaël Carré <funman@videolan.org>
-
Rémi Denis-Courmont authored
-
Rémi Denis-Courmont authored
-
Jean-Baptiste Kempf authored
-
Benjamin Drung authored
Close #7436 Signed-off-by: Rafaël Carré <funman@videolan.org>
-
Rémi Denis-Courmont authored
-
Rémi Denis-Courmont authored
-
Rémi Denis-Courmont authored
-
Ludovic Fauvet authored
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
-
Denis Charmet authored
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
-
Benjamin Drung authored
-
Francois Cartegnie authored
-
Rémi Denis-Courmont authored
-
Rémi Denis-Courmont authored
-
Rafaël Carré authored
They were spotted by ruby-elf's cowstats (gem install ruby-elf) Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
-
- 05 Sep, 2012 6 commits
-
-
Rémi Denis-Courmont authored
-
Rafaël Carré authored
-
Mark Lee authored
Signed-off-by: Rafaël Carré <funman@videolan.org>
-
Edward Wang authored
Some mingw-w64 3.0 toolchains like mine, for example, have those DLLs in the second or possibly third field of LIBRARY_PATH. Skipping these could lead to weird C++ dlopen errors. Signed-off-by: Rafaël Carré <funman@videolan.org>
-
Rafaël Carré authored
-
Diego Elio Pettenò authored
By changing the description from a pointer to an array of characters, the long tables don't need to be relocated. Unfortunately this adds 30KB of padding zeroes for empty strings. Since Lookup() returns a custom entry_t object, with a possibly modified description, this is implemented by creating a new structure type. With more work it would be possible to simply filling all the entries with final description and remove the extra matching and entry manipulation. Signed-off-by: Diego Elio Pettenò <flameeyes@flameeyes.eu> Signed-off-by: Rafaël Carré <funman@videolan.org>
-
- 04 Sep, 2012 19 commits
-
-
Jean-Baptiste Kempf authored
-
Denis Charmet authored
Close #7401 Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
-
Francois Cartegnie authored
If psz_album is NULL you have no data to insert. If you have an id then the value is already in database and only empty string can erase it.
-
Francois Cartegnie authored
-
David Fuhrmann authored
-
David Fuhrmann authored
close #6019
-
David Fuhrmann authored
-
Luca Barbato authored
Shorten build time. On x86_64 linux build: LIBTOOL make -j2 60,82s user 13,98s system 112% cpu 1:06,59 total make -j2 56,83s user 12,72s system 110% cpu 1:03,20 total DOLT make -j2 44,32s user 11,02s system 108% cpu 51,215 total make -j2 42,15s user 11,04s system 106% cpu 50,155 total Signed-off-by: Rafaël Carré <funman@videolan.org>
-
Francois Cartegnie authored
-
Francois Cartegnie authored
-
Francois Cartegnie authored
-
Diego Elio Pettenò authored
Signed-off-by: Diego Elio Pettenò <flameeyes@flameeyes.eu> Signed-off-by: Rafaël Carré <funman@videolan.org>
-
Rémi Denis-Courmont authored
As agreed last week with upstream, that's basicailly the least insane way to deal with the ever changing and expanding V4L2 API. This allows compiling VLC with old Linux headers and run it well on a new Linux kernel. This also avoids a pile of backward compatibility hacks. A few other userspace tool already copy videodev2.h like that.
-
Rémi Denis-Courmont authored
CROPCAP is no longer mandatory, and many drivers failed to provide it. If there is no CROPCAP, then there is no cropping.
-
Rafaël Carré authored
Some formats can only be converted to FI32, and then a FI32->FL32 converter can be used.
-
Rafaël Carré authored
-
Felix Paul Kühne authored
-
Martin Storsjö authored
If we need to use pts for timestamps, but pts isn't set (the avi demuxer seems to set it to 0 for cases where it isn't known, while dts has a nonzero value), use dts instead. This fixes things on OMAP4 (the only platform where b_use_pts is set at the moment), when playing back files that have only dts, no pts. Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
-
Jean-Baptiste Kempf authored
Bored as hell of those crashes...
-