- 26 Jul, 2014 3 commits
-
-
Martin Storsjö authored
For input, libui.so from a 4.0 build is used, and libui-hc.so from a 3.2 build. The generated symbol list in libui.symbols contains symbols from both, to allow linking for all generations of libui.so. Signed-off-by:
Martin Storsjö <martin@martin.st>
-
Martin Storsjö authored
Previously this used a checkout from ICS (which worked fine so far). There are no honeycomb tags in this repo, so we can only guess and pick commits based on where important changes were made. This set of headers is from commit 41693fa, which is the last commit before gralloc.h started including system/window.h (which caused issues with the system/core headers from ICS, which provided a duplicate definition of android_native_buffer_t. Signed-off-by:
Martin Storsjö <martin@martin.st>
-
Martin Storsjö authored
The honeycomb history hasn't been published with branches and tags, but nowadays there is a tag "android-3.2.4_r1" in frameworks/base, allowing picking headers much closer to what honeycomb devices actually used than the previous random guess. This is necessary for working IOMX direct rendering on honeycomb, where the GraphicsBuffer class ABI is slightly different from what it was in the previously guessed (randomly picked) honeycomb header checkout. Signed-off-by:
Martin Storsjö <martin@martin.st>
-
- 25 Jul, 2014 12 commits
-
-
Jean-Baptiste Kempf authored
-
Paulo Vitor Magacho da Silva authored
Signed-off-by:
Jean-Baptiste Kempf <jb@videolan.org>
-
Jean-Baptiste Kempf authored
-
Paulo Vitor Magacho da Silva authored
Signed-off-by:
Jean-Baptiste Kempf <jb@videolan.org>
-
Paulo Vitor Magacho da Silva authored
Signed-off-by:
Jean-Baptiste Kempf <jb@videolan.org>
-
Paulo Vitor Magacho da Silva authored
Signed-off-by:
Jean-Baptiste Kempf <jb@videolan.org>
-
Thomas Guillem authored
This is useful for debugging. Don't warn if unable to load these debug libs. Signed-off-by:
Martin Storsjö <martin@martin.st>
-
Thomas Guillem authored
Signed-off-by:
Martin Storsjö <martin@martin.st>
-
Martin Storsjö authored
The "auto" option doesn't enable any acceleration at all until 4.3, so this is quite safe - only people who intentionally enable it will get it. (And for older vlc.git versions, such as the vlc-2.2 branch, it won't have any effect.) Signed-off-by:
Martin Storsjö <martin@martin.st>
-
Thomas Guillem authored
The new iomx_hwbuffer.c file currently only exists in the master branch in vlc.git, not in the 2.2 release branch. Linking the libiomx-*.so to libcutils/liblog/libui even though we don't use any functions from those (if iomx_hwbuffer.c isn't found) doesn't hurt, and some of them are necessary even if iomx_hwbuffer.c isn't built, since we still provide ANDROID_API to iomx.cpp (which causes the HAS_USE_BUFFER codepaths to be built). Signed-off-by:
Martin Storsjö <martin@martin.st>
-
Thomas Guillem authored
Signed-off-by:
Martin Storsjö <martin@martin.st>
-
Thomas Guillem authored
This allows building most of the iomx hwbuffer code for gingerbread as well, even though it's not hooked up completely yet. Signed-off-by:
Martin Storsjö <martin@martin.st>
-
- 24 Jul, 2014 6 commits
-
-
Martin Storsjö authored
For the normal, non-direct-rendering iomx, the headers from ICS work on 4.1 and 4.2, but not on 4.3. Signed-off-by:
Martin Storsjö <martin@martin.st>
-
Thomas Guillem authored
Only libiomx libs that are specified by LIBVLC_LIBS will be built. Signed-off-by:
Martin Storsjö <martin@martin.st>
-
Edward Wang authored
It is not accurate but it works as a last resort when real frequency is unavailable.
-
Edward Wang authored
-
flx42 authored
-
flx42 authored
-
- 22 Jul, 2014 2 commits
-
-
Edward Wang authored
-
flx42 authored
-
- 20 Jul, 2014 1 commit
-
-
Jean-Baptiste Kempf authored
Close #11794
-
- 18 Jul, 2014 16 commits
-
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
And catch all exceptions
-
Jean-Baptiste Kempf authored
And yes, the try{} catch in the finally is necessary, because .close() can resend an IOException.
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
-
Jean-Baptiste Kempf authored
So, the intent can be null, but the default implementation cannot cope with it?
-
Edward Wang authored
-
Edward Wang authored
-