On Wed, 22 Aug 2012 12:05:45 -0700, Alex Converse
wrote:
> ---
> Changelog|1 +
> doc/filters.texi | 33 +
> libavfilter/Makefile |1 +
> libavfilter/af_alength.c | 178
> ++
> libavfilter/allfilters.c |
On Wed, 22 Aug 2012, daniel.d.k...@gmail.com wrote:
> +; void pixels8_l2_mmx2(uint8_t *dst, uint8_t *src1, uint8_t *src2, int
> dstStride, int src1Stride, int h)
> +%macro PIXELS8_L2 1
> +%define OP op_%1
> +cglobal %1_pixels8_l2, 6,6
> +test r5d, 1
> +je .loop
> +movam0, [r1]
>
On Wed, Aug 22, 2012 at 4:04 PM, Diego Biurrun wrote:
> On Wed, Aug 22, 2012 at 08:56:46PM +0200, Luca Barbato wrote:
>> On 8/22/12 8:45 PM, Måns Rullgård wrote:
>> >Daniel Kang writes:
>> >
>> >>Why drop this:
>> >>As mentioned in the commit message, 3dnow is deprecated by AMD. I see
>> >>no rea
On Wed, Aug 22, 2012 at 08:56:46PM +0200, Luca Barbato wrote:
> On 8/22/12 8:45 PM, Måns Rullgård wrote:
> >Daniel Kang writes:
> >
> >>Why drop this:
> >>As mentioned in the commit message, 3dnow is deprecated by AMD. I see
> >>no reason to support a deprecated instruction set. Also, it's a
> >>m
On Wed, Aug 22, 2012 at 03:01:06PM -0400, Derek Buitenhuis wrote:
>
> --- a/tests/fate/lossless-video.mak
> +++ b/tests/fate/lossless-video.mak
> @@ -36,3 +36,12 @@ fate-zlib: CMD = framecrc -i $(SAMPLES)/lcl/zlib-1frame.avi
>
> FATE_SAMPLES_AVCONV += fate-zerocodec
> fate-zerocodec: CMD = fra
A "-f avi" was left over from the old file hashing test base. This
should fix the failing tests on the valgrinding fate machines.
---
tests/fate/utvideo.mak |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/tests/fate/utvideo.mak b/tests/fate/utvideo.mak
index ac67944..2236
---
Changelog|1 +
doc/filters.texi | 33 +
libavfilter/Makefile |1 +
libavfilter/af_alength.c | 178 ++
libavfilter/allfilters.c |1 +
5 files changed, 214 insertions(+), 0 deletions(-)
create mode 10
Signed-off-by: Derek Buitenhuis
---
tests/fate/lossless-video.mak |9 +
tests/ref/fate/cllc-argb |5 +
tests/ref/fate/cllc-rgb | 15 +++
3 files changed, 29 insertions(+)
create mode 100644 tests/ref/fate/cllc-argb
create mode 100644 tests/ref/fate/c
Files are located here:
http://chromashift.org/sample-cllc-argb.avi
http://chromashift.org/sample-cllc-rgb.avi
md5sums:
8fbf9e6d099bf0cbc2ca42593f478a2a *sample-cllc-argb.avi
5d24481938c27685cb05c0b70ec0bce4 *sample-cllc-rgb.avi
Derek Buitenhuis (1):
FATE: Add Canopus Lossless t
On 8/22/12 8:45 PM, Måns Rullgård wrote:
Daniel Kang writes:
Why drop this:
As mentioned in the commit message, 3dnow is deprecated by AMD. I see
no reason to support a deprecated instruction set. Also, it's a
maintenance burden for anyone who makes functional changes to the
code.
It is depr
Daniel Kang writes:
> Why drop this:
> As mentioned in the commit message, 3dnow is deprecated by AMD. I see
> no reason to support a deprecated instruction set. Also, it's a
> maintenance burden for anyone who makes functional changes to the
> code.
It is deprecated on current CPUs, not on the
On Wed, Aug 22, 2012 at 10:19 AM, Diego Biurrun wrote:
> On Wed, Aug 22, 2012 at 10:15:14AM -0700, Daniel Kang wrote:
>> On Wed, Aug 22, 2012 at 9:14 AM, Diego Biurrun wrote:
>> >
>> > On Tue, Aug 21, 2012 at 10:08:02PM -0500, daniel.d.k...@gmail.com wrote:
>> > > From: Daniel Kang
>> > >
>> > >
On Wed, Aug 22, 2012 at 01:40:31PM -0400, Justin Ruggles wrote:
> Fixes crash on x86 due to alignment requirements for w_data in
> lpc_apply_welch_window_sse2().
> ---
> libavcodec/lpc.c | 15 +++
> libavcodec/lpc.h |1 +
> 2 files changed, 8 insertions(+), 8 deletions(-)
>
> di
Fixes crash on x86 due to alignment requirements for w_data in
lpc_apply_welch_window_sse2().
---
libavcodec/lpc.c | 15 +++
libavcodec/lpc.h |1 +
2 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/libavcodec/lpc.c b/libavcodec/lpc.c
index 6ce5dde..cc3ce44 100644
---
On Wed, Aug 22, 2012 at 10:15:14AM -0700, Daniel Kang wrote:
> On Wed, Aug 22, 2012 at 9:14 AM, Diego Biurrun wrote:
> >
> > On Tue, Aug 21, 2012 at 10:08:02PM -0500, daniel.d.k...@gmail.com wrote:
> > > From: Daniel Kang
> > >
> > > The only CPUs that have 3dnow and don't have mmxext are 12 year
On Wed, Aug 22, 2012 at 9:14 AM, Diego Biurrun wrote:
>
> On Tue, Aug 21, 2012 at 10:08:02PM -0500, daniel.d.k...@gmail.com wrote:
> > From: Daniel Kang
> >
> > The only CPUs that have 3dnow and don't have mmxext are 12 years old.
> >
> > Moreover, AMD has deprecated 3dnow.
> > ---
> > libavcode
On Wed, Aug 22, 2012 at 11:35:56AM -0400, Justin Ruggles wrote:
> On 08/04/2012 02:33 PM, Justin Ruggles wrote:
> > Also, do not allow seek-by-byte, as there is no way to find the next packet
> > boundary.
> > ---
> > libavformat/idcin.c | 21 +
> > 1 files changed, 21 insert
On Wed, Aug 22, 2012 at 11:35:31AM -0400, Justin Ruggles wrote:
> On 08/01/2012 05:36 PM, Justin Ruggles wrote:
> > ---
> > libavformat/idcin.c |1 +
> > 1 files changed, 1 insertions(+), 0 deletions(-)
> >
> > diff --git a/libavformat/idcin.c b/libavformat/idcin.c
> > index 0e93a59..638377c
On Wed, Aug 22, 2012 at 11:35:19AM -0400, Justin Ruggles wrote:
> On 08/04/2012 02:33 PM, Justin Ruggles wrote:
> > Also, use 1 / sample_rate for audio stream time_base.
> > ---
> > libavformat/idcin.c | 15 +++
> > 1 files changed, 7 insertions(+), 8 deletions(-)
>
> ping.
LGTM
__
On Wed, Aug 22, 2012 at 11:36:48AM -0400, Justin Ruggles wrote:
> On 08/04/2012 02:34 PM, Justin Ruggles wrote:
> > Add some additional checks for EOF and print error messages on an incomplete
> > header or packet.
> > ---
> > libavformat/idcin.c | 37 ++---
> > 1
On 08/21/2012 11:45 AM, Justin Ruggles wrote:
> ---
> libavcodec/vorbisenc.c | 20 +---
> 1 files changed, 9 insertions(+), 11 deletions(-)
ping
-Justin
___
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailm
On Wed, Aug 22, 2012 at 11:36:23AM -0400, Justin Ruggles wrote:
> On 08/04/2012 02:33 PM, Justin Ruggles wrote:
> > chunk_size is unsigned 32-bit, but av_get_packet() takes a signed int as the
> > packet size.
> > ---
> > libavformat/idcin.c |4
> > 1 files changed, 4 insertions(+), 0 del
---
libavcodec/libmp3lame.c | 105 +-
1 files changed, 39 insertions(+), 66 deletions(-)
diff --git a/libavcodec/libmp3lame.c b/libavcodec/libmp3lame.c
index d754f46..b09b61c 100644
--- a/libavcodec/libmp3lame.c
+++ b/libavcodec/libmp3lame.c
@@ -33,6 +
On Wed, Aug 22, 2012 at 12:27:52PM -0400, Justin Ruggles wrote:
> ---
> tests/Makefile |1 +
> tests/fate/alac.mak | 15 +++
> 2 files changed, 16 insertions(+), 0 deletions(-)
> create mode 100644 tests/fate/alac.mak
LGTM
Diego
___
Hi,
On Tue, Aug 21, 2012 at 8:08 PM, wrote:
> From: Daniel Kang
>
> The only CPUs that have 3dnow and don't have mmxext are 12 years old.
>
> Moreover, AMD has deprecated 3dnow.
> ---
> libavcodec/x86/dsputil_mmx.c | 142
> +
> libavcodec/x86/dsputil_
---
tests/Makefile |1 +
tests/fate/alac.mak | 15 +++
2 files changed, 16 insertions(+), 0 deletions(-)
create mode 100644 tests/fate/alac.mak
diff --git a/tests/Makefile b/tests/Makefile
index 0b11102..2af9a91 100644
--- a/tests/Makefile
+++ b/tests/Makefile
@@ -26,6 +26
On 08/13/2012 02:53 PM, Diego Biurrun wrote:
> ---
> libavcodec/mpegvideo.c |4 ++--
> libavcodec/mpegvideo.h |2 +-
> libavcodec/x86/Makefile |2 +-
> libavcodec/x86/{mpegvideo_mmx.c => mpegvideo.c} |2 +-
>
On Tue, Aug 21, 2012 at 10:08:02PM -0500, daniel.d.k...@gmail.com wrote:
> From: Daniel Kang
>
> The only CPUs that have 3dnow and don't have mmxext are 12 years old.
>
> Moreover, AMD has deprecated 3dnow.
> ---
> libavcodec/x86/dsputil_mmx.c | 142
> +---
On Wed, Aug 22, 2012 at 10:43:46AM -0500, daniel.d.k...@gmail.com wrote:
> From: Daniel Kang
>
> --- /dev/null
> +++ b/libavcodec/x86/h264_qpel.asm
> @@ -0,0 +1,852 @@
> +
> +INIT_MMX mmx2
> +PIXELS8_L2_SHIFT5 put
> +PIXELS8_L2_SHIFT5 avg
> +
> +
> +
> +
> +
> +%if ARCH_X86_64 && HAVE_SSSE3
> +%m
On Wed, Aug 22, 2012 at 11:43:26AM -0400, Daniel Kang wrote:
> On Wed, Aug 22, 2012 at 5:01 AM, Diego Biurrun wrote:
> > On Tue, Aug 21, 2012 at 10:08:03PM -0500, daniel.d.k...@gmail.com wrote:
> > > From: Daniel Kang
> > >
> > > +%if ARCH_X86_64
> > > +; Is there a has ssse3 flag?
> >
> > Yes of
Diego Biurrun writes:
> On Tue, Aug 21, 2012 at 12:06:25AM +0100, Måns Rullgård wrote:
>> Diego Biurrun writes:
>> > --- a/libavcodec/Makefile
>> > +++ b/libavcodec/Makefile
>> > @@ -33,6 +33,9 @@ OBJS = allcodecs.o
>> >\
>> > # parts neede
On Tue, Aug 21, 2012 at 12:06:25AM +0100, Måns Rullgård wrote:
> Diego Biurrun writes:
> > --- a/libavcodec/Makefile
> > +++ b/libavcodec/Makefile
> > @@ -33,6 +33,9 @@ OBJS = allcodecs.o
> > \
> > # parts needed for many different codecs
> >
On Fri, Aug 17, 2012 at 12:13:25AM +0200, Diego Biurrun wrote:
> On Mon, Aug 13, 2012 at 02:53:34PM +0200, Diego Biurrun wrote:
> > ---
> > libavcodec/mpegvideo.c |4 ++--
> > libavcodec/mpegvideo.h |2 +-
> > libavcodec/x86/Makefile
On Fri, Aug 17, 2012 at 12:13:02AM +0200, Diego Biurrun wrote:
> On Mon, Aug 13, 2012 at 02:53:33PM +0200, Diego Biurrun wrote:
> > ---
> >
> > For now this is only done for x86, other arches will follow after
> > first review round.
> >
> > libavcodec/mpegvideo.h |
---
libavcodec/dnxhdenc.c |5 ++---
libavcodec/dnxhdenc.h |2 +-
libavcodec/x86/Makefile|2 +-
libavcodec/x86/{dnxhd_mmx.c => dnxhdenc.c} |2 +-
4 files changed, 5 insertions(+), 6 deletions(-)
rename libavcodec/x86/{dn
On Wed, Aug 22, 2012 at 5:01 AM, Diego Biurrun wrote:
>
> On Tue, Aug 21, 2012 at 10:08:03PM -0500, daniel.d.k...@gmail.com wrote:
> > From: Daniel Kang
> >
> > ---
> > libavcodec/x86/Makefile |3 +-
> > libavcodec/x86/dsputil.asm| 138
> > libavcodec/
On 08/04/2012 02:34 PM, Justin Ruggles wrote:
> Add some additional checks for EOF and print error messages on an incomplete
> header or packet.
> ---
> libavformat/idcin.c | 37 ++---
> 1 files changed, 30 insertions(+), 7 deletions(-)
ping.
-Justin
__
---
libavcodec/libschroedingerenc.c | 13 +
1 file changed, 13 insertions(+)
diff --git a/libavcodec/libschroedingerenc.c b/libavcodec/libschroedingerenc.c
index 865b093..53dbb13 100644
--- a/libavcodec/libschroedingerenc.c
+++ b/libavcodec/libschroedingerenc.c
@@ -74,6 +74,9 @@ typ
On 08/04/2012 02:33 PM, Justin Ruggles wrote:
> chunk_size is unsigned 32-bit, but av_get_packet() takes a signed int as the
> packet size.
> ---
> libavformat/idcin.c |4
> 1 files changed, 4 insertions(+), 0 deletions(-)
>
> diff --git a/libavformat/idcin.c b/libavformat/idcin.c
> inde
On 08/04/2012 02:33 PM, Justin Ruggles wrote:
> Also, do not allow seek-by-byte, as there is no way to find the next packet
> boundary.
> ---
> libavformat/idcin.c | 21 +
> 1 files changed, 21 insertions(+), 0 deletions(-)
ping.
-Justin
___
On 08/01/2012 05:36 PM, Justin Ruggles wrote:
> ---
> libavformat/idcin.c |1 +
> 1 files changed, 1 insertions(+), 0 deletions(-)
>
> diff --git a/libavformat/idcin.c b/libavformat/idcin.c
> index 0e93a59..638377c 100644
> --- a/libavformat/idcin.c
> +++ b/libavformat/idcin.c
> @@ -292,6 +29
On 08/04/2012 02:33 PM, Justin Ruggles wrote:
> Also, use 1 / sample_rate for audio stream time_base.
> ---
> libavformat/idcin.c | 15 +++
> 1 files changed, 7 insertions(+), 8 deletions(-)
ping.
-Justin
___
libav-devel mailing list
lib
On Wed, Aug 22, 2012 at 02:58:41PM +, Babic, Nedeljko wrote:
>
> Anyhow, currently we did not plane resources for this project to support
> two versions of the same code (inline assembly and external asm).
What do you mean by "two versions"? There is no need to support both
inline and extern
On Wed, Aug 22, 2012 at 04:18:54PM +0200, Jordi Ortiz wrote:
> 2012/8/22 Diego Biurrun :
> > On Wed, Aug 22, 2012 at 01:24:34PM +0200, Jordi Ortiz wrote:
> >> From: Jordi Ortiz
>
> ^^
> It seems that I created the original patch at uni =s
>
> > LGTM,
"Babic, Nedeljko" writes:
> Your list is more or less correct in general terms.
>
> We agree that inline assembler is a problem for clang (and other
> compilers that do not support it).
>
> However, for our architecture and code that we are delivering for this
> project rest of the points from th
On Wed, Aug 22, 2012 at 7:58 AM, Babic, Nedeljko wrote:
> Your list is more or less correct in general terms.
>
> We agree that inline assembler is a problem for clang (and other compilers
> that
> do not support it).
>
> However, for our architecture and code that we are delivering for this proj
On 8/22/12 4:58 PM, Babic, Nedeljko wrote:
Your list is more or less correct in general terms.
We agree that inline assembler is a problem for clang (and other compilers that
do not support it).
However, for our architecture and code that we are delivering for this project
rest of the points fr
Your list is more or less correct in general terms.
We agree that inline assembler is a problem for clang (and other compilers that
do not support it).
However, for our architecture and code that we are delivering for this project
rest of the points from the list are simply not valid (mips doesn'
2012/8/22 Diego Biurrun :
> On Wed, Aug 22, 2012 at 01:24:34PM +0200, Jordi Ortiz wrote:
>> From: Jordi Ortiz
^^
It seems that I created the original patch at uni =s
>
> LGTM, queued.
May I still change that? Can you amend it to nenjo...@gmail.com?
A
On Wed, Aug 22, 2012 at 01:24:34PM +0200, Jordi Ortiz wrote:
> From: Jordi Ortiz
>
> ---
> libavcodec/dirac.c | 80
> +++-
> 1 file changed, 61 insertions(+), 19 deletions(-)
LGTM, queued.
Diego
___
On Wed, Aug 22, 2012 at 12:57:25PM +0100, Måns Rullgård wrote:
> Benjamin Larsson writes:
>
> >> ping
> >>
> > LGTM
>
> It's great that it looks good to you, but how does it look to someone
> who knows the ins and outs of x86 performance?
Like a puzzle, I guess.
Benjamin Larsson writes:
>> ping
>>
> LGTM
It's great that it looks good to you, but how does it look to someone
who knows the ins and outs of x86 performance?
--
Måns Rullgård
m...@mansr.com
___
libav-devel mailing list
libav-devel@libav.org
https:/
On Wed, Aug 22, 2012 at 12:50:18PM +0100, Måns Rullgård wrote:
> Mans Rullgard writes:
>
> > GCC 4.3 and later do the right thing with the plain C code. Earlier
> > versions in 32-bit mode generate one extra instruction, needlessly
> > zeroing what would be the high half of the shifted value. A
ping
LGTM
___
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-devel
Mans Rullgard writes:
> GCC 4.3 and later do the right thing with the plain C code. Earlier
> versions in 32-bit mode generate one extra instruction, needlessly
> zeroing what would be the high half of the shifted value. At least
> two gcc configurations miscompile the inline asm in some situat
---
libavcodec/libschroedingerenc.c |9 +
1 file changed, 9 insertions(+)
diff --git a/libavcodec/libschroedingerenc.c b/libavcodec/libschroedingerenc.c
index 63d354e..ccfb145 100644
--- a/libavcodec/libschroedingerenc.c
+++ b/libavcodec/libschroedingerenc.c
@@ -220,6 +220,15 @@ stati
---
libavcodec/libschroedingerenc.c | 10 ++
1 file changed, 10 insertions(+)
diff --git a/libavcodec/libschroedingerenc.c b/libavcodec/libschroedingerenc.c
index 63d354e..1c7a418 100644
--- a/libavcodec/libschroedingerenc.c
+++ b/libavcodec/libschroedingerenc.c
@@ -220,6 +220,16 @@ sta
From: Jordi Ortiz
---
libavcodec/dirac.c | 80 +++-
1 file changed, 61 insertions(+), 19 deletions(-)
diff --git a/libavcodec/dirac.c b/libavcodec/dirac.c
index 3b5d177..489b1f5 100644
--- a/libavcodec/dirac.c
+++ b/libavcodec/dirac.c
@@ -58,6 +
On Tue, Aug 21, 2012 at 10:08:03PM -0500, daniel.d.k...@gmail.com wrote:
> From: Daniel Kang
>
> ---
> libavcodec/x86/Makefile |3 +-
> libavcodec/x86/dsputil.asm| 138
> libavcodec/x86/dsputil_mmx.c | 79 +-
> libavcodec/x86/dsputil_mmx
From: Alex Converse
Signed-off-by: Luca Barbato
---
Changelog|1 +
doc/filters.texi | 33 ++
libavfilter/Makefile |1 +
libavfilter/af_alength.c | 160 ++
libavfilter/allfilters.c |1 +
libavfilter/ve
On Wed, Aug 22, 2012 at 10:41:16AM +0200, Anton Khirnov wrote:
> Also fill in missing dates and hashes.
> ---
> doc/APIchanges | 22 --
> 1 file changed, 16 insertions(+), 6 deletions(-)
OK
Diego
___
libav-devel mailing list
libav
Also fill in missing dates and hashes.
---
doc/APIchanges | 22 --
1 file changed, 16 insertions(+), 6 deletions(-)
diff --git a/doc/APIchanges b/doc/APIchanges
index 5e67e0d..f919e53 100644
--- a/doc/APIchanges
+++ b/doc/APIchanges
@@ -13,28 +13,38 @@ libavutil: 2011-04
On Tue, Aug 21, 2012 at 06:34:00PM -0700, Alex Converse wrote:
> ---
> Changelog |1 +
> doc/filters.texi | 30 ++
> libavfilter/Makefile |1 +
> libavfilter/af_asetnsamples.c | 210
> +
> libavfilter
On 8/22/12 10:16 AM, Anton Khirnov wrote:
Double does not have enough precision to represent all int64 numbers
exactly.
Seems fine.
___
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-devel
Double does not have enough precision to represent all int64 numbers
exactly.
---
libavcodec/options_table.h |6 +++---
libavfilter/af_asyncts.c |2 +-
libavresample/options.c|4 ++--
libavutil/opt.c| 11 ++-
4 files changed, 12 insertions(+), 11 deletions(-
On Wed, 22 Aug 2012 08:57:13 +0200, Luca Barbato wrote:
> From: Alex Converse
>
> Signed-off-by: Luca Barbato
> ---
> Changelog|1 +
> doc/filters.texi | 33 +
> libavfilter/Makefile |1 +
> libavfilter/af_alength.c | 162
> +
On Tue, 21 Aug 2012 21:58:42 -0700, Alex Converse
wrote:
> ---
> Changelog|1 +
> doc/filters.texi | 33 ++
> libavfilter/Makefile |1 +
> libavfilter/af_alength.c | 161
> ++
> libavfilter/allfilters.c
On 8/22/12 3:34 AM, Alex Converse wrote:
From: Stefano Sabatini
This filter changes the number of samples on single output operation.
I wonder if having those options in the resample filter wouldn't be nice.
On the other hand might be nice have a complex filter so it picks the
duration fro
68 matches
Mail list logo