This is quite a significant update for CPAL, including a number of breaking changes. Here is a list of the breaking changes along with links to where you can find more information: - A `Host` API has been introduced in #289 along with a follow-up refactor in #295. Please see the examples for a demonstration of how to update your code. The necessary changes should hopefully be minimal. If this has caused you any major difficulty please let us know in an issue! - An ASIO host has been introduced in #292. This adds support for Steinberg's ASIO audio driver API on Windows. Please see the ASIO section of the README for more information on how to setup CPAL with ASIO support for your project. - The user callback API has been overhauled to emit `StreamEvent`s rather than buffers in order to support handling stream errors. #288. - Error handling in general was overhauled in #286. This meant taking advantage of the `failure` crate and adding support for backend-specific errors with custom messages. Many unnecessary `panic!`s have been removed, but a few remain that would indicate bugs in CPAL. In general, checking out the updated examples will be the easiest way to get a quick overview on how you can update your own code for these changes. The CHANGELOG.md has been updated to include these changes.
3.3 KiB
3.3 KiB
Unreleased
Version 0.10.0 (2019-07-05)
- core-foundation-sys and coreaudio-rs version bumps.
- Add an ASIO host, available under Windows.
- Introduce a new Host API, adding support for alternative audio APIs.
- Remove sleep loop on macOS in favour of using a
Condvar. - Allow users to handle stream callback errors with a new
StreamEventtype. - Overhaul error handling throughout the crate.
- Remove unnecessary Mutex from ALSA and WASAPI backends in favour of channels.
- Remove
panic!from OutputBuffer Deref impl as it is no longer necessary.
Version 0.9.0 (2019-06-06)
- Better buffer handling
- Fix logic error in frame/sample size
- Added error handling for unknown ALSA device errors
- Fix resuming a paused stream on Windows (wasapi).
- Implement
default_output_formatfor emscripten backend.
Version 0.8.1 (2018-03-18)
- Fix the handling of non-default sample rates for coreaudio input streams.
Version 0.8.0 (2018-02-15)
- Add
record_wav.rsexample. Records 3 seconds to$CARGO_MANIFEST_DIR/recorded.wavusing default input device. - Update
enumerate.rsexample to display default input/output devices and formats. - Add input stream support to coreaudio, alsa and windows backends.
- Introduce
StreamDatatype for handling either input or output streams inEventLoop::runcallback. - Add
Device::supported_{input/output}_formatsmethods. - Add
Device::default_{input/output}_formatmethods. - Add
default_{input/output}_devicefunctions. - Replace usage of
VoicewithStreamthroughout the crate. - Remove
Endpointin favour ofDevicefor supporting both input and output streams.
Version 0.7.0 (2018-02-04)
- Rename
ChannelsCounttoChannelCount. - Rename
SamplesRatetoSampleRate. - Rename the
min_samples_ratefield ofSupportedFormattomin_sample_rate - Rename the
with_max_samples_rate()method ofSupportedFormattowith_max_sample_rate() - Rename the
samples_ratefield ofFormattosample_rate - Changed the type of the
channelsfield of theSupportedFormatstruct fromVec<ChannelPosition>toChannelCount(an alias tou16) - Remove unused ChannelPosition API.
- Implement
EndpointandFormatEnumeration for macos. - Implement format handling for macos
build_voicemethod.
Version 0.6.0 (2017-12-11)
- Changed the emscripten backend to consume less CPU.
- Added improvements to the crate documentation.
- Implement
pauseandplayfor ALSA backend. - Reduced the number of allocations in the CoreAudio backend.
- Fixes for macos build (#186, #189).
Version 0.5.1 (2017-10-21)
- Added
Sample::to_i16(),Sample::to_u16()andSample::from.
Version 0.5.0 (2017-10-21)
- Removed the dependency on the
futureslibrary. - Removed the
VoiceandSamplesStreamtypes. - Added
EventLoop::build_voice,EventLoop::destroy_voice,EventLoop::play, andEventLoop::pausethat can be used to create, destroy, play and pause voices. - Added a
VoiceIdstruct that is now used to identify a voice owned by anEventLoop. - Changed
EventLoop::run()to take a callback that is called whenever a voice requires sound data. - Changed
supported_formats()to produce a list ofSupportedFormatinstead ofFormat. ASupportedFormatmust then be turned into aFormatin order to build a voice.