aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Check that atomic_load works with a const _AtomicChris Robinson2014-09-051-1/+1
| | | | | | | The original C11 spec does not allow atomic_load to work on const _Atomic variables, which we sometimes do in alGet* methods, despite the fact that it does not modify the variable. An update to the C spec corrects this oversight, and GCC 4.9 has allowed it anyway, while Clang 3.4 does not.
* Fix the __get_cpuid cmake checkChris Robinson2014-09-051-1/+1
|
* Use a standard pointer-sized integer typeChris Robinson2014-09-041-1/+1
|
* Make ExchangeInt and ExchangePtr non-atomicChris Robinson2014-09-031-41/+23
|
* Make the fontsound's buffer and link fields atomicChris Robinson2014-09-034-19/+29
|
* Protect alProcessUpdatesSOFT with a lockChris Robinson2014-09-031-2/+2
|
* Use proper atomics for the thunk arrayChris Robinson2014-09-031-12/+13
|
* Make the buffer's pack and unpack properties atomicChris Robinson2014-09-032-11/+11
|
* Fix Neon mixer definitionChris Robinson2014-08-311-2/+2
|
* Use al_calloc/al_free to allocate contexts and voicesChris Robinson2014-08-301-6/+6
|
* Setup the HRTF format before tracing the pre-reset formatChris Robinson2014-08-291-17/+17
|
* Check mmdevapi device ids to match the default deviceChris Robinson2014-08-281-19/+36
| | | | | Seems Windows can return different IMMDevice object pointers for the same endpoint.
* Return the correct default capture device nameChris Robinson2014-08-261-1/+1
|
* Check the given CoreAudio capture device nameChris Robinson2014-08-261-0/+5
|
* Remove a couple unnecessary typedefsChris Robinson2014-08-242-5/+2
|
* Convert the wave writer backend to the new APIChris Robinson2014-08-244-115/+182
|
* Use al_malloc/al_free for default allocatorsChris Robinson2014-08-241-2/+2
|
* Include the common sources when building staticallyChris Robinson2014-08-231-1/+5
|
* Rename activesource to voiceChris Robinson2014-08-218-210/+208
|
* Use an array of objects for active sources instead of pointersChris Robinson2014-08-215-66/+48
|
* Use a NULL source for inactive activesourcesChris Robinson2014-08-217-52/+64
| | | | Also only access the activesource's source field once per update.
* Use the current binary dir for shared function checks tooChris Robinson2014-08-201-7/+7
|
* Use directories relative to current project in CheckFileOffsetBits.cmakeJesper Särnesjö2014-08-201-4/+4
| | | | Enables building OpenAL Soft in a subdirectory of another project.
* Support brace-enclosed environment variable namesChris Robinson2014-08-191-0/+9
| | | | | | | This makes it possible to append alpha-numeric characters directly to an environment variable value, e.g. ${FOO}bar will use "FOO" as the variable name and keep the "bar" as-is, whereas $FOObar will take "FOObar" as the variable name.
* Update COPYING to the latest ↵François Cami2014-08-1847-105/+102
| | | | https://www.gnu.org/licenses/old-licenses/lgpl-2.0.txt to fix the FSF' address Fix the FSF' address in the source
* Use the POSITION_INDEPENDENT_CODE target property when availableChris Robinson2014-08-161-7/+12
|
* Don't force specific compile flags for the different build typesChris Robinson2014-08-161-17/+7
|
* Don't try to include stdalign.h if C11 _Alignas isn't availableChris Robinson2014-08-161-1/+1
| | | | | | Some compilers will allow including stdalign.h, and even define alignas to _Alignas, even if that C11 feature is unavailable (e.g. because it requires a suitable -std= setting).
* Search for the correct include and lib directories for the DX SDKChris Robinson2014-08-152-5/+4
| | | | And only set them when needed.
* Add a changelog, detailing notable changes for each releaseChris Robinson2014-08-151-0/+100
|
* Release 1.16.0openal-soft-1.16.0Chris Robinson2014-08-141-2/+2
|
* ALC_SOFT_pause_device is finishedChris Robinson2014-08-123-11/+11
|
* Assume SSE is available if building with support and no run-time checkingChris Robinson2014-08-111-0/+12
|
* Try the __cpuid intrinsic if GCC's __get_cpuid isn't availableChris Robinson2014-08-113-10/+59
|
* Check for GCC's __get_cpuid before using itChris Robinson2014-08-113-2/+14
|
* Only support dsound and mmdevapi on WindowsChris Robinson2014-08-111-36/+35
|
* Remove a couple unnecessary HAVE_*_H header definesChris Robinson2014-08-111-6/+0
| | | | | The xmmintrin.h and arm_neon.h headers are always available with SSE and Neon support, respectively.
* Don't try to generate .def or .lib files when building staticallyChris Robinson2014-08-101-18/+20
|
* Add an option to build an import .lib with dlltool on mingwChris Robinson2014-08-101-22/+48
| | | | | | Also make sure the ordinals are stripped from the .def file (using sed), so that the generated .lib will link apps to the DLL using function names instead of ordinals.
* Simplify some vector size range checksChris Robinson2014-08-102-17/+19
|
* Give NULL for null-vector iteratorsChris Robinson2014-08-102-3/+5
|
* Pass pointer-to-vector types as char* instead of void*Chris Robinson2014-08-102-17/+16
| | | | | C aliasing rules only allow char* to alias an otherwise-incompatible type, rather than void*.
* Properly set FluidSynth's include dir for older cmake versionsChris Robinson2014-08-101-1/+1
|
* Create a .def file when building with MinGWChris Robinson2014-08-101-0/+3
| | | | | Preferably we'd tell CMake to make it itself instead of specifiying compiler- specific flags, but there doesn't seem to be a way to.
* Use VECTOR_FIND_IF and VECTOR_FOR_EACH instead of manual loopsChris Robinson2014-08-094-106/+57
|
* Use the default input device for portaudio's default capture deviceChris Robinson2014-08-081-1/+4
|
* Modify CPU extensions and effect options to be enabled when checkedChris Robinson2014-08-072-63/+105
| | | | | It looks nicer and more featured for options to be checked when they're allowed, rather than when disallowed.
* Add a 48khz HRTF data setChris Robinson2014-08-072-0/+1
| | | | | Made by simply using makehrtf's resampling option with the 44.1hz KEMAR data set.
* Disable the autowah effectChris Robinson2014-08-062-0/+6
| | | | | | | | There's apparently some issues with it causing noise or killing the output. It might be due to the per-sample changes being too harsh for the filter to keep up with, but it's not something I can take care of in time for release. This commit should be reverted after release when work on fixing it can resume.
* Make the DYNLOAD LoadFSynth function non-inlineChris Robinson2014-08-051-1/+1
|
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-15 12:08:31 +0000