The icy_metadata_headers string never gets initialized, so,
during the first call to av_strlcatf() in parse_icy(),
strlen() will be called on a pointer to uninitialized memory.
At best this causes some garbage data to be left at the
start of the string.
By initializing icy_metadata_headers to the empty string, the
first call to strlen() will always return 0, so that data is
appended from the start of the string.
Anton Khirnov [Thu, 14 Nov 2013 11:19:04 +0000 (12:19 +0100)]
Work around broken floating point limits on some systems.
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
breaks code (e.g. in vf_fps) like
double f = DBL_MAX;
[...]
if (f == DBL_MAX) { // f has not been changed yet
[....]
}
Martin Storsjö [Sun, 9 Mar 2014 20:44:10 +0000 (22:44 +0200)]
mxfdec: Validate parameters to strftime
The MSVCRT version of strftime calls the invalid parameter handler
if the struct values in struct tm are invalid. In case no invalid
parameter handler is set for the process, the process is aborted.
This fixes fate failures on MSVC builds since 570af382.
Tim Walker [Mon, 3 Mar 2014 14:53:40 +0000 (14:53 +0000)]
movenc: use 'hev1' tag for HEVC in MODE_MOV.
'hvc1' requires that parameter set NAL units be
present only in the samples entry, but not in the
samples themselves, requiring that additional
parameter sets, if present, be filtered out of the
samples and placed in new, additional sample entries
if they override or otherwise conflict with the
parameter sets present in the first sample entry.
We do not have any way of doing this at present, so
the files we produce can only comply with the
restrictions set for the 'hev1' sample entry name in
ISO/IEC 14496-15.
movdec: handle 0x7fff langcode as macintosh per the specs
The correct point that seperates ISO and MAC language codes is 0x400
according to the current QT spec. Old QT specs did not list where this
seperation is but apparently only defined the meaning of the first 137.
Mark Himsley [Fri, 1 Nov 2013 11:22:53 +0000 (11:22 +0000)]
isom: lpcm in mov default to big endian
It is my understanding that "Unless otherwise stated, all data in a
QuickTime movie is stored in big-endian byte ordering" [1] in MOV files.
I have a couple of thousand files, which technically are invalid because
their sound sample description element 4CC is 'lpcm' but its version is
0 - and "Version 0 supports only uncompressed audio in raw ('raw ') or
twos-complement ('twos') format" [2]
Because isom.c only contains a mapping for 4CC 'lpcm' to
AV_CODEC_ID_PCM_S16LE, these files have their audio decoded as LE when
it is actually BE.
This commit adds AV_CODEC_ID_PCM_S16BE as the first match for 4CC 'lpcm'.
Janne Grunau [Fri, 24 Jan 2014 00:15:31 +0000 (01:15 +0100)]
timer: use mach_absolute_time as high resolution clock on darwin
Not guaranteed to be in nanosecond resolution. On iOS 7 the duration
of one tick is 125/3 ns which is still more than an order of magnitude
better then microseconds.
Replace decicycles with the neutral UNITS. Decicycles is strange but
tenths of a nanosecond and unspecific "deci"-ticks for mach_absolute_time
is just silly.
Janne Grunau [Wed, 29 Jan 2014 13:48:22 +0000 (14:48 +0100)]
arm: get_cabac inline asm
Based on the aarch64 asm. CPU cycle counts on cortex-a9 compared to
gcc 4.8.2:
before: 475 decicycles in get_cabac_noinline, 67106035 runs, 2829 skips
after: 393 decicycles in get_cabac_noinline, 67106474 runs, 2390 skips
Overall speedup is above 2%. Code generated by clang 3.4 is slower on
the same hardware and the relative change is a little larger.
Janne Grunau [Mon, 13 Jan 2014 00:06:20 +0000 (01:06 +0100)]
aarch64: get_cabac inline asm
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().
Janne Grunau [Sat, 8 Mar 2014 10:52:14 +0000 (11:52 +0100)]
arm: hpeldsp: fix put_pixels8_y2_{,no_rnd_}armv6
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.
Martin Storsjö [Wed, 5 Mar 2014 14:07:37 +0000 (16:07 +0200)]
rtmppkt: Rename the ts_delta field to ts_field
Based on a suggestion by Martin Panter. This is more descriptive,
since it's the actual timestamp field from the RTMP packet,
which might or might not be a delta depending on context (in
some packets it's a delta, in some packets it's an absolute
timestamp, and in some packets it's 0xffffff to indicate that
the actual delta or absolute timestamp is transmitted separately).
Martin Panter [Wed, 5 Mar 2014 04:04:39 +0000 (04:04 +0000)]
rtmppkt: Handle extended timestamp field even for one-byte header
Related fix in "rtmpdump":
https://repo.or.cz/w/rtmpdump.git/commitdiff/79459a2
Adobe's RTMP specification (21 Dec 2012), section 5.3.1.3 ("Extended
Timestamp"), says "this field is present in Type 3 chunks". Type 3 chunks are
those with the one-byte header size.
This resolves intermittent hangs and segfaults caused by the read function,
and also includes an untested fix for the write function.
The read function was tested with ABC (Australia) News 24 streams, however
they are probably restricted to only Australian internet addresses. Some of
the packets at the start of these streams seem to contain junk timestamp
fields, often requiring the extended field. Test command:
projects / ffmpeg / log