Nothing useful was ever written to this file. It did, however,
often cause spurious failures when the file was left behind by
another user on the system.
Originally committed as revision 19938 to svn://svn.ffmpeg.org/ffmpeg/trunk
This adds a function pointer for forward MDCT to FFTContext and
initialises it with the existing C function. ff_calc_mdct() is
changed to an inline function calling the selected version as
done for other fft/mdct functions.
Originally committed as revision 19818 to svn://svn.ffmpeg.org/ffmpeg/trunk
libx264 recently started mangling the name of x264_encoder_open() to
prevent version mismatches, breaking our test. Checking for another
function instead makes it work again.
Originally committed as revision 19754 to svn://svn.ffmpeg.org/ffmpeg/trunk
The DECLARE_ALIGNED_8 macro is defined to align to 16 bytes instead
the 8 suggested by the name on some CPUs. None of the uses of this
macro ever need 16-byte alignment, cases which once did having been
changed to always specify 16 bytes explicitly.
Originally committed as revision 19737 to svn://svn.ffmpeg.org/ffmpeg/trunk
The first member of struct ScanTable is a simple pointer, extra alignment
of which serves no purpose. The alignment specifier was added along with
some Altivec optimisations also adding a 16-byte-aligned array at the end
of struct ScanTable. Presumably the redundant, outer alignment was added
by mistake. The inner one is clearly sufficient.
Originally committed as revision 19724 to svn://svn.ffmpeg.org/ffmpeg/trunk
These arrays are passed to functions for which we promise 16-byte
alignment, so make sure we fulfil that promise.
Originally committed as revision 19716 to svn://svn.ffmpeg.org/ffmpeg/trunk
Storing a single element from a vector where all elements have the same
value does not require an aligned destination. Which element is stored
depends on the alignment of the destination address, but since they all
have the same value, the result is the same regardless of the alignment.
Originally committed as revision 19696 to svn://svn.ffmpeg.org/ffmpeg/trunk
Instead of filling a local array with the desired value and loading it,
load a single element and vec_splat() it to fill the vector.
Originally committed as revision 19691 to svn://svn.ffmpeg.org/ffmpeg/trunk
This makes comma-separated patters to --disable-decoder and friends
work correctly with POSIX-compliant shells.
Originally committed as revision 19688 to svn://svn.ffmpeg.org/ffmpeg/trunk
Using this flag enables position-independent code even when not strictly
required. It is impossible to use --disable-pic to forcibly disable PIC
when other properties mandate it.
Originally committed as revision 19672 to svn://svn.ffmpeg.org/ffmpeg/trunk
There is a theoretical possibility to pass a very long string to ff_parse,
which could crash if allocated from the stack. This allows the allocation
to be checked properly.
Originally committed as revision 19670 to svn://svn.ffmpeg.org/ffmpeg/trunk
pred_order can never exceed 32, so always allocating that amount is safe
and not very wasteful.
Originally committed as revision 19669 to svn://svn.ffmpeg.org/ffmpeg/trunk
Recognise AVR32 processor names as well as the generic "ap" and "uc"
family names as values for --cpu. Also define two subtypes, avr32_ap
and avr32_uc.
Originally committed as revision 19663 to svn://svn.ffmpeg.org/ffmpeg/trunk
The per-arch headers can define any combination of B/L/N variants.
This ensures that whatever is defined in an arch header gets used
for all equivalents not defined there. E.g. on a little-endian
machine, AV_RN and AV_RL should give the same code.
Originally committed as revision 19658 to svn://svn.ffmpeg.org/ffmpeg/trunk
When running behind ccache, the output from -MMD is corrupted unless
we also use the -MF and -MT flags.
Since ccache is difficult to detect and gcc 2.x doesn't support the
-MF and -MT flags, we always use the old dependency generation method
when gcc 2.x is detected.
Originally committed as revision 19620 to svn://svn.ffmpeg.org/ffmpeg/trunk
The argument to these options is now a comma-separated list of shell
patterns, e.g. --disable-decoder='indeo*,rv*'
Originally committed as revision 19604 to svn://svn.ffmpeg.org/ffmpeg/trunk
The name can now be a standard shell pattern. For example,
--disable-encoder=* disables all encoders.
Originally committed as revision 19603 to svn://svn.ffmpeg.org/ffmpeg/trunk
This requires a gcc version with -march=native or -mcpu=native support,
which is somewhat random for non-x86.
Originally committed as revision 19602 to svn://svn.ffmpeg.org/ffmpeg/trunk
Rationale:
- Current versions of XLC accept gcc flags
- XLC can't compile FFmpeg
Originally committed as revision 19488 to svn://svn.ffmpeg.org/ffmpeg/trunk
It is possible that the compiler requires flags set by the --cpu
option, so we should not attempt to run it before those flags have
been added.
Originally committed as revision 19452 to svn://svn.ffmpeg.org/ffmpeg/trunk
We do not issue similar warnings for other architectures, and we
generally assume users know what they want.
Originally committed as revision 19445 to svn://svn.ffmpeg.org/ffmpeg/trunk
The auto-vectoriser in gcc is enabled by default at -O3 since version
4.3. However, it provides no speed benefit, but does produce incorrect
code on many targets. Disabling it for gcc should give more reliable
builds.
If the adventurous want it back, they can edit the makefile themselves.
Originally committed as revision 19439 to svn://svn.ffmpeg.org/ffmpeg/trunk
Altivec, NEON, and SSE require more alignment than standard malloc()
guarantees, whereas MMX does not. Update checks to reflect this.
Originally committed as revision 19438 to svn://svn.ffmpeg.org/ffmpeg/trunk
Måns Rullgård