* minimizes diff due to cosmetic changes
* reorders get_nal_start and get_nalsize to match the original if/else
and directly break after get_nal_start at the end of buffer instead
of calling get_nalsize.
* rename get_nal_start -> find_start_code and
get_nalsize -> get_avc_nalsi
On 2014-04-06 20:41:04 +0100, Derek Buitenhuis wrote:
> Signed-off-by: Derek Buitenhuis
> ---
> Should not be pushed until the following goes in on Monday:
>
> https://mailman.videolan.org/pipermail/x265-devel/2014-April/004116.html
> ---
> configure| 4 ++--
> libavcodec/libx265.c |
On 2014-04-05 03:41:14 +0200, Luca Barbato wrote:
> Two less read and somehow match the style of other functions with similar
> purpose in the codebase.
only in the unlikely case of damaged streams, one read more for
the expected case.
> ---
> libavcodec/h264.c | 21 +
> 1 fi
On 2014-04-05 03:41:13 +0200, Luca Barbato wrote:
> ---
> libavcodec/h264.c | 422
> +++---
> 1 file changed, 209 insertions(+), 213 deletions(-)
>
> diff --git a/libavcodec/h264.c b/libavcodec/h264.c
> index 4c432ae..1a04cc1 100644
> --- a/libavco
From: Janne Grunau
8% faster h264 decoding on Apple A7.
---
libavcodec/aarch64/Makefile| 3 +++
libavcodec/aarch64/videodsp.S | 29 +
libavcodec/aarch64/videodsp_init.c | 32
libavcodec/videodsp.c | 2
From: Janne Grunau
---
Makefile| 2 +-
arch.mak| 1 +
configure | 3 +++
libavutil/aarch64/cpu.c | 5 +++--
libavutil/aarch64/cpu.h | 1 +
libavutil/cpu.c | 2 ++
libavutil/cpu.h | 1 +
7 files changed, 12 insertions(+), 3 deletions
On 2014-04-05 14:35:00 +0200, Diego Biurrun wrote:
> On Sat, Apr 05, 2014 at 02:16:47PM +0200, Janne Grunau wrote:
> > From: Janne Grunau
> >
> > ---
> > libavutil/aarch64/cpu.c | 3 ++-
> > libavutil/aarch64/cpu.h | 1 +
> > libavutil/cpu.c
From: Janne Grunau
8% faster h264 decoding on Apple A7.
---
libavcodec/aarch64/Makefile| 2 ++
libavcodec/aarch64/videodsp.S | 28
libavcodec/aarch64/videodsp_init.c | 32
libavcodec/videodsp.c | 2
From: Janne Grunau
---
libavutil/aarch64/cpu.c | 3 ++-
libavutil/aarch64/cpu.h | 1 +
libavutil/cpu.c | 2 ++
libavutil/cpu.h | 1 +
4 files changed, 6 insertions(+), 1 deletion(-)
diff --git a/libavutil/aarch64/cpu.c b/libavutil/aarch64/cpu.c
index 2cc2ccd..4ad19b4 100644
Unbreaks compilation on arm due to conflicting types for
'ff_decode_block_coeffs_armv6'.
---
libavcodec/arm/vp8.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavcodec/arm/vp8.h b/libavcodec/arm/vp8.h
index cd3428b..93b2788 100644
--- a/libavcodec/arm/vp8.h
+++ b/libavcode
Fixes NULL pointer dereference with disabled threading. Should be
squashed.
---
libavcodec/utils.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/libavcodec/utils.c b/libavcodec/utils.c
index e04f455..d9832e2 100644
--- a/libavcodec/utils.c
+++ b/libavcodec/utils.c
@@ -2292,
On 2014-04-03 09:01:02 +0300, Martin Storsjö wrote:
> On Thu, 3 Apr 2014, Tim Walker wrote:
>
> >On 02 Apr 2014, at 12:15, Janne Grunau wrote:
> >
> >>The first entry in @ARGV not starting with a dash or everything after
> >>'--' is handled as gcc_
On 2014-04-02 12:32:24 +0200, Luca Barbato wrote:
> On 02/04/14 12:15, Janne Grunau wrote:
> > ---
> > gas-preprocessor.pl | 4
> > 1 file changed, 4 insertions(+)
>
> Probably ok, new code coming?
unrelated, this fix was necessary for the float_dsp code an
---
gas-preprocessor.pl | 4
1 file changed, 4 insertions(+)
diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
index 6dd9cc6..1abe9fe 100755
--- a/gas-preprocessor.pl
+++ b/gas-preprocessor.pl
@@ -803,6 +803,10 @@ sub handle_serialized_line {
$line =~ s/\b(saddl2?)(\s+v
The first entry in @ARGV not starting with a dash or everything after
'--' is handled as gcc_cmd command. Makes ./gas-preprocessor.pl -help
work as intended.
---
gas-preprocessor.pl | 33 -
1 file changed, 8 insertions(+), 25 deletions(-)
diff --git a/gas-preproces
On 2014-04-02 00:48:28 +0200, Janne Grunau wrote:
> ---
> it still spews out a "64bpp not supported by yuv2rgb" in
> ff_yuv2rgb_c_init_tables(). no idea why that is called at all.
> But seems to work fine. it produces the same output with and without
> -filter 'for
---
it still spews out a "64bpp not supported by yuv2rgb" in
ff_yuv2rgb_c_init_tables(). no idea why that is called at all.
But seems to work fine. it produces the same output with and without
-filter 'format=pix_fmts=rgba64be,showinfo,format=pix_fmts=rgba64le'
libswscale/swscale_unscaled.c | 2 +
---
any ideas how to auto-generate this from pixfmt.h? or using pixel format
descriptions.
libavutil/pixdesc.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/libavutil/pixdesc.c b/libavutil/pixdesc.c
index d0e6919..1b34fc9 100644
--- a/libavutil/pixdesc.c
+++ b/libavutil/pixdesc.c
@@ -165
On 2014-03-26 21:57:08 +0200, Martin Storsjö wrote:
> The previous instructions were woefully outdated.
> ---
> Using xcrun to find the currently installed SDK path, mentioning
> iOS instead of explicitly saying e.g. iPhone 3gs.
> ---
> README | 9 ++---
> 1 file changed, 2 insertions(+), 7 de
On 2014-03-24 19:38:03 +0200, Martin Storsjö wrote:
> ---
> gas-preprocessor.pl | 5 +
> test.S | 11 +++
> 2 files changed, 16 insertions(+)
>
> diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
> index d875d76..6dd9cc6 100755
> --- a/gas-preprocessor.pl
> +++ b/ga
On 2014-03-24 14:46:10 +0200, Martin Storsjö wrote:
> ---
> gas-preprocessor.pl | 11 ++-
> test.S | 6 ++
> 2 files changed, 16 insertions(+), 1 deletion(-)
>
> diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
> index 16e91ab..6d8e57b 100755
> --- a/gas-preprocess
On 2014-03-26 12:04:28 +0100, Diego Biurrun wrote:
> This fixes FATE with mmxext CPUFLAGS set.
> ---
> libavcodec/x86/h264_qpel.c |2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/libavcodec/x86/h264_qpel.c b/libavcodec/x86/h264_qpel.c
> index aa6e9d5..b285c1f 100644
> -
On 2014-03-24 21:20:27 +0100, Luca Barbato wrote:
> On 24/03/14 20:26, wm4 wrote:
> > In practice, it's called when the decoder is flushed, i.e. on seeking.
> > And yes, this is a bit of a bother. Freeing all hardware surfaces and
> > reallocating them etc. makes seeking quite a bit slower this way
On 2014-03-24 14:38:11 +0200, Martin Storsjö wrote:
> A branch instruction in thumb mode might either be narrow or
> wide, depending on the offset to the branch location.
>
> armasm seems to have a bug when calculating the difference
> between two symbols, where there's such a branch inbetween.
>
On 2014-03-24 14:38:10 +0200, Martin Storsjö wrote:
> ---
> gas-preprocessor.pl | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
> index 8864cfd..42e1c7f 100755
> --- a/gas-preprocessor.pl
> +++ b/gas-preprocessor.pl
> @@ -936,6 +936,8 @@ sub ha
On 2014-03-24 14:38:09 +0200, Martin Storsjö wrote:
> ---
> gas-preprocessor.pl | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
> index f013c71..8864cfd 100755
> --- a/gas-preprocessor.pl
> +++ b/gas-preprocessor.pl
> @@ -616,
On 2014-03-24 14:38:08 +0200, Martin Storsjö wrote:
> ---
> gas-preprocessor.pl | 4
> 1 file changed, 4 insertions(+)
>
> diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
> index e291094..f013c71 100755
> --- a/gas-preprocessor.pl
> +++ b/gas-preprocessor.pl
> @@ -849,6 +849,10 @@ sub
On 2014-03-24 14:38:07 +0200, Martin Storsjö wrote:
> There is no longer any need to do the processing in two passes.
> ---
> gas-preprocessor.pl | 100
> +---
> 1 file changed, 49 insertions(+), 51 deletions(-)
looks good
Janne
__
On 2014-03-24 14:38:06 +0200, Martin Storsjö wrote:
> Use two separate comment variables - one for the input source
> format and one used to comment out directives from the output.
>
> This is necessary if the two input passes are merged.
> ---
> gas-preprocessor.pl | 6 +++---
> 1 file changed,
On 2014-03-24 14:38:05 +0200, Martin Storsjö wrote:
> All comments with the original comment character have already
> been stripped out.
> ---
> gas-preprocessor.pl | 4
> 1 file changed, 4 deletions(-)
>
> diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
> index 957a774..10c7203 10075
On 2014-03-24 14:38:04 +0200, Martin Storsjö wrote:
> We could have injected extra lines with temporary labels before
> this one, if the line consisted of a numeric label before the
> branch instruction.
> ---
> gas-preprocessor.pl | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff
On 2014-03-24 14:38:03 +0200, Martin Storsjö wrote:
> ---
> gas-preprocessor.pl | 27 +--
> 1 file changed, 17 insertions(+), 10 deletions(-)
ok
Janne
___
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/ma
On 2014-03-24 14:38:02 +0200, Martin Storsjö wrote:
> ---
> gas-preprocessor.pl | 14 +++---
> 1 file changed, 11 insertions(+), 3 deletions(-)
>
> diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
> index 27a84b7..3bdc7dd 100755
> --- a/gas-preprocessor.pl
> +++ b/gas-preprocessor.p
On 2014-03-24 14:38:01 +0200, Martin Storsjö wrote:
> ---
> gas-preprocessor.pl | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
ok
Janne
___
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-dev
On 2014-03-24 14:38:00 +0200, Martin Storsjö wrote:
> ---
> gas-preprocessor.pl | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
> index 5498b7c..6274571 100755
> --- a/gas-preprocessor.pl
> +++ b/gas-preprocessor.pl
> @@ -914,7
On 2014-03-24 14:37:59 +0200, Martin Storsjö wrote:
> Also add test cases that show that macros can use commas
> or omit them.
> ---
> gas-preprocessor.pl | 2 +-
> test.S | 18 ++
> 2 files changed, 19 insertions(+), 1 deletion(-)
ok
Janne
_
On 2014-03-24 14:37:58 +0200, Martin Storsjö wrote:
> This avoids prematurely evaluating .if conditions within
> repetitions/macros, that previously were erroneously
> evaluated too early if the enclosing environment was wrapped
> in an .if.
> ---
> gas-preprocessor.pl | 33 +++
On 2014-03-24 14:37:55 +0200, Martin Storsjö wrote:
> Instead of handling macros and repetitions in two passes, with
> .if handling spread out over the two, handle both at the same
> time.
> ---
> gas-preprocessor.pl | 171
> ++--
> test.S
On 2014-03-24 14:37:57 +0200, Martin Storsjö wrote:
> This fixes cases of evaluating x-y when y is a negative number.
> ---
> gas-preprocessor.pl | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
> index 190ef87..5917d2b 100755
> --
On 2014-03-24 14:37:56 +0200, Martin Storsjö wrote:
> Since the second pass loop doesn't handle expanding repetitions
> any longer, there's no need to have this as a separate function.
> ---
> gas-preprocessor.pl | 18 ++
> 1 file changed, 6 insertions(+), 12 deletions(-)
>
> diff
On 2014-03-24 14:49:34 +0200, Martin Storsjö wrote:
> On Mon, 24 Mar 2014, Janne Grunau wrote:
>
> >On 2014-03-24 14:12:47 +0200, Martin Storsjö wrote:
> >>On Thu, 20 Mar 2014, Ben Avison wrote:
> >>
> >>>+function ff_mlp_pack_output_inorder_\chan
On 2014-03-24 14:37:54 +0200, Martin Storsjö wrote:
> By comparing the disassembled output of gas-preprocessor to
> the disassembly from plain gas, we can easily test and make sure
> we handle the gas features in a similar way as the actual
> assembler.
> ---
> .gitignore | 3 +++
> Makefile |
On 2014-03-24 14:37:53 +0200, Martin Storsjö wrote:
> It is no longer commented out at this point.
> ---
> gas-preprocessor.pl | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
> index 14a6309..d6779cd 100755
> --- a/gas-preprocesso
On 2014-03-24 14:37:52 +0200, Martin Storsjö wrote:
> From: Janne Grunau
>
> Although it is allowed it does not make much sense since the immediate
> is already 8bits.
> ---
> gas-preprocessor.pl | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --
On 2014-03-24 14:12:47 +0200, Martin Storsjö wrote:
> On Thu, 20 Mar 2014, Ben Avison wrote:
>
> >Profiling results for overall decode and the output_data function in
> >particular are as follows:
> >
> > Before After
> > Mean StdDev Mean StdDev Confidence
On 2014-03-23 20:23:42 +0100, Vittorio Giovara wrote:
>
> Attached patch simplifies iOS configure lines.
simplifies how? I don't think I have anything in my configure command
which could be removed after this patch.
see https://fate.libav.org/armv7-apple-darwin-xcode5/20140320182122 or
https://f
On 2014-03-22 14:38:53 +0100, Vittorio Giovara wrote:
> On Tuesday, March 18, 2014, Luca Barbato wrote:
>
> > On 12/03/14 18:27, Vittorio Giovara wrote:
> > > ---
> > > libavfilter/vf_scale.c | 20 +++-
> > > 1 file changed, 7 insertions(+), 13 deletions(-)
> >
> > I'm not sure w
From: Luca Barbato
---
libavcodec/Makefile | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/libavcodec/Makefile b/libavcodec/Makefile
index ce6fab7..cf5166d 100644
--- a/libavcodec/Makefile
+++ b/libavcodec/Makefile
@@ -193,10 +193,10 @@ OBJS-$(CONFIG_H263_DECODER)
From: Luca Barbato
---
configure | 2 +-
libavcodec/Makefile | 6 +-
2 files changed, 2 insertions(+), 6 deletions(-)
diff --git a/configure b/configure
index fced12f..bd1d703 100755
--- a/configure
+++ b/configure
@@ -1907,7 +1907,7 @@ wmv3_vaapi_hwaccel_select="vc1_vaapi_hwaccel
Hi,
here's the updated patch set for splitting MB and slice decoding
from h264.c. Manually verified with appropriate git diff options
that changes are just cosmetic,
git diff -B --minimal COMMIT~1:libavcodec/h264.c..COMMIT:libavcodec/NEW_FILE
shows mostly deletes and some changes due to function
On 2014-03-20 09:33:10 +0200, Martin Storsjö wrote:
> On Thu, 20 Mar 2014, Ben Avison wrote:
>
> >diff --git a/libavcodec/arm/mlpdsp_arm.S b/libavcodec/arm/mlpdsp_arm.S
> >new file mode 100644
> >index 000..9e0bf57
> >--- /dev/null
> >+++ b/libavcodec/arm/mlpdsp_arm.S
> >@@ -0,0 +1,435 @@
> >+
On 2014-03-19 16:25:10 +0200, Martin Storsjö wrote:
> On Wed, 19 Mar 2014, Janne Grunau wrote:
>
> >Although it is allowed it does not make much sense since the immediate
> >is already 8bits.
> >---
> >gas-preprocessor.pl | 2 +-
> >1 file changed, 1 insertion(+
Although it is allowed it does not make much sense since the immediate
is already 8bits.
---
gas-preprocessor.pl | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
index 914c41c..692e788 100755
--- a/gas-preprocessor.pl
+++ b/gas-preproces
On 2014-03-19 14:51:27 +0200, Martin Storsjö wrote:
> This fixes fate-float_dsp-test on RVCT 4.0.
> ---
> libavutil/float_dsp.c | 48 +---
> 1 file changed, 25 insertions(+), 23 deletions(-)
ok
Janne
___
liba
It is not supported by all compilers on FATE. Fixes "some test were
skipped" errors.
---
libavutil/float_dsp.c | 18 +-
1 file changed, 9 insertions(+), 9 deletions(-)
diff --git a/libavutil/float_dsp.c b/libavutil/float_dsp.c
index 23468ad..a83caed 100644
--- a/libavutil/float_ds
2c
--- /dev/null
+++ b/libavutil/aarch64/float_dsp_neon.S
@@ -0,0 +1,202 @@
+/*
+ * ARM NEON optimised Float DSP functions
+ * Copyright (c) 2008 Mans Rullgard
+ * Copyright (c) 2014 Janne Grunau
+ *
+ * This file is part of Libav.
+ *
+ * Libav is free software; you can redistribute it and/or
+ * m
---
added the fate test as it looked stable enough on oracle
fixed the indentation and commit message
libavutil/Makefile | 1 +
libavutil/float_dsp.c| 283 +++
tests/fate/libavutil.mak | 6 +
3 files changed, 290 insertions(+)
diff --git
..776542c
--- /dev/null
+++ b/libavutil/aarch64/float_dsp_neon.S
@@ -0,0 +1,202 @@
+/*
+ * ARM NEON optimised Float DSP functions
+ * Copyright (c) 2008 Mans Rullgard
+ * Copyright (c) 2014 Janne Grunau
+ *
+ * This file is part of Libav.
+ *
+ * Libav is free software; you can redistribute
Avoids a warning of the unused function 'autocorrelate'.
---
libavcodec/sbrdsp.c | 12
1 file changed, 8 insertions(+), 4 deletions(-)
diff --git a/libavcodec/sbrdsp.c b/libavcodec/sbrdsp.c
index 4d07af5..b7917dc 100644
--- a/libavcodec/sbrdsp.c
+++ b/libavcodec/sbrdsp.c
@@ -109,6 +1
The code using it was remove in d66e305bd1b.
---
libavcodec/error_resilience.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/libavcodec/error_resilience.c b/libavcodec/error_resilience.c
index d648f09..6ab5a98 100644
--- a/libavcodec/error_resilience.c
+++ b/libavcodec/error_resilience.c
@@ -
On 2014-03-18 11:36:19 +0200, Martin Storsjö wrote:
> ---
> gas-preprocessor.pl | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
> index 36cef69..914c41c 100755
> --- a/gas-preprocessor.pl
> +++ b/gas-preprocessor.pl
> @@ -143,
On 2014-03-17 12:41:54 +0200, Martin Storsjö wrote:
> ---
> gas-preprocessor.pl | 6 --
> 1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
> index 0417f47..eefe571 100755
> --- a/gas-preprocessor.pl
> +++ b/gas-preprocessor.pl
> @@ -28
On 2014-03-17 14:28:13 +0200, Martin Storsjö wrote:
> This does some rudimentary conversions of instructions that aren't
> available in thumb mode.
>
> This allows building OpenH264 for Windows Phone 8 (and Windows RT),
> which only supports thumb mode, and the OpenH264 arm assembly is
> hardcoded
On 2014-03-17 12:17:46 +0200, Martin Storsjö wrote:
> On Mon, 17 Mar 2014, Janne Grunau wrote:
>
> >On 2014-03-17 10:56:49 +0200, Martin Storsjö wrote:
> >>This does some rudimentary conversions of instructions that aren't
> >>available in thumb mode.
> &g
On 2014-03-17 10:56:48 +0200, Martin Storsjö wrote:
> This is for use with OpenH264, which specifies more MSVC specific
> compiler flags than libav does.
> ---
> gas-preprocessor.pl | 14 ++
> 1 file changed, 14 insertions(+)
>
> diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
On 2014-03-17 10:56:47 +0200, Martin Storsjö wrote:
> Libav itself doesn't output these unless targeting ELF (where
> they don't need to be commented out), but other projects (e.g.
> OpenH264) have them unconditionally in the source.
> ---
> gas-preprocessor.pl | 1 +
> 1 file changed, 1 insertion
On 2014-03-17 10:56:49 +0200, Martin Storsjö wrote:
> This does some rudimentary conversions of instructions that aren't
> available in thumb mode.
>
> This allows building OpenH264 for Windows Phone 8 (and Windows RT),
> which only supports thumb mode, and the OpenH264 arm assembly is
> hardcoded
On 2014-03-17 10:46:34 +0200, Martin Storsjö wrote:
> The RE for matching apple- was typoed in ea29a0e6c. Additionally,
> reverse the order of the two checks, since we use $3 and $4 to refer
> to parts of one of the REs. Prior to ea29a0e6c the as_type check
> was a plain string comparison, not a RE
On 2014-03-17 00:10:46 +0200, Martin Storsjö wrote:
> This allows the main part of the preprocessor to work with source
> in the actual input format and not have to look for directives
> in their translated form (which can be different depending on the
> target).
> ---
> Tested with armasm, linux-g
On 2014-03-12 23:02:38 +0200, Martin Storsjö wrote:
> This syntax is supported by the official arm tools and
> by Microsoft's assembler.
>
> This currently only supports microsoft's assembler, the
> armasm assembler in RVCT requires a few more tweaks to
> be able to build libav.
>
> The preproces
On 2014-03-12 22:48:49 +0200, Martin Storsjö wrote:
> On Wed, 12 Mar 2014, Martin Storsjö wrote:
>
> >On Wed, 12 Mar 2014, Janne Grunau wrote:
> >
> >>On 2014-03-12 13:24:42 +0200, Martin Storsjö wrote:
> >>>@@ -533,7 +614,7 @@ foreach my $line (@pass1_
On 2014-03-16 19:21:57 +0100, Diego Biurrun wrote:
> Module: libav
> Branch: refs/tags/dev14.2
> Commit: 4d7ab5cfebef91820af2933ef2f622ea598e6b53
deleted for now, Anton rightfully remarked that those number might become
confusing when there's a v14...
I wanted to avoid a v11_pre0, I'm open to su
On 2014-03-11 16:09:21 +0200, Martin Storsjö wrote:
> For live audio streams, requiring 500 frames for a stream to
> be detected is a bit overkill.
>
> This allows live ADTS streams that aren't frame-aligned
> to start up more quickly, e.g.
> http://mp3.streampower.be/radio1.aac.
> ---
> Since the
On 2014-03-11 16:09:20 +0200, Martin Storsjö wrote:
> If a portion of the probe buffer seem to resemble ADTS frames,
> but some data at the end is a mismatch, disregard the whole
> probing attempt. If it actually is ADTS data, there shouldn't be
> any mismatches within the sequential frame data.
>
On 2014-03-12 13:24:42 +0200, Martin Storsjö wrote:
> This syntax is supported by the official arm tools and
> by Microsoft's assembler.
>
> This currently only supports microsoft's assembler, the
> armasm assembler in RVCT requires a few more tweaks to
> be able to build libav.
>
> The preproces
On 2014-03-11 00:28:55 +0100, Luca Barbato wrote:
> On 11/03/14 00:25, Janne Grunau wrote:
> > From: Janne Grunau
> >
> > The integrated arm assembler in clang-503.0.38 (Xcode-5.1) seems
> > to get confused by the offset of 4 and decides to use a non-wide
> > thu
From: Janne Grunau
The integrated arm assembler in clang-503.0.38 (Xcode-5.1) seems
to get confused by the offset of 4 and decides to use a non-wide
thumb encoding. That fails since the labels are out of range of
the limited offset a 16-bit thumb encoding offers.
---
libavcodec/arm/vp8dsp_armv6
On 2014-03-11 00:08:20 +0100, Diego Biurrun wrote:
> On Mon, Mar 10, 2014 at 04:20:52PM +0100, Janne Grunau wrote:
> > On 2014-03-06 14:55:06 +0100, Diego Biurrun wrote:
> > > --- a/configure
> > > +++ b/configure
> > > @@ -3805,19 +3805,19 @@ elif check_func
On 2014-03-10 17:03:14 +0100, Tim Walker wrote:
> From: Michael Niedermayer
>
> Signed-off-by: Michael Niedermayer
> Signed-off-by: Tim Walker
> ---
>
> This supersedes the previous version of the patch.
>
> libavformat/hevc.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff
On 2014-03-10 17:03:13 +0100, Tim Walker wrote:
> Use ff_hevc_annexb2mp4 instead of duplicating
> its functionality, and update the documentation
> to match the new behavior.
> ---
>
> Suggested by Janne on IRC.
>
> libavformat/hevc.c | 43 +++
> libavform
On 2014-03-06 14:55:06 +0100, Diego Biurrun wrote:
> Group checks into logical groups and sort alphabetically within those groups
> unless checks have close coupling. Keep closely coupled checks together.
> ---
> configure | 49 +++--
> 1 file changed,
On 2014-03-06 14:55:05 +0100, Diego Biurrun wrote:
> Sort lists of components. Add some whitespace where it helps readability.
> Consistently use double instead of single quotes around component lists.
> ---
> configure | 67
> +++--
> 1 f
Options for gas-preprocessor need to be separated with ' -- ' from the
assembler command line.
The target architecture can now be set explicitly with the added '-arch'
option.
---
gas-preprocessor.pl | 128 ++--
1 file changed, 84 insertions(+), 44 d
Hi,
this patchset adds support for gas/clang on linux as assembler. The main
reason for this change is to support external asm with clang's integrated
assembler on aarch64. Still works with clang from Xcode 5 targeting armv7
and arm64. Tested aarch64/clang-3.4 and armv7/gas linux as arch/as type
c
Tested with libav's fate using gas on armv7 and clang-3.4 on aarch64.
---
gas-preprocessor.pl | 25 +
1 file changed, 13 insertions(+), 12 deletions(-)
diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
index 65e507b..a8f119a 100755
--- a/gas-preprocessor.pl
+++ b/gas-
When no input file was found but the compiler command contains -v or
--version just execute the command. Used by Libav's configure to
probe type and version of the backing assembler.
---
gas-preprocessor.pl | 5 +
1 file changed, 5 insertions(+)
diff --git a/gas-preprocessor.pl b/gas-preproce
---
gas-preprocessor.pl | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
index bae5e55..c31045a 100755
--- a/gas-preprocessor.pl
+++ b/gas-preprocessor.pl
@@ -454,6 +454,8 @@ my @rept_lines;
my %literal_labels; # for ldr , =
This seems to be unnecessary with current setups and actively breaks
things on arm without --enable-pic.
This reverts commit 76a72f00e0af8b8bacb4417e6806561d4c596dcc.
---
gas-preprocessor.pl | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/gas-preprocessor.pl b/gas-preproces
Allows differating between assemblers and will allow using
gas-preprocessor on linux.
---
gas-preprocessor.pl | 5 +
1 file changed, 5 insertions(+)
diff --git a/gas-preprocessor.pl b/gas-preprocessor.pl
index c31045a..65e507b 100755
--- a/gas-preprocessor.pl
+++ b/gas-preprocessor.pl
@@ -24,
From: Michael Niedermayer
Bug-Id: 649
CC: libav-sta...@libav.org
---
libavcodec/utvideodec.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavcodec/utvideodec.c b/libavcodec/utvideodec.c
index 3492595..cad2655 100644
--- a/libavcodec/utvideodec.c
+++ b/libavcodec/utvideod
Was missed in aeaf268e52fc11c1f64914a319e0edddf1346d6a when integrating
clear_blocks into the idct.
---
libavcodec/arm/vp3dsp_init_arm.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavcodec/arm/vp3dsp_init_arm.c b/libavcodec/arm/vp3dsp_init_arm.c
index dfd6078..1ab0852 10
Based on the x86 branchless get_cabac asm. get_cabac_noinline() gets
approximately 20% faster (no cycle counts available) compared to clang
from Xcode 5.1 beta5. More than 6% faster overall. A part of the overall
speedup might be explained by additional inlining of get_cabac().
---
moves the extern
On 2014-03-07 17:47:09 +0100, Vittorio Giovara wrote:
> From: Michael Niedermayer
>
> Fix assertion failure.
> Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
> ---
> libavcodec/hevc_ps.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/libavcodec/hevc_ps.c b/
The overread avoidance fix in cbddee1cca0ebd01e8c5aa694d31228eb4de4b41
broke the computation for the last row since it prevented the safe
reading from the height+1-th row.
---
I'm not fully sure how I came to the conclusion that those values are
not used during the last iteration. I either checked
On 2014-03-07 11:56:22 +0100, Vittorio Giovara wrote:
> ---
> libavcodec/mpegvideo.c | 17 +
> libavcodec/mpegvideo.h | 2 +-
> libavcodec/svq3.c | 3 +--
> 3 files changed, 11 insertions(+), 11 deletions(-)
>
> diff --git a/libavcodec/mpegvideo.c b/libavcodec/mpegvideo.c
>
On 2014-03-07 11:56:23 +0100, Vittorio Giovara wrote:
> ---
> libavcodec/h264.c | 10 +-
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/libavcodec/h264.c b/libavcodec/h264.c
> index c76abf7..1ce772f 100644
> --- a/libavcodec/h264.c
> +++ b/libavcodec/h264.c
> @@ -212,8
On 2014-03-06 13:11:02 +0100, Diego Biurrun wrote:
> The SVQ3 decoder reuses large parts of the H.264 decoder so it
> makes no sense to enable the former but not the latter.
> ---
> configure |4 ++--
> libavcodec/Makefile |5 +
> 2 files changed, 3 insertions(+), 6 deletions
On 2014-03-06 13:11:01 +0100, Diego Biurrun wrote:
> ---
> libavcodec/Makefile |2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/libavcodec/Makefile b/libavcodec/Makefile
> index 44d5ae6..0a92c7c 100644
> --- a/libavcodec/Makefile
> +++ b/libavcodec/Makefile
> @@ -335,7
On 2014-03-06 09:24:13 +0100, Anton Khirnov wrote:
> The values of {FLT,DBL}_{MAX,MIN} macros on some systems (older musl
> libc, some BSD flavours) are not exactly representable, i.e.
> (double)DBL_MAX == DBL_MAX is false
> This violates (at least some interpretations of) the C99 standard and
> br
On 2014-03-05 14:10:24 +0100, Diego Biurrun wrote:
> Some optimized functions reference optimized symbols, so the functions
> must be explicitly disabled when those symbols are unavailable.
> ---
>
> Now with a more general #if, as suggested by James.
I think --disable-optimizations --disable-sse
701 - 800 of 2711 matches
Mail list logo