After the removal of the OS SSE detection stuff ia32_cpu_info()
and x86_64_cpu_info() became very similar. Merging them makes
sense.
Patch-from: lvqcl <lvqcl.mail@gmail.com>
Removes FLAC__lpc_restore_signal_16_intrin_sse2() which was faster
than than C code, but not faster than MMX-accelerated ASM functions.
It's also slower than the new SSE4.1 functions that were added by
the previous patch.
Patch-from: lvqcl <lvqcl.mail@gmail.com>
This patch removes FLAC__lpc_restore_signal_16_intrin_sse2().
It's faster than C code, but not faster than MMX-accelerated
ASM functions. It's also slower than the new SSE4.1 functions
that were added by the previous patch.
So this function wasn't very useful before, and now it's
even less useful. I don't see a reason to keep it.
Patch-from: lvqcl <lvqcl.mail@gmail.com>
Accelerates decoding of non-Subset 24-bit FLAC files (where lpc_order
> 12).
The improved function is FLAC__lpc_restore_signal_wide_intrin_sse41().
It requires SSE4.1 and it's used only by 32-bit libFLAC.
Patch-from: lvqcl <lvqcl.mail@gmail.com>
GCC version 4.6 was throwing an error on `__has_attribute` which
seems to be Clang specific. Solution was to un-nest the CPP code
so that Clang is confirmed before the `__has_attribute` is hit.
When encoding a 24 bit multichannel W64 file on a 32 bit system, a
calculation could overflow resulting in a "got partial sample"
error and early termination of the encoding process.
The fix was to replace the cast to `size_t` (which is only 32 bits
on a 32 bit system) with a cast to `uin64_t`.
Thanks to Janne Hyvärinen for helping to track down the problem.
Closes: https://sourceforge.net/p/flac/bugs/453
The underscores are wrong. The comment is also correct.
Also remove the configure.ac option. Otherwise it tries to compile the
windows unicode stuff which POSIX(cygwin) does not understand.
Signed-off-by: Erik de Castro Lopo <erikd@mega-nerd.com>
Replace
a = _mm_alignr_epi8(b, a, n);
with
a = _mm_alignr_epi8(a, b, n);
The resulting code is very slightly faster and the binaries slightly
smaller.
Patch-from: lvqcl <lvqcl.mail@gmail.com>
Previously this function would abort or crash if passed `NULL` for
the encoder. Now it just returns without crashing or aborting which
is far more sane behaviour (ie much like FLAC__stream_encoder_delete).
* Furthermore, use NDEBUG globally to detect the presence
of building with more debug output information.
AX_CHECK_ENABLE_DEBUG is easier to use, and nowadays
Gnome has also switched to it from its own custom solution.
Signed-off-by: Erik de Castro Lopo <erikd@mega-nerd.com>
* Remove <stdint.h> because MSVC 2005 (and probably 2008) don't have it.
* Fix pointer difference caclulation between `nameend` and `nextchar`.
Patch-from: lvqcl <lvqcl.mail@gmail.com>
The orignal code had `unsigned long` inside `#ifdef MSVC`. It was then changed
to `uint64_t` which was incorrect because on Windows `sizze long == 4`.
Change it now to `uint32_t` which is always correct regardless of OS, compiler
or architecture.
* HTML files should be installed to $(htmldir), and $(docdir) should
not be changed, as this is a user flag in the GNU conventions.
Signed-off-by: Erik de Castro Lopo <erikd@mega-nerd.com>
These fixes are in the public API, but FLAC would have been extremely
unlikely to work in any system where `unsigned` was not 32 bits.
Therefore we do not treat this as an API change.
As pointed out by Ozkan Sezer, on some platforms `int32_t` is actually
a typedef for `long` so `unsigned` cannot be used interchangably with
`FLAC__uint32`. Fix is to switch from `unsigned` to explicit sized ISO
C types defined in <stdint.h>.
Not all linux hosts match the *-pc-linux-gnu wildcard, causing build
failures for older glibc versions where we need to link with -lrt for
clock_gettime - E.G.:
- arm-unknown-linux-musleabihf
- powerpc-unknown-linux-gnuspe
- bfin-linux-linux-uclibc
..
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Erik de Castro Lopo <erikd@mega-nerd.com>
