On Thu, 04 Oct 2012 00:12:38 +0200, Luca Barbato lu_z...@gentoo.org wrote:
On 10/03/2012 11:37 PM, Martin Storsjö wrote:
Before, the chained muxer reused the AVStreams array from
the outer muxer, which made it impossible to use the proper
public functions (such as av_write_frame) when
On Thu, 4 Oct 2012, Anton Khirnov wrote:
On Thu, 04 Oct 2012 00:12:38 +0200, Luca Barbato lu_z...@gentoo.org wrote:
On 10/03/2012 11:37 PM, Martin Storsjö wrote:
Before, the chained muxer reused the AVStreams array from
the outer muxer, which made it impossible to use the proper
public
On Wed, Oct 03, 2012 at 10:11:37PM -0400, Sean McGovern wrote:
On Wed, Oct 3, 2012 at 7:06 PM, Diego Biurrun di...@biurrun.de wrote:
--- a/libavutil/x86/cpu.c
+++ b/libavutil/x86/cpu.c
@@ -108,10 +108,7 @@ int ff_get_cpu_flags_x86(void)
-cpuid(0, max_std_level, ebx, ecx, edx);
-
On Thu, Oct 04, 2012 at 12:44:12AM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
On Thu, Oct 04, 2012 at 12:27:09AM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
--- a/libavutil/x86/cpu.c
+++ b/libavutil/x86/cpu.c
@@ -108,10 +108,7 @@ int
On 10/04/2012 07:29 AM, Anton Khirnov wrote:
On Thu, 04 Oct 2012 00:21:12 +0200, Luca Barbato lu_z...@gentoo.org wrote:
On 10/03/2012 10:56 PM, Anton Khirnov wrote:
Right now it incorrectly assumes that the frames are sent in the proper
order, which worked with old ffmpeg and avconv versions
Diego Biurrun di...@biurrun.de writes:
On Wed, Oct 03, 2012 at 10:11:37PM -0400, Sean McGovern wrote:
On Wed, Oct 3, 2012 at 7:06 PM, Diego Biurrun di...@biurrun.de wrote:
--- a/libavutil/x86/cpu.c
+++ b/libavutil/x86/cpu.c
@@ -108,10 +108,7 @@ int ff_get_cpu_flags_x86(void)
-
On 10/02/2012 11:28 PM, Luca Barbato wrote:
---
Ping...
doc/formats/nut.texi | 130
+++
1 file changed, 130 insertions(+)
create mode 100644 doc/formats/nut.texi
diff --git a/doc/formats/nut.texi b/doc/formats/nut.texi
new file mode
On 10/02/2012 11:21 AM, Luca Barbato wrote:
---
libavcodec/raw.c | 2 ++
libavformat/nut.c | 2 ++
2 files changed, 4 insertions(+)
Ping.
___
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-devel
On Thu, 4 Oct 2012 11:43:54 +0300 (EEST), Martin Storsjö mar...@martin.st
wrote:
On Thu, 4 Oct 2012, Anton Khirnov wrote:
On Thu, 04 Oct 2012 00:12:38 +0200, Luca Barbato lu_z...@gentoo.org wrote:
On 10/03/2012 11:37 PM, Martin Storsjö wrote:
Before, the chained muxer reused the
On Thu, 4 Oct 2012 00:37:22 +0300, Martin Storsjö mar...@martin.st wrote:
This makes sure new inline headers are emitted when the next
packet is written. This allows segmenting mpegts without calling
write_header/write_trailer (nor freeing/reiniting the muxer)
for each segment.
---
On Thu, Oct 4, 2012 at 6:14 AM, Måns Rullgård m...@mansr.com wrote:
Diego Biurrun di...@biurrun.de writes:
On Wed, Oct 03, 2012 at 10:11:37PM -0400, Sean McGovern wrote:
On Wed, Oct 3, 2012 at 7:06 PM, Diego Biurrun di...@biurrun.de wrote:
--- a/libavutil/x86/cpu.c
+++
---
libavformat/segment.c |4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/libavformat/segment.c b/libavformat/segment.c
index 1e928a7..375df58 100644
--- a/libavformat/segment.c
+++ b/libavformat/segment.c
@@ -98,7 +98,7 @@ static int segment_start(AVFormatContext *s,
---
The #define indirection will be reused by the yasm version of cpuid_test,
so I'm adding it here already to avoid code churn.
libavutil/x86/cpu.c | 27 ++-
1 files changed, 18 insertions(+), 9 deletions(-)
diff --git a/libavutil/x86/cpu.c b/libavutil/x86/cpu.c
Now that there is CPU detection in YASM, there will always be one of
inline or external assembly enabled, which obviates the need to fall
back on CPU detection through compiler intrinsics.
---
configure |6 --
libavutil/x86/cpu.c | 39 ---
2
---
This was done for old suncc versions, but these miscompile libav anyway.
libavutil/x86/cpu.c |5 +
1 files changed, 1 insertions(+), 4 deletions(-)
diff --git a/libavutil/x86/cpu.c b/libavutil/x86/cpu.c
index 9acc867..9cf68e4 100644
--- a/libavutil/x86/cpu.c
+++
This allows detecting CPU features with builds that have neither
gcc inline assembly nor the right compiler intrinsics enabled.
---
Changed to implement have_cpuid() instead of emulating the macros
used by the inline/intrinsics versions.
libavutil/x86/Makefile |3 +-
libavutil/x86/cpu.c
ff_get_cpu_flags_x86() requires cpuid(), which is conditionally defined
elsewhere in the file. Surrounding the function body with ifdefs allows
building even when cpuid is not defined. An empty cpuflags mask is
returned in this case.
---
libavutil/x86/cpu.c |5 +
1 files changed, 5
Diego Biurrun di...@biurrun.de writes:
---
This was done for old suncc versions, but these miscompile libav anyway.
libavutil/x86/cpu.c |5 +
1 files changed, 1 insertions(+), 4 deletions(-)
diff --git a/libavutil/x86/cpu.c b/libavutil/x86/cpu.c
index 9acc867..9cf68e4 100644
Diego Biurrun di...@biurrun.de writes:
---
The #define indirection will be reused by the yasm version of cpuid_test,
so I'm adding it here already to avoid code churn.
libavutil/x86/cpu.c | 27 ++-
1 files changed, 18 insertions(+), 9 deletions(-)
diff --git
Diego Biurrun di...@biurrun.de writes:
Now that there is CPU detection in YASM, there will always be one of
inline or external assembly enabled, which obviates the need to fall
back on CPU detection through compiler intrinsics.
---
configure |6 --
libavutil/x86/cpu.c |
Diego Biurrun di...@biurrun.de writes:
ff_get_cpu_flags_x86() requires cpuid(), which is conditionally defined
elsewhere in the file. Surrounding the function body with ifdefs allows
building even when cpuid is not defined. An empty cpuflags mask is
returned in this case.
---
Ok
___
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-devel
ff_get_cpu_flags_x86() requires cpuid(), which is conditionally defined
elsewhere in the file. Surrounding the function body with ifdefs allows
building even when cpuid is not defined. An empty cpuflags mask is
returned in this case.
---
libavutil/x86/cpu.c |5 +
1 files changed, 5
Diego Biurrun di...@biurrun.de writes:
ff_get_cpu_flags_x86() requires cpuid(), which is conditionally defined
elsewhere in the file. Surrounding the function body with ifdefs allows
building even when cpuid is not defined. An empty cpuflags mask is
returned in this case.
---
On Sun, Sep 2, 2012 at 7:03 PM, Luca Barbato lu_z...@gentoo.org wrote:
On systems supported have the default lock manager enabled by default.
---
Now with free and _cp - _cb typo fixed.
Yet to be discussed if/how it could be of use.
libavcodec/pthread.c | 27 +++
On Thu, Oct 04, 2012 at 02:04:27PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
--- a/libavutil/x86/cpu.c
+++ b/libavutil/x86/cpu.c
@@ -96,15 +113,7 @@ int ff_get_cpu_flags_x86(void)
#if ARCH_X86_32
-x86_reg a, c;
-
-/* Check if CPUID is supported by
Diego Biurrun di...@biurrun.de writes:
On Thu, Oct 04, 2012 at 02:04:27PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
--- a/libavutil/x86/cpu.c
+++ b/libavutil/x86/cpu.c
@@ -96,15 +113,7 @@ int ff_get_cpu_flags_x86(void)
#if ARCH_X86_32
-x86_reg a, c;
-
On Thu, Oct 04, 2012 at 03:30:17PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
On Thu, Oct 04, 2012 at 02:04:27PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
--- a/libavutil/x86/cpu.c
+++ b/libavutil/x86/cpu.c
@@ -96,15 +113,7 @@ int
Diego Biurrun di...@biurrun.de writes:
On Thu, Oct 04, 2012 at 03:30:17PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
On Thu, Oct 04, 2012 at 02:04:27PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
--- a/libavutil/x86/cpu.c
+++
On Thu, Oct 04, 2012 at 03:36:40PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
On Thu, Oct 04, 2012 at 03:30:17PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
On Thu, Oct 04, 2012 at 02:04:27PM +0100, Måns Rullgård wrote:
Diego Biurrun
Diego Biurrun di...@biurrun.de writes:
On Thu, Oct 04, 2012 at 03:36:40PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
On Thu, Oct 04, 2012 at 03:30:17PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
On Thu, Oct 04, 2012 at 02:04:27PM +0100,
ff_get_cpu_flags_x86() requires cpuid(), which is conditionally defined
elsewhere in the file. Surrounding the function body with ifdefs allows
building even when cpuid is not defined. An empty cpuflags mask is
returned in this case.
---
libavutil/x86/cpu.c |5 +
1 files changed, 5
Diego Biurrun di...@biurrun.de writes:
ff_get_cpu_flags_x86() requires cpuid(), which is conditionally defined
elsewhere in the file. Surrounding the function body with ifdefs allows
building even when cpuid is not defined. An empty cpuflags mask is
returned in this case.
---
---
libavutil/x86/cpu.c | 34 +++---
1 files changed, 23 insertions(+), 11 deletions(-)
diff --git a/libavutil/x86/cpu.c b/libavutil/x86/cpu.c
index 9cf68e4..bf1ed0b 100644
--- a/libavutil/x86/cpu.c
+++ b/libavutil/x86/cpu.c
@@ -86,16 +86,16 @@
#endif /*
Diego Biurrun di...@biurrun.de writes:
---
libavutil/x86/cpu.c | 34 +++---
1 files changed, 23 insertions(+), 11 deletions(-)
diff --git a/libavutil/x86/cpu.c b/libavutil/x86/cpu.c
index 9cf68e4..bf1ed0b 100644
--- a/libavutil/x86/cpu.c
+++
On Wed, Sep 26, 2012 at 07:59:36PM +0200, Diego Biurrun wrote:
On Tue, Sep 11, 2012 at 10:30:44AM +0200, Diego Biurrun wrote:
On Fri, Sep 07, 2012 at 11:17:38AM +0200, Diego Biurrun wrote:
On Fri, Aug 24, 2012 at 07:06:26PM +0200, Diego Biurrun wrote:
A new hidden config variable is
On 02/10/2012 5:21 AM, Luca Barbato wrote:
---
libavcodec/raw.c | 2 ++
libavformat/nut.c | 2 ++
2 files changed, 4 insertions(+)
Looks OK. Matches what's in FFmpeg's nut code too.
- Derek
___
libav-devel mailing list
libav-devel@libav.org
ff_get_cpu_flags_x86() requires cpuid(), which is conditionally defined
elsewhere in the file. Surrounding the function body with ifdefs allows
building even when cpuid is not defined. An empty cpuflags mask is
returned in this case.
---
libavutil/x86/cpu.c |5 +
1 files changed, 5
---
libavutil/x86/cpu.c | 32 +---
1 files changed, 21 insertions(+), 11 deletions(-)
diff --git a/libavutil/x86/cpu.c b/libavutil/x86/cpu.c
index 9cf68e4..5de6014 100644
--- a/libavutil/x86/cpu.c
+++ b/libavutil/x86/cpu.c
@@ -86,16 +86,14 @@
#endif /*
Diego Biurrun di...@biurrun.de writes:
---
libavutil/x86/cpu.c | 32 +---
1 files changed, 21 insertions(+), 11 deletions(-)
diff --git a/libavutil/x86/cpu.c b/libavutil/x86/cpu.c
index 9cf68e4..5de6014 100644
--- a/libavutil/x86/cpu.c
+++
avplay.c:3015:20: warning: assignment discards ‘const’ qualifier from pointer
target type
---
avplay.c |3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/avplay.c b/avplay.c
index 5ba8155..cd8c386 100644
--- a/avplay.c
+++ b/avplay.c
@@ -2959,6 +2959,7 @@ static void
Diego Biurrun di...@biurrun.de writes:
avplay.c:3015:20: warning: assignment discards ‘const’ qualifier from pointer
target type
---
avplay.c |3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/avplay.c b/avplay.c
index 5ba8155..cd8c386 100644
--- a/avplay.c
+++
Signed-off-by: Mans Rullgard m...@mansr.com
---
libavcodec/ppc/fmtconvert_altivec.c | 25 ++---
1 file changed, 18 insertions(+), 7 deletions(-)
diff --git a/libavcodec/ppc/fmtconvert_altivec.c
b/libavcodec/ppc/fmtconvert_altivec.c
index db2c258..129891a 100644
---
With all the VLAs gone, make sure they never come back.
Signed-off-by: Mans Rullgard m...@mansr.com
---
configure | 1 +
1 file changed, 1 insertion(+)
diff --git a/configure b/configure
index 760b747..7032f98 100755
--- a/configure
+++ b/configure
@@ -3522,6 +3522,7 @@ elif enabled gcc; then
Instead of using an evil VLA, fall back to C version when edge
emulation is needed. MPEG4 GMC is a rarely used fringe feature
so the speed loss is an acceptable cost for cleaner code.
Signed-off-by: Mans Rullgard m...@mansr.com
---
libavcodec/x86/dsputil_mmx.c | 58
This gets rid of the variable-length scratch buffer by filtering 16
pixels at a time and writing directly to the destination. The extra
loads this requires to load the source values are compensated by not
doing a round-trip to memory before shifting.
Signed-off-by: Mans Rullgard m...@mansr.com
On Thu, Oct 04, 2012 at 05:43:57PM +0100, Mans Rullgard wrote:
With all the VLAs gone, make sure they never come back.
Fine with me once the rest of the set is in.
Diego
___
libav-devel mailing list
libav-devel@libav.org
On Thu, Oct 04, 2012 at 05:36:58PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
avplay.c:3015:20: warning: assignment discards ‘const’ qualifier from
pointer target type
--- a/avplay.c
+++ b/avplay.c
@@ -2959,6 +2959,7 @@ static void opt_input_file(void *optctx,
Diego Biurrun di...@biurrun.de writes:
On Thu, Oct 04, 2012 at 05:36:58PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
avplay.c:3015:20: warning: assignment discards ‘const’ qualifier from
pointer target type
--- a/avplay.c
+++ b/avplay.c
@@ -2959,6 +2959,7 @@
On Thu, Oct 04, 2012 at 06:03:57PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
On Thu, Oct 04, 2012 at 05:36:58PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
avplay.c:3015:20: warning: assignment discards ‘const’ qualifier from
pointer
Diego Biurrun di...@biurrun.de writes:
This allows detecting CPU features with builds that have neither
gcc inline assembly nor the right compiler intrinsics enabled.
---
Changed to implement have_cpuid() instead of emulating the macros
used by the inline/intrinsics versions.
Diego Biurrun di...@biurrun.de writes:
On Thu, Oct 04, 2012 at 06:03:57PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
On Thu, Oct 04, 2012 at 05:36:58PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
avplay.c:3015:20: warning: assignment
On 04/10/2012 12:43 PM, Mans Rullgard wrote:
This gets rid of the variable-length scratch buffer by filtering 16
pixels at a time and writing directly to the destination. The extra
loads this requires to load the source values are compensated by not
doing a round-trip to memory before
---
libavcodec/Makefile |7 +
libavcodec/arm/mathops.h| 24 ++
libavcodec/inverse.c|1 -
libavutil/inverse.c = libavcodec/mathops.c | 14 --
libavcodec/mathops.h|
---
libavutil/libavutil.v |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/libavutil/libavutil.v b/libavutil/libavutil.v
index eb16ae1..e9f04cb 100644
--- a/libavutil/libavutil.v
+++ b/libavutil/libavutil.v
@@ -1,4 +1,4 @@
LIBAVUTIL_$MAJOR {
-global: av*; ff_*;
+
---
libswscale/libswscale.v |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/libswscale/libswscale.v b/libswscale/libswscale.v
index c68f153..8b9a96a 100644
--- a/libswscale/libswscale.v
+++ b/libswscale/libswscale.v
@@ -1,4 +1,4 @@
LIBSWSCALE_$MAJOR {
-global:
---
libavcodec/aacps.c|2 +-
libavcodec/bmpenc.c |2 +-
libavcodec/rawdec.c |2 +-
libavcodec/tiertexseqv.c |2 +-
libavcodec/utils.c|2 +-
libavcodec/xsubdec.c |2 +-
libavcodec/xsubenc.c |2 +-
libavformat/matroskadec.c |4
On 04/10/2012 12:43 PM, Mans Rullgard wrote:
Instead of using an evil VLA, fall back to C version when edge
emulation is needed. MPEG4 GMC is a rarely used fringe feature
so the speed loss is an acceptable cost for cleaner code.
While, IIRC, 'Scene' rules ban GMC usage, I've seen plenty of
Derek Buitenhuis derek.buitenh...@gmail.com writes:
On 04/10/2012 12:43 PM, Mans Rullgard wrote:
This gets rid of the variable-length scratch buffer by filtering 16
pixels at a time and writing directly to the destination. The extra
loads this requires to load the source values are
Diego Biurrun di...@biurrun.de writes:
---
libavcodec/Makefile |7 +
libavcodec/arm/mathops.h| 24 ++
libavcodec/inverse.c|1 -
libavutil/inverse.c = libavcodec/mathops.c | 14 --
Diego Biurrun di...@biurrun.de writes:
---
libswscale/libswscale.v |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/libswscale/libswscale.v b/libswscale/libswscale.v
index c68f153..8b9a96a 100644
--- a/libswscale/libswscale.v
+++ b/libswscale/libswscale.v
@@ -1,4
On 04/10/2012 1:44 PM, Måns Rullgård wrote:
Why are they visible now?
The file was never touched after its creation... by you.
Perhaps you should ask yourself :D
- Derek
___
libav-devel mailing list
libav-devel@libav.org
Derek Buitenhuis derek.buitenh...@gmail.com writes:
On 04/10/2012 1:44 PM, Måns Rullgård wrote:
Why are they visible now?
The file was never touched after its creation... by you.
I may have committed the file, but I did not create it. Now stop being
snide.
--
Måns Rullgård
m...@mansr.com
On Thu, Oct 04, 2012 at 06:44:06PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
--- a/libswscale/libswscale.v
+++ b/libswscale/libswscale.v
@@ -1,4 +1,4 @@
LIBSWSCALE_$MAJOR {
-global: swscale_*; sws_*; ff_*;
+global: swscale_*; sws_*;
On Thu, Oct 04, 2012 at 06:43:16PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
---
libavcodec/Makefile |7 +
libavcodec/arm/mathops.h| 24 ++
libavcodec/inverse.c|1
From: Baptiste Coudurier baptiste.coudur...@gmail.com
Setting the average bitrate to 0 in esds makes the produced files
not compatible with certain players.
According to 14496-1:2010 7.2.6.6 DecoderConfigDescriptor:
avgBitrate – is the average bitrate in bits per second of this
elementary
On 10/04/2012 07:14 PM, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
On Thu, Oct 04, 2012 at 06:03:57PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
On Thu, Oct 04, 2012 at 05:36:58PM +0100, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
On 10/04/2012 06:43 PM, Mans Rullgard wrote:
Signed-off-by: Mans Rullgard m...@mansr.com
---
libavcodec/ppc/fmtconvert_altivec.c | 25 ++---
1 file changed, 18 insertions(+), 7 deletions(-)
Ok.
___
libav-devel mailing list
On 10/04/2012 07:38 PM, Måns Rullgård wrote:
Derek Buitenhuis derek.buitenh...@gmail.com writes:
On 04/10/2012 12:43 PM, Mans Rullgard wrote:
This gets rid of the variable-length scratch buffer by filtering 16
pixels at a time and writing directly to the destination. The extra
loads this
Luca Barbato lu_z...@gentoo.org writes:
From: Baptiste Coudurier baptiste.coudur...@gmail.com
Setting the average bitrate to 0 in esds makes the produced files
not compatible with certain players.
According to 14496-1:2010 7.2.6.6 DecoderConfigDescriptor:
avgBitrate – is the average
This gets rid of the variable-length scratch buffer by filtering 16
pixels at a time and writing directly to the destination. The extra
loads this requires to load the source values are compensated by not
doing a round-trip to memory before shifting.
Signed-off-by: Mans Rullgard m...@mansr.com
On 10/04/2012 08:52 PM, Måns Rullgård wrote:
Luca Barbato lu_z...@gentoo.org writes:
From: Baptiste Coudurier baptiste.coudur...@gmail.com
Setting the average bitrate to 0 in esds makes the produced files
not compatible with certain players.
According to 14496-1:2010 7.2.6.6
---
libavcodec/bmpenc.c |2 +-
libavcodec/rawdec.c |2 +-
libavcodec/utils.c |2 +-
libavutil/imgutils.c |4 ++--
libavutil/imgutils.h |2 +-
5 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/libavcodec/bmpenc.c b/libavcodec/bmpenc.c
index 37cf6e1..791d846
The table is so small that the space gain is not worth the
performance overhead of cross-library access.
---
libavcodec/aacps.c |1 +
libavcodec/mathops.c|2 ++
libavcodec/mathops.h|1 +
libavutil/Makefile |1 +
libavutil/log2_tab.c| 30
On 10/04/2012 01:46 AM, Luca Barbato wrote:
From: Ronald S. Bultje rsbul...@gmail.com
The variable is copied to subsequent threads at the same time, so this
may cause wrong ref_count[] values to be copied to subsequent threads.
This bug was found using TSAN.
Signed-off-by: Luca Barbato
The table is so small that the space gain is not worth the
performance overhead of cross-library access.
---
Now with each duplicated table in a separate file.
libavcodec/Makefile |1 +
libavcodec/log2_tab.c |1 +
libavutil/Makefile |1 +
libavutil/log2_tab.c| 30
On Tue, 20 Sep 2011, Måns Rullgård wrote:
Diego Biurrun di...@biurrun.de writes:
If the build_only variable is set in the configuration file, the FATE
client will skip running tests and just compile all targets.
---
tests/fate.sh |1 +
1 files changed, 1 insertions(+), 0 deletions(-)
---
Now making the function public instead of marking it avpriv_.
libavcodec/bmpenc.c |2 +-
libavcodec/rawdec.c |2 +-
libavcodec/utils.c |2 +-
libavutil/imgutils.c |4 ++--
libavutil/imgutils.h |2 +-
5 files changed, 6 insertions(+), 6 deletions(-)
diff --git
Diego Biurrun di...@biurrun.de writes:
---
Now making the function public instead of marking it avpriv_.
I'm not at all convinced this function needs to be exposed at all.
It probably shouldn't even exist.
--
Måns Rullgård
m...@mansr.com
___
On Thu, Sep 13, 2012 at 08:18:56PM +0100, Mans Rullgard wrote:
Some compilers are extra strict about register usage in main(),
disallowing ebp in inline asm there while allowing it elsewhere.
This change makes the test better reflect actual usage.
--- a/configure
+++ b/configure
@@ -865,6
Diego Biurrun di...@biurrun.de writes:
On Thu, Sep 13, 2012 at 08:18:56PM +0100, Mans Rullgard wrote:
Some compilers are extra strict about register usage in main(),
disallowing ebp in inline asm there while allowing it elsewhere.
This change makes the test better reflect actual usage.
---
Derek Buitenhuis derek.buitenh...@gmail.com writes:
On 04/10/2012 12:43 PM, Mans Rullgard wrote:
Instead of using an evil VLA, fall back to C version when edge
emulation is needed. MPEG4 GMC is a rarely used fringe feature
so the speed loss is an acceptable cost for cleaner code.
While,
On 04/10/2012 5:42 PM, Måns Rullgård wrote:
The gmc function itself is only mmx. The old code has multiple wrappers
using mmx or sse versions of the edge emulation function. Since the
patch changes it to fall back to C code when edge emulation is needed,
these wrappers are gone, leaving only
This updates the declaration to match the actual size of the
table.
---
This is a troll-patch in the hope that someone will suggest a better
home for that declaration.
---
libavcodec/jpegls.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavcodec/jpegls.h
Hi,
On Thu, Oct 4, 2012 at 12:59 PM, Luca Barbato lu_z...@gentoo.org wrote:
On 10/04/2012 01:46 AM, Luca Barbato wrote:
From: Ronald S. Bultje rsbul...@gmail.com
The variable is copied to subsequent threads at the same time, so this
may cause wrong ref_count[] values to be copied to
On Thu, Oct 4, 2012 at 4:27 PM, Ronald S. Bultje rsbul...@gmail.com wrote:
Hi,
On Thu, Oct 4, 2012 at 12:59 PM, Luca Barbato lu_z...@gentoo.org wrote:
On 10/04/2012 01:46 AM, Luca Barbato wrote:
From: Ronald S. Bultje rsbul...@gmail.com
The variable is copied to subsequent threads at the
On 10/05/2012 01:29 AM, Ronald S. Bultje wrote:
On Thu, Oct 4, 2012 at 4:27 PM, Ronald S. Bultje rsbul...@gmail.com wrote:
Oh right you mean for performance?
fate-h264-conformance-cama2_vtc_b
fate-h264-conformance-cama3_vtc_b
should be ok.
I'd push your patch + my style nits soon, I'll
Segment set AVFMT_GLOBALHEADER since it is assumed that global headers might
be used, it interacts badly with x264 assumption on that flag and mpegts.
The two patches aim to partially workaround another option is adding a
AVFMT_ANYHEADER or making the flag not part of the immutable format
Currently this option is bound to the container format requesting global
headers or not, in a number of situation global headers and and repeated
headers are not mutually exclusive.
---
libavcodec/libx264.c | 5 +
1 file changed, 5 insertions(+)
diff --git a/libavcodec/libx264.c
avconv provides an option to set it externally and there isn't a way for
the outer demuxer to report the inner demuxer flag.
Solves a bad interaction between mpegts and x264, but requires
additional setting for the user.
---
libavformat/segment.c | 2 +-
1 file changed, 1 insertion(+), 1
It has been superseded by lavr.
---
libavcodec/avcodec.h | 10 ++
libavcodec/resample.c |4
libavcodec/resample2.c |4
libavcodec/version.h |3 +++
4 files changed, 21 insertions(+)
diff --git a/libavcodec/avcodec.h b/libavcodec/avcodec.h
index
This is more consistent with what the rest of Libav does.
This breaks API.
---
avplay.c|4 ++--
libavfilter/af_asyncts.c|8
libavfilter/af_resample.c |6 +++---
libavresample/audio_data.c |2 +-
libavresample/audio_data.h |
---
Changelog |2 ++
RELEASE |2 +-
2 files changed, 3 insertions(+), 1 deletion(-)
diff --git a/Changelog b/Changelog
index cc5eaeb..935cd49 100644
--- a/Changelog
+++ b/Changelog
@@ -3,6 +3,8 @@ releases are sorted from youngest to oldest.
version next:
+version 9_beta1:
+
-
---
doc/RELEASE_NOTES | 109 ++---
1 file changed, 61 insertions(+), 48 deletions(-)
diff --git a/doc/RELEASE_NOTES b/doc/RELEASE_NOTES
index 8c642b9..6acb230 100644
--- a/doc/RELEASE_NOTES
+++ b/doc/RELEASE_NOTES
@@ -1,29 +1,53 @@
Release Notes
On 10/05/2012 01:27 AM, Anton Khirnov wrote:
This is more consistent with what the rest of Libav does.
This breaks API.
---
avplay.c|4 ++--
libavfilter/af_asyncts.c|8
libavfilter/af_resample.c |6 +++---
On 10/05/2012 01:27 AM, Anton Khirnov wrote:
---
doc/APIchanges |5 +
libavresample/version.h |4 ++--
2 files changed, 7 insertions(+), 2 deletions(-)
diff --git a/doc/APIchanges b/doc/APIchanges
index 4558ca7..314622d 100644
--- a/doc/APIchanges
+++
On Mon, 1 Oct 2012 01:42:24 -0400, Justin Ruggles justin.rugg...@gmail.com
wrote:
---
alternative patch
tests/lavf-regression.sh |4
tests/ref/lavf/ffm |3 ---
2 files changed, 0 insertions(+), 7 deletions(-)
delete mode 100644 tests/ref/lavf/ffm
diff --git
97 matches
Mail list logo