Skip to content

Releases: Vercidium/openal-soft-github-actions

OpenAL Soft v-f6e6df25

Pre-release

Choose a tag to compare

@github-actions github-actions released this 01 Sep 23:53

Commits

  • 10d962c: Add comments for the strong number's difference type (Chris Robinson)
  • 50c24c1: Add some math functions for strong number types (Chris Robinson)
  • a719fcc: Make f64 a strong number type (Chris Robinson)
  • 8c25ee8: Make f32 a strong number type (Chris Robinson)
  • 55d1c21: Make i32 and u32 strong number types (Chris Robinson)
  • 2665380: Avoid u32 for the macOS-compatible atomic flag (Chris Robinson)
  • b4a09b5: Support fp-to-integer in .saturate_as() (Chris Robinson)
  • 9a144be: Use strong numbers to generate bsinc tables (Chris Robinson)
  • e40a32b: Use copy_n instead of views::counted (Chris Robinson)
  • 6ea1723: Prefer formatting strong numbers as unsigned int when possible (Chris Robinson)
  • e78dcbb: Use strong numbers for BsincState (Chris Robinson)
  • c73f461: Use standard parameter names (Chris Robinson)
  • 4db8716: Allow implicit construction from weak numbers (Chris Robinson)
  • a619947: Remove an unneeded difference_type operation (Chris Robinson)
  • 1e7eae4: Limit filter reference frequency to 0.49 (Chris Robinson)
  • f06e70f: Avoid an unnecessary cast (Chris Robinson)
  • f120be6: Avoid 'using' for operator= (Chris Robinson)
  • 6621b93: Rename a template parameter for clarity (Chris Robinson)
  • 222a90b: Try to improve handling constants with strong numbers (Chris Robinson)
  • 30339ef: Rename can_narrow to might_narrow for clarity (Chris Robinson)
  • 50a777b: Implement 3D processing for the distortion effect (Chris Robinson)
  • dc0e3db: Implement 3D processing for the chorus/flanger effects (Chris Robinson)
  • 4578b74: Implement 3D processing for the pitch shifter effect (Chris Robinson)
  • cecc2ea: Implement 3D processing for the frequency shifter effect (Chris Robinson)
  • baa2048: Adding support for X714ChanMap to CoreAudio (kcat#1245) (Colin Cornaby) #1245
  • 06b8f74: Fix an outdated comment (Chris Robinson)
  • 8f0054f: Adjust the A-Format conversion for some effects processing (Chris Robinson)
  • eea2757: Apply the pitch shifter directly to B-Format (Chris Robinson)
  • c7ffa88: Fix an old incorrect comment (Chris Robinson)
  • 2e5cdf0: Fix double-processing of first sample in Multi2Mono channel converter (kcat#1246) (Christian Schelle) #1246
  • b2255be: Duplicate and recombine a couple loops to remove a temp buffer (Chris Robinson)
  • a4da1d6: Rename some template parameter names for clarity (Chris Robinson)
  • 94d9f3b: Move extension loading for examples to a shared function (Chris Robinson)
  • 0a3870f: Add a time stretch option (-tstretch) to alffplay (Chris Robinson)
  • 30eb04b: Avoid recreating the pitch shifter for every track in alffplay (Chris Robinson)
  • 32759d3: Make sure the effect slot is deleted when quitting (Chris Robinson)
  • a9cdd5f: Get the actual output size of swr_convert (Chris Robinson)
  • f2c3f05: Fix a potential crash with builds using older PipeWire headers (Chris Robinson)
  • 6af654c: Use a newer libmysofa release tag for CI builds (Chris Robinson)
  • c4e1984: Update zlib version used by libmysofa (Chris Robinson)
  • 4de16fc: Remove an outdated DLL from the CI artifacts (Chris Robinson)
  • d06a87d: Add a macro for specifying a function as nonblocking (Chris Robinson)
  • 19b81ab: Mark the filter functions as nonblocking (Chris Robinson)
  • 06bb49e: Remove an explicit static member function (Chris Robinson)
  • 9f5deac: Make isize a strong number type (Chris Robinson)
  • c1884b5: Use strong number types for decoding ADPCM (Chris Robinson)
  • a51260c: Fix a cast when compiling against PipeWire 0.3.50 headers (Chris Robinson)
  • 8780402: Add an ubuntu-22.04 build to CI (Chris Robinson)
  • 810c8b1: Remove constexpr from a function that can't be (Chris Robinson)
  • b48db84: Try to workaround a GCC 11 issue with string_view construction (Chris Robinson)
  • bb2f88e: Try to work around more older GCC issues (Chris Robinson)
  • 2b38916: Changes to support arm64ec on msvc. (kcat#1250) (aalekhm) #1250
  • 1c31c0f: Check for NEON first (Chris Robinson)
  • 00fed14: Use GCC 12 for the Ubuntu 22.04 CI build (Chris Robinson)
  • 53db0bf: Cast a value to what it's being accumulated into (Chris Robinson)
  • 8edf24b: Remove -Warith-conversion (Chris Robinson)
  • 6f39e42: Rename make_from to from (Chris Robinson)
  • bbe3466: Support ops between strong and compatible weak number types (Chris Robinson)
  • 21a1abc: Add a -tune option to alffplay (Chris Robinson)
  • 3eb2da7: Rename UInt and make it an alias (Chris Robinson)
  • 42b56d7: Avoid some explicit template parameters (Chris Robinson)
  • 2ac547c: Define the common type for strong numbers (Chris Robinson)
  • a22a0f1: Make bit_pack methods for 1, 4, and 8-byte types (Chris Robinson)
  • 2a425c0: Avoid an unnecessary size_t (Chris Robinson)
  • 7939abe: Remove excess requires clauses (Chris Robinson)
  • 29a96ab: Add common types for equally-sized fp and integer types (Chris Robinson)
  • 3dc03e2: Reduce the magnitude of the phase difference (Chris Robinson)
  • f0e7747: Simplify the phase delta calculation a bit (Chris Robinson)
  • 171360a: Fix some mismatched numeric range types (Chris Robinson)
  • 3bae338: Share the Hann window between effects that use it (Chris Robinson)
  • aa2ea6c: Make the Hann window object itself an array (Chris Robinson)
  • 4e8d62a: Separate phase_shifter.h from allpass_conv.hpp (Chris Robinson)
  • 95c6ccf: Rename phase_shifter.h to phase_shifter.hpp (Chris Robinson)
  • aad6889: Enable vector acceleration on OpenPOWER (powerpc64le) systems (kcat#1252) (Timothy Pearson) #1252
  • 7466303: Make a module for phase_shifter (Chris Robinson)
  • cc4ed9c: Make some functions constexpr (Chris Robinson)
  • 9b972a4: Remove an incorrect self-include (Chris Robinson)
  • 18197fa: Clean up some includes (Chris Robinson)
  • a68d0c7: Add a module for logging functions (Chris Robinson)
  • 2620c42: Fix fp->integer saturating casts (Chris Robinson)
  • b3e89a7: Avoid some uses of usize (Chris Robinson)
  • c13fa64: Move DecoderBase to a separate header (Chris Robinson)
  • db74ed0: Update the type of a couple arrays (Chris Robinson)
  • 20b7711: Add a Super Stereo decoder for TSME output (Chris Robinson)
  • 24fa1fe: Make aliases for specific struct templates (Chris Robinson)
  • 06803e9: Improve calculating the Blackman window (Chris Robinson)
  • 7ed9031: Don't include gsl in altypes.hpp (Chris Robinson)
  • d10a1ea: Check for types derived from isize for common_type (Chris Robinson)
  • e091b64: Adjust the Hann window scaling (Chris Robinson)
  • 1a1cbac: Calculate a half-size once as needed (Chris Robinson)
  • f12ba6c: Avoid some usize types (Chris Robinson)
  • 8c59075: Don't include unused headers with modules enabled (Chris Robinson)
  • 54cc96d: Use file sets in CMake for headers when supported (Chris Robinson)
  • fb5d9f4: Use a separate target for the router modules (Chris Robinson)
  • 2e8b5a4: Put generated headers in a file set with the correct base dir (Chris Robinson)
  • 8683495: Update some GitHub CI actions (Chris Robinson)
  • 87d0877: Rename some CMake targets for clarity (Chris Robinson)
  • 0b74bf5: Move router.cppm (Chris Robinson)
  • e7d2221: Mark a module function delcaration as a normal C++ function (Chris Robinson)
  • 9927d89: Make the generated HRTF header a proper header (Chris Robinson)
  • 1bfe68e: Correctly depend on HRTF data file targets (Chris Robinson)
  • 8a1212d: Avoid extraneous CMake variables (Chris Robinson)
  • c8d6e49: Use list(APPEND ...) to append items to lists in cmake (Chris Robinson)
  • 5b65cb1: Make and start using a bitset for scoped enum types (Chris Robinson)
  • a8e6b8d: Use scoped-enum-aware bitsets (Chris Robinson)
  • c3e1d04: Give the enum type itself to al::bitset (Chris Robinson)
  • 7a038f3: Avoid a template specialization (Chris Robinson)
  • 615c8c5: Pass the declaration attribute to the DECL_FUNC* macros (Chris Robinson)
  • b5d0fb0: Use a forwarding macro to define AL functions (Chris Robinson)
  • 5a7f551: Avoid using protected visibility (Chris Robinson)
  • 126aa77: Only make hidden symbol aliases for Linux (Chris Robinson)
  • 0e64488: Try to work around an MSVC issue with VA_ARGS (Chris Robinson)
  • 1d52120: Fix missing break statement for UWP (Chris Robinson)
  • f676c2b: Allow disabling dynamic loading of optional libs on Windows (kcat#1256) (Alexander) #1256
  • 1fed18e: Give the EAX extension functions C linkage (Chris Robinson)
  • ed87ca3: Simplify creating hidden symbol aliases (Chris Robinson)
  • 234c60d: Convert usize to a strong number type (Chris Robinson)
  • e9c479e: Workaround MSVC not converting an int to ConstantNum (Chris Robinson)
  • 4ee390c: Avoid convert_to for conversions that can't narrow (Chris Robinson)
  • 3ae4cff: Fix signed->unsigned strong integer saturating cast (Chris Robinson)
  • 3a34204: Move some alffplay main state to a separate struct (Chris Robinson)
  • c08bd77: Remove an unnecessary final action (Chris Robinson)
  • 4263f85: Make the strong number types themselves nodiscard (Chris Robinson)
  • e7934e5: Pass compressor params as a struct (Chris Robinson)
  • a2956d6: Use strong numbers in the main compressor (Chris Robinson)
  • 55ce30e: Include the value and type in narrowing exception messages (Chris Robinson)
  • a043756: Move formatting and exception code out of altypes.hpp (Chris Robinson)
  • 416b3f8: Add some missing includes (Chris Robinson)
  • c949d21: Fix a using declaration (Chris Robinson)
  • 390069d: Add another missing include (Chris Robinson)
  • 33e2286: Append a counter to PipeWire devices with duplicate names (Chris Robinson)
  • 5fbcf96: Use generate to initialize an array object's field (Chris Robinson)
  • e56f298: Make isize and usize aliases for other strong types (Chris Robinson)
  • 1b2a52d: Move a struct definition to where it's used (Chris Robinson)
  • a41fbba: Fix the source's default super stereo width value (Chris Robinson)
  • 595c3c8: Use an enum class for the source property enums (Chris Robinson)
  • fba0c5e: Avoid some uses o...
Read more