Use a versioned gcc container image for more consistent
test results. This is the same version we're using for
other projects, but of course it will need to be bumped
periodically.

The current gcc release is 10.2. The oldest supported
release is 8.4, so this is in the middle of the support
window.

Signed-off-by: Marvin Scholz <epirat07@gmail.com>

sezero authored 5 years ago and Jean-Marc Valin committed 4 years ago

.. just like the way it is done in celt/x86/x86cpu.c.

Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>

Use the fuzzed sub-length of the input data instead of the whole input.

The autotools build doesn't set OPUS_HAVE_RTCD for arm targets,
assuming all the supported intrinsics will work on the runtime
cpu.

The cmake build however defines this by default when the neon
extension is available on the target. On Linux, the runtime
cpu detection reads /proc/cpuinfo, so removing stdio.h from
celt/os_support.h meant that the cmake build for arm targets
failed.

We don't currently have ci runs for that configuration, so
this only became apparent through manual testing.

Signed-off-by: Marcus Asteborg <maastebo@microsoft.com>
Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>

Marcus Asteborg authored 4 years ago and Ralph Giles committed 4 years ago

Signed-off-by: Ralph Giles <giles@thaumas.net>

Marcus Asteborg authored 4 years ago and Jean-Marc Valin committed 4 years ago

Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>

willson-chen authored 5 years ago and Ralph Giles committed 4 years ago

Add newer source directories to the destdir file tree so we
can include all sources referenced from opus_sources.mk.

Signed-off-by: Ralph Giles <giles@thaumas.net>

willson-chen authored 4 years ago and Ralph Giles committed 4 years ago

[[ ]], the compound command is not supported by all
shell interpreter. [ ], the buildin command is more
common.

Signed-off-by: Ralph Giles <giles@thaumas.net>

Marcus Asteborg authored 4 years ago and Mark Harris committed 4 years ago

Signed-off-by: Mark Harris <mark.hsj@gmail.com>

Davide Beatrici authored 4 years ago and Mark Harris committed 4 years ago

CMAKE_SOURCE_DIR corresponds to the top project's source directory.
CMAKE_BINARY_DIR corresponds to the top project's binary directory.

The usage of these variables doesn't cause any problems when Opus is built as a standalone project.

This is not the case when Opus is added as submodule: the variables are set by the project that calls "add_subdirectory()".

The fix consists in using PROJECT_SOURCE_DIR and PROJECT_BINARY_DIR, which always refer to the current project.

Signed-off-by: Mark Harris <mark.hsj@gmail.com>

Marcus Asteborg authored 4 years ago and Mark Harris committed 4 years ago

Signed-off-by: Mark Harris <mark.hsj@gmail.com>

Marcus Asteborg authored 4 years ago and Mark Harris committed 4 years ago

Signed-off-by: Mark Harris <mark.hsj@gmail.com>

Marcus Asteborg authored 4 years ago and Mark Harris committed 4 years ago

Signed-off-by: Mark Harris <mark.hsj@gmail.com>

Marcus Asteborg authored 4 years ago and Mark Harris committed 4 years ago

Signed-off-by: Mark Harris <mark.hsj@gmail.com>

Marcus Asteborg authored 4 years ago and Mark Harris committed 4 years ago

Signed-off-by: Mark Harris <mark.hsj@gmail.com>

Marcus Asteborg authored 4 years ago and Mark Harris committed 4 years ago

Signed-off-by: Mark Harris <mark.hsj@gmail.com>

Simon Jackson authored 4 years ago and Mark Harris committed 4 years ago

Signed-off-by: Mark Harris <mark.hsj@gmail.com>

This lighter-weight intrinsics header is available starting
with Microsoft Visual Studio 2017, so the previous change
to allow this header failed when building with Visual
Studio 2015.

Restores the appveyor continuous integration build.

Marcus Asteborg authored 4 years ago and Jean-Marc Valin committed 4 years ago

Build time improvement, for MSVC use intrin0.h instead of intrin.h and remove usage of stdio.h in production code

Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>

Marcus Asteborg authored 4 years ago and Jean-Marc Valin committed 4 years ago

Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>

Marcus Asteborg authored 4 years ago and Jean-Marc Valin committed 4 years ago

Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>

Marcus Asteborg authored 4 years ago and Ralph Giles committed 4 years ago

Signed-off-by: Ralph Giles <giles@thaumas.net>

Add doc/trivial_example.c to the autotools build so we get
some minimal verification that the code compiles.

Signed-off-by: Mark Harris <mark.hsj@gmail.com>

Silence a gcc warning by checking the return value of the fread()
call instead of the feof() guard. This prevents an infinite loop
in the case of a read error. Otherwise, when end-of-file is reached
fread() will certainly return a smaller number of elements read
than requested, so both cases are handled now.

Add a comment to clarify that we're dropping a partial frame on
purpose to keep the code simple.

Also add more braces around conditional bodies for less error-prone
style.

Signed-off-by: Mark Harris <mark.hsj@gmail.com>

The simple codec round-trip example file in the doc directory
opens an input and output pcm file. It was working fine on
POSIX systems, but not on Windows, which treats text files
differently.

This is confusing in a example, so it's better to add an
explicit binary flag to the fopen() calls. This does nothing
on unix-like systems, but should make the example work for
developers on Windows.

Thanks to Wavesonics who reported this on irc.

Signed-off-by: Mark Harris <mark.hsj@gmail.com>

Hector Martin authored 4 years ago and Mark Harris committed 4 years ago

Signed-off-by: Hector Martin <marcan@marcan.st>
Signed-off-by: Mark Harris <mark.hsj@gmail.com>

Describe builds for the gitlab continuous integration
service runners. This does a trial build under both
autotools and cmake, so we get some coverage for changes
on that hosting platform.

After the same script in the vorbis and ogg projects.

Signed-off-by: Mark Harris <mark.hsj@gmail.com>

Hua, Chunbo authored 4 years ago and Ralph Giles committed 4 years ago

Signed-off-by: Ralph Giles <giles@thaumas.net>

Marcus Asteborg authored 4 years ago and Jean-Marc Valin committed 4 years ago

Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>

Marcus Asteborg authored 4 years ago and Jean-Marc Valin committed 4 years ago

Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>

Marcus Asteborg authored 4 years ago and Jean-Marc Valin committed 4 years ago

Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>

Nathaniel R. Lewis authored 4 years ago and Jean-Marc Valin committed 4 years ago

    - Fix typo in OPUS_USE_NEON description.

    - Set OPUS_PRESUME_NEON for iOS platforms as all armv7 and higher
      iOS devices support NEON.

    - Fix detection of aarch64 for OPUS_CPU_ARM and adding sources from
      celt_sources_arm. (previously would miss armcpu.c and arm_celt_map.c)

    - Change "armv7-a" to "arm" in MATCHES checks against
      CMAKE_SYSTEM_PROCESSOR as systems like the RPi3 report as
      "armv7l".

    - Rename OPUS_MAY_SUPPORT_NEON to OPUS_MAY_HAVE_NEON as this name is
      used everywhere else in the CMake build system. Without this,
      runtime capability detection is broken on aarch64.

Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca>