* Don't write UT features until the end of TX/RX.
* Attempt 2 at preventing spurious Linux asan failures.
* Add PR #975 to changelog.
* Revert "Attempt 2 at preventing spurious Linux asan failures."
This reverts commit 8f53d1a980.
* Another attempt at preventing spurious asan failures.
* Revert back to k6aq.net hosted wheels.
* Resolve warnings in pipeline tests.
* ResampleStep: skip processing if no samples available.
* TapStep: execute in separate thread.
* Fix issue causing mic audio to stop when using Speex.
* Add a bit more time between loads.
* Use -O3 for libsamplerate to see if that helps.
* Revert "Add a bit more time between loads."
This reverts commit 1273d6531e.
* Disable content caching during CI build.
* Add printout of sysctl for debugging.
* Try speexdsp resampling to see if we can reduce CPU usage.
* Slightly increase quality.
* Increase quality again.
* Fix monitor functions not producing audio.
* Reenable Instruments tracing.
* Experiment: update librade to grab sync and SNR while running RX.
* Update SNR retrieval to make sure it happens on RX thread.
* Bump up resample quality again.
* Go back to libsamplerate.
* Use same semaphore wait time calculation for macOS as with other platforms.
* Try 20ms time quantum.
* Try 40ms.
* Back to 60ms.
* Wait needs to be based on start of work.
* Revert "Go back to libsamplerate."
This reverts commit 86d1c640ff.
* Try maximum quality on speex resampler first.
* Go back to calling fwrite() immediately on FIFO read.
* Revert resampler changes, again.
* Try experimental RADE branch again.
* Out of left field: try 2.9ms quantum like what Chrome uses.
* Improve reliability of Linux rade_loss test.
* Smooth out longer than expected semaphore waits on macOS.
* Fix compiler errors.
* Put quantum back to 60ms.
* Fix compiler error.
* More experimentation with time quantums.
* Use relaxed atomic semantics for macOS worker counter.
* Back to 2.9ms due to mpp test failures.
* Disable processing stats.
* Time quantum should match buffer size.
* Remove unneeded variable.
* Time quantum was accidentally getting set to 0.
* Try to reduce standard deviation for macOS wait time.
* Revert previous change, try just halving the wait time.
* Try to reduce likelihood of scheduler contention on macOS.
* macOS: Set lower thread QoS for non-audio threads.
* Using infinite wait for ThreadedObject.
* Force lowest possible latency.
* We can reference the activity ID directly.
* Avoid crash on stop from previous changes.
* Revert recent PulseAudioDevice changes in case those are causing 22.04 tests to fail.
* Try different macOS flag to improve responsiveness.
* Remove memset call from TX handling.
* Disable real-time threading on macOS.
* Use setiopolicy_np() to make sure writing RADE features doesn't interfere with the rest of the app.
* Resample input file for rade_loss test to lower CPU usage.
* Try soxr resampler for further CPU usage reductions.
* Forgot missed file.
* On macOS CI runs, we should also use the Homebrew soxr package.
* Revert soxr adoption, does not meet real time safety guidelines.
* Try deferring feature file writes until the end again.
* Turn on PYTHONOPTIMIZE.
* Read/write entire blocks from FIFO, not just single samples.
* Warning cleanup.
* Allow partial output if FIFO gets empty.
* Remove unneeded memset().
* Fix race condition preventing user from properly being hidden from FreeDV Reporter on startup.
* Downgrade additional pipeline threads to Utility QoS.
* Try malloc-free FIFO for RADE steps.
* Use 2048 samples instead.
* Use PreAllocatedFIFO in RADE steps to avoid cache misses.
* Use MKL enabled version of numpy for Intel macOS.
* Try Intel MKL version of PyTorch for macOS.
* Fix GH compiler issue on macOS ARM.
* Replace sleep_for with semaphore for TapStep.
* Only signal semaphore every 100ms like with previous sleep_based impl.
* Replace additional sleeps with semaphore waits.
* Fix compiler error.
* Adjust timing thresholds for dumping feature debug data.
* macOS: reenable real-time threading as a test.
* Make sure both RX and TX threads start processing at the same time.
* Update missed shared_ptr reference.
* Remove hidden allocation due to use of lambdas.
* Undo FIFO size reduction.
* Suppress spurious PlaybackStep completion messages.
* Fix CI compiler error.
* Use Time Profiler to get a better idea as to why RADE steps intermittently take longer to execute.
* g_voice_keyer_tx should use acquire/release.
* Revert "Use Time Profiler to get a better idea as to why RADE steps intermittently take longer to execute."
This reverts commit 711f83be0c.
* Only grab TX status once through the GUI timer.
* Only get sync once per timer run.
* Remove use of wxThread from TxRxThread.
* Use relaxed for retrieving sync status.
* Test: disable Mic In plot.
* Try using relaxed memory ordering.
* Revert "Use relaxed for retrieving sync status."
This reverts commit 364bc62e8c.
* Revert "Try using relaxed memory ordering."
This reverts commit bac5a8a799.
* One more attempt at using Speex resampler.
* Add optimizations for PlotScalar in case it's impacting test runs.
* Reenable lower quality level for plots.
* Try quality level 6.
* Actually fix spurious playback complete log messages.
* Back to quality level 10.
* macOS: Begin wait from end of work.
* Try DISPATCH_TIME_FOREVER.
* Revert "Try DISPATCH_TIME_FOREVER."
This reverts commit a913e8573e.
* Revert "macOS: Begin wait from end of work."
This reverts commit 7c719a3e3c.
* Add crazy hack to inhibit the UI thread while the pipeline is executing.
* Try checking pipeline spinlock every time we have to redraw something.
* Try actually sleeping instead of yielding.
* Revert crazy hack.
* Use separate RX thread instead of ThreadedTimer.
* Begin splitting GUI timer work into separate timers.
* Only update sync and TX state on the timers that need it.
* Defer item updates until highlights are updated.
* Delete accidentally added files.
* Try smaller block sizes.
* Try bigger block size for comparison.
* Re-disable real-time threading.
* Preallocate what we need to output RX/TX features.
* No need to duplicate bitmap to draw onto graphics context. (Reduces number of mmap() calls req'd on macOS.)
* Reenable real time threading.
* Force 50ms time quantum regardless of buffer size.
* Update reporting timers to attempt to guarantee that only one is active at a time.
* Add safety offset to BlackHole virtual audio device.
* WIP: Reduce CPU usage in PlotWaterfall.
* Fix compiler errors and rendering bugs.
* Fix compile error for real this time.
* Don't update FreeDV Reporter window if it's not actually visible.
* Back to 60ms time quantum.
* Back to 50ms.
* Slightly adjust how IsShownOnScreen() suppresses updates.
* Just use an array instead of map for PlotScalar.
* Prevent propagation of changes to FreeDV Reporter window unless something actually changed.
* Tweak Speex resampler quality again.
* Resample level 7.
* Don't need to recreate ThreadedTimer thread every time we need to restart.
* Try not running xctrace to see if we're close to fixing the issues.
* Wait a bit more time between loop iterations when playing UT recordings.
* Revert "Try not running xctrace to see if we're close to fixing the issues."
This reverts commit decff7a5b2.
* Add Time Profiler to xctrace files.
* Make all FreeDV Reporter rows invisible when window is not visible.
* Make FreeDV Reporter object synchronous.
* Add support for queuing on parent object thread instead of own.
* Fix Linux segfault.
* Use macOS Dispatch for ThreadedObject.
* Try disabling threading in TapStep.
* Reenable threading for TapStep.
* More quality level tweaking.
* Re-disable separate threading.
* Only allocate memory for feature storage if necessary.
* Use GCD for TapStep.
* Fix compiler error on Linux.
* Tweak maximum sizes of FIFOs based on recorded max usage during ctests.
* Additional adjustments to FIFO sizes.
* More tweaks of FIFO sizes for Linux.
* Disable workflow telemtry for macOS runs.
* Fix syntax error.
* Revert "Use GCD for TapStep."
This reverts commit e1522a7ab3.
* Fix memory violations when using PlotScalar.
* Fix rendering bugs from previous optimization.
* Try only returning numInputSamples at a time from vocoders.
* Revert src/pipeline to last passed build.
* Quality level 8 works better locally.
* Try 40ms blocks on macOS.
* Reduce the number of context switches while handling FreeDV Reporter traffic.
* Try reducing FIFO sizes again.
* Try reverting PyTorch upgrade.
* Didn't fully revert PyTorch changes.
* SoX is ignoring errors from Core Audio, use buffer sizes known to work.
* SoX divides by 8 to produce frame size sent to CoreAudio.
* Reset resampler back to last known working level.
* Update Linux/macOS audio code to match Windows implementation more closely.
* Try 40ms blocks again.
* Add middle click reset support for speaker/mic level.
* Revert "Try 40ms blocks again."
This reverts commit 3ca16e0335.
* Have FreeDV itself load test.wav due to performance limitations in GH environment.
* GitHub Actions: Disable iCloud (!)
* Need to set config value and not the slider (middle-click).
* Disable xctrace collection.
* Disable mic/speaker level slider unless modem is active.
* Return to libsamplerate for ResampleStep.
* MacAudioDevice cleanup.
* Remove unneeded warning message.
* Need memory_order_acquire for legacy sync state.
* Try RADE main branch again.
* Debugging: see if anything else is running that shouldn't be.
* GitHub: disable Spotlight too.
* Fix typo in README.
* Reenable macOS runner debugging.
* Include mimalloc for a bit extra boost.
* Disable mimalloc ctests.
* Disable macOS debugging again.
* mimalloc wasn't actually being used on macOS.
* mimalloc: decrease purge interval.
* Fix hidden memory allocation during TX output handling.
* Fix additional hidden memory allocations.
* Unit test: make sure rade_loss shows sync state.
* More hidden memory allocation removals.
* Further reduce duplicate memory allocations.
* Fix voice keyer bug reported by @barjac.
* More cleanup of unneeded allocations.
* Fix mini versions of scalar plots (discovered during testing).
* Remove some wxAsyncCallEventFunctor instances from gprof profile.
* Fix compiler error.
* Additional refactoring and memory allocation reduction.
* Extend purge delay a bit just in case.
* Allow GitHub Actions to re-execute test in case of an intermittent failure.
* Fix typo in rade_reporting test.
* Minor reduction in allocations during FreeDV Reporter RX handling.
* Use std::atomic for PlaybackStep.
* Revert "Use std::atomic for PlaybackStep."
This reverts commit 0ee65a38eb.
* Return nullptr to PlaybackStep if not actively playing.
* PlaybackStep: wait for main thread to finish stopping TX instead of previous approach.
* Fix rendering issue going between TX and RX.
* Fix issue where VK monitor plays previously TX audio prior to current VK cycle.
* Fix segfault in Hamlib on pressing Stop.
* Add missed destroying check while setting PTT.
* Additional cleanup in rig_control due to previous commits.
* Fix PortAudio compile issue on macOS.
* Tweak macOS audio device initialization.
* Add resample shortcut logic to avoid going through libsamplerate if in/out rates are the same.
* Initialize TX audio devices before RX in case sample rates change.
* Add optional debugging output for AudioPipeline.
* Avoid use of ParallelStep for RADE.
* Add extra protection in RADE TX step just in case.
* Partial revert of createReceivePipeline changes.
* Rewrite macOS audio code to use traditional Core Audio instead of AVAudioEngine.
* Try again at removing ParallelStep from RADE RX.
* Fix missed change in GH actions.
* Fix Linux test failure.
* Previous change was in the wrong place.
* Bring back CFString method of retrieving device names.
* Use higher device latencies if we detect a Bluetooth device.
* Try 30ms interval time to avoid GH failures.
* Revert "Try 30ms interval time to avoid GH failures."
This reverts commit f6bd8e1a57.
* Use integer math for calculating block sizes.
* Also use integer math for FRAME_DURATION.
* Add additional timeout for Windows builds.
* Rename MacAudioDevice.mm -> .cpp.
* Ignore required misspelling.
* Add PR #971 to changelog.
* Fix Registry update issue during GH Windows tests.
* Allow rade_loss ctest to work better in RT context.
* Attempt to avoid the spurious Reporter-related LeakSanitizer failures on shutdown.
* We don't need to have latency so high for Bluetooth.
* Let's try memcpy for GenericFIFO.
* Another attempt at eliminating the LeakSanitizer error.
* Try previous RX implementation again.
* Revert "Try previous RX implementation again."
This reverts commit cc0f08f2e9.
* Slightly increase block times.
* Make explicit various codec2 dependencies to avoid compile race condition.
* Add PR #957 to changelog.
* No need to use build_rade as the dependency if we're using rade.
* Forgot to check in a change.
* Looks like we need the larger FIFO after all.
* Use SRC_SINC_FASTEST for plots.
Should be low to no-impact as previous RADE testing showed only a slight quality
decrease vs. SRC_SINC_MEDIUM_QUALITY.
* Update changelog.
* Back to SRC_LINEAR.
* Update comments.
* Don't fail fast for Linux builds.
* Add libsamplerate deps where appropriate.
* Add sndfile deps.
* Revert "Add sndfile deps."
This reverts commit dc00a3b7cd.
* Revert "Add libsamplerate deps where appropriate."
This reverts commit 65c4766c20.
* Reenable processing stats.
* Try building with -O3.
* Try building using LTO.
* Revert "Try building using LTO."
This reverts commit 2536a15dd2.
* Revert "Try building with -O3."
This reverts commit 8b9b824a88.
* Revert "Reenable processing stats."
This reverts commit 5b0d31b1a2.
* Experimental changes to resampling to reduce pipeline execution jitter.
* Cap the number of times through the TX and RX loops.
* Faster resampler for ComputeRfSpectrum step (RX only).
* Ensure we're not overrunning FIFOs on RX.
* Force all Linux tests to run to see how many end up failing now.
* Fix RADE reporting test failures.
* Revert "Force all Linux tests to run to see how many end up failing now."
This reverts commit 6ba267b758.
* Only try processing FRAME_DURATION ms at a time for TX.
* Fix intermittent RADE reporting failure.
* Make sure runner can exit ASAP when ParallelStep is being destroyed.
* Make sure we can only do one stop() at a time.
* Force all Linux tests to run to see how many end up failing now.
* Remove caps on number of cycles through TX and RX loops.
* Reenable all tests on Ubuntu 22.04.
* We don't need to call opus_select_arch() every single time through the loop.
* Python 3.11 works better on 22.04.
* Forgot to update AppRun for 3.11.
* Revert fail-fast for AppImage tests.
* Revert other fail-fast.
* Slightly increase PulseAudio latency to avoid RADE dropouts.
* Test change.
* Test change.
* Test change.
* Revert "Test change."
This reverts commit 2eee1e538d.
* Revert "Test change."
This reverts commit 8205c3b097.
* Revert "Test change."
This reverts commit c9ba8b45b2.
* Revert latency changes.
* For some reason, increasing the wait in the UT code makes the RADE test pass more reliably on 22.04.
* Increase wait time to 5s.
* Revert "Increase wait time to 5s."
This reverts commit 1ca459a02a.
* Revert "For some reason, increasing the wait in the UT code makes the RADE test pass more reliably on 22.04."
This reverts commit f267814b50.
* Try another way of increasing sleep time.
* Revert "Try another way of increasing sleep time."
This reverts commit 5a11c59167.
* Re-arrange sleeps.
* Increase time again.
* Revert "Increase time again."
This reverts commit 44073477ef.
* Revert "Re-arrange sleeps."
This reverts commit edbe451601.
* Wrap globals relevant to TX in atomic.
* Slightly increase default FIFO size.
* Trigger build.
* Trigger build.
* Trigger build.
* Trigger build.
* Trigger build.
* Revert "Trigger build."
This reverts commit 7210e745de.
* Revert "Trigger build."
This reverts commit 3fc2059d76.
* Revert "Trigger build."
This reverts commit 66d64cfe8a.
* Revert "Trigger build."
This reverts commit df745e8a77.
* Revert "Trigger build."
This reverts commit 80945de3b8.
* Add macOS 15 to GitHub tests.
* Code cleanup.
* Create ifdefs for disabled logic.
* Add PR #949 to changelog.
* See if reducing default FIFO size will work.
* Initial check-in of files from Peter's repo.
* NO_STRIP isn't necessary on Ubuntu.
* Build AppImage as part of CI workflow.
* Try copying user's Python libraries.
* Force setting PYTHONHOME/PYTHONPATH inside FreeDV code.
* Fix PYTHONHOME paths.
* Clean up unneeded code.
* Have GitHub sanity check AppImage on clean build agents.
* Forgot to install pipewire and X inside test agent.
* Build AppImage on 22.04 instead due to glibc incompatibilities.
* apt update prior to building AppImage.
* Need GTK if we're going to build our own wxWidgets.
* Need to use Python 3.10 when building AppImages on 22.04.
* Restructure Linux tests to hopefully take less time to run.
* Forgot keyword.
* Fix typo.
* Need to run UT in debug mode.
* ASan needs to be run with gcc.
* Use PulseAudio instead of pipewire when testing AppImage on 22.04.
* Include version number in AppImage filename.
* Fix typo.
* Fix filename for AppImage testing.
* Build dependencies ourselves.
* Forgot that we duplicated the Sanity Check step.
* Ensure sanity check step fails if we lose sync.
* Run unit tests on 22.04 as well.
* Add missed GDK_BACKEND export.
* Transmit at least 2x during rade_reporting_* tests.
This is because the logic to calculate RTT between "rig" and FreeDV doesn't
fully complete until after the first TX->RX transition. By transmitting 2x,
we ensure that this logic is also exercised and thus reduce the number of
failures in the GitHub environment.
* latency_msec doesn't work with PulseAudio.
* On 22.04, limit tests to fullduplex_* to avoid GH environment problem.
* Add minimum requirements to user manual.
* Fix rendering issue while debugging test problem in local Ubuntu 22.04 environment.
* Forgot to add a tee command to the 24.04 test.
* Try 10ms latency for Pulse audio.
* Remove std::vector from RADE TX/RX steps just in case that's causing problems.
* CLOCK_MONOTONIC seems to have higher reliability for Linux tests.
* Fix compiler warning.
* WIP user manual update.
* Revert UI fix as it breaks things on macOS.
* Experiment: try to test AppImage in Fedora on GitHub.
* Try again without rtkit.
* Forgot to remove systemctl calls
* FUSE is required for AppImage.
* Use alternate method of running AppImage.
* Try third-party Fedora 41 systemd image to stop dropouts.
* Pass --init to Fedora container.
* Try another container for Fedora.
* No ps in container.
* Use podman to run container instead.
* Add missed DBus packages.
* Forgot to start pipewire.
* Start pipewire after dbus is set up.
* Use different way to start pipewire.
* podman might need to run as root.
* Forgot quote.
* Fix environment not getting passed to podman.
* Experiment: see if Fedora 41 can be tested too.
* Revert "Experiment: see if Fedora 41 can be tested too."
This reverts commit c751ab2c77.
* Bring back previous GUI bugfix but limit it to wxWidgets < 3.2.
* Fix timespec bug discovered during performance profiling.
* Forgot to capture rv.
* Revert back to CLOCK_REALTIME.
* Remove additional sleep on semaphore timeout.
* Need to use sem_clockwait if we're going to use monotonic time.
* Remove comment that no longer applies.
* Add PR #937 to changelog.
* Fix memory leak reported by ASan.
* Deploy GTK dependencies as part of AppImage.
* Add additional protection to avoid dropouts.
* Use deque instead of vector for function queues due to better worst-case complexity.
* Warning cleanup.
* Documentation update.
* Fix macOS compiler error.
* Eliminate the need to install Octave in GitHub environment.
* Have all macOS jobs run at once to save time.
* Fix compiler errors when using wxWidgets 3.0.
* Add GH action to make sure we can compile on Ubuntu 22.04.
* Fix package name.
* Add PR #914 to changelog.
* Ensure only one build per branch happens at a time.
* Make sure RTSan-related code isn't added if not installed.
* Enable RTSan checks in GH Actions.
* Fix syntax error in GH action.
* Fix GCC compile failures.
* Use different GH workflow to get correct Clang installed.
* Split RTSan test into separate build and test steps
* Move LLVM installation to beginning in case there are side effects.
* Prevent dynamic resampling in AudioPipeline when playing back files.
* Mixed up sample calculation, causing tests to fail.
* Additional changes to see if it still hangs.
* Update cmake-macos.yml
* Do completion check after queuing to output FIFO.
* Fix buffer overflow.
* Some additional changes based on local testing.
* GH tests weren't actually being run through the resampler.
* Remove ExclusiveAccessStep files as they're not being used.
* Remove some logging that shouldn't be in RT context.
* Add reset() to IPipelineStep to avoid repeated deallocations and reallocations of the TX pipeline.
* Fix segfault from previous changes.
* Remove RADE specific logging in RT context and suppress RTSan for RADE EOO handling.
* Disable RTSan for RADE text handling.
* Try disabling real-time for macOS.
* Force Windows 2022.
* Revert "Try disabling real-time for macOS."
This reverts commit c8ee871406.
* Experiment: use O(1) heap allocator for Codec2.
* Fix patch commands.
* thread_local is only in C++11 and above.
* Force codec2 folder to be removed every call.
* Update codec2 patch.
* Fix Windows compiler error.
* Oops, should be _aligned_malloc.
* Integrate codec2 into source tree so we don't need to deal with patch.
* Update Codec2 README.
* mbest needs to use o1heap too.
* Remove rtsan disable from FreeDV TX/RX steps.
* Bring back AudioPipeline change previously removed by master merge.
* Result resampler needs to be initialized on construction.
* Create resampler only if input and output rates don't match.
* Increase Codec2 RT heap size.
* Fix Windows failures.
* Ensure separate threads spawned from PipelineStep are also checked by RTSan.
* Support RTSan builds on macOS if user has appropriate LLVM.
* Remove no longer needed patch.
* Remove unneeded resettingThread property.
* Fix issue discovered during testing.
* Test: match Linux and macOS requested latency to Windows.
* Add PR #900 to changelog.
* Remove 800XA, 700C, 2020 and 2020B from UI,
* Remove additional 700C references in UI.
* Remove 2020/AVX detection logic from codebase.
* Remove LPCNet from build scripts.
* Add PR #889 to changelog.
* Forgot to remove LPCNet from Windows installer generation.
* Generate Release build for GH workflows.
* Disable PSK Reporter reporting during test execution as invalid callsigns are used.
* Build type was being overwritten.
* Missed mutex removal.
* Temporarily use Instruments to trace macOS CI execution.
* Minor update.
* Use 2 minute window in xctrace to actually capture wanted events.
* Disable xctrace execution but keep in tests just in case.
* Hide main window during tests to prevent GUI from interfering.
* Experiment: set infifo2 max length to 30 seconds
* Oops, forgot to remove division from previous commit.
* Deallocation of PulseAudio stream should occur while mainloop is locked.
* Revert "Hide main window during tests to prevent GUI from interfering."
This reverts commit 69689bb927.
* Based on online code snippets nullptr for neg is fine.
* Clear PYTHONHOME if set on the user's system (Windows/macOS only).
* Bring back locking in TxRxThread just in case.
* Remove unneeded backslash.
* Revert "Bring back locking in TxRxThread just in case."
This reverts commit fc6a9f8797.
* Manually back out the changes instead of using git checkout.
* Use shared_ptr for LinkStep.
* Try removing the TX mode change mutex again.
* Try moving actual codec activity out of RT context.
* Fix compiler error.
* Fix issue preventing legacy modes from transmitting.
* Another attempt at preventing RX dropouts.
* Try putting mode TX into separate threads again.
* Only limit output samples if input samples were provided.
* Put codec RX back into RT context.
* Put TX on RT context.
* Remove unused code.
* Deallocation should occur only in the destructor.
* Fix Windows spectrum plot formatting issues.
* Prevent buffer overflow during TX/RX processing.
* Bring FRAME_DURATION back down to 10ms.
* We want to feed samples into the RADE/FreeDV TX/RX FIFOs regardless of whether there's room in the output.
* ParallelStep should only feed FRAME_DURATION ms worth of samples at a time.
* Sleep fallback should track FRAME_DURATION.
* Don't execute the parallel steps if there isn't enough room in the output FIFO.
* We really can only inject a portion of EOO samples at a time.
* Queue EOO as a single unit to ensure it gets transmitted as a unit.
* Allow ParallelStep to provide up to 4x the number of input samples as output.
* Back to processing 20ms frames.
* Run GH Linux tests using asan enabled.
* Suppress Python related memory leaks and Hamlib related invalid memory accesses.
* Forgot suppressions= in test execution.
* Add more leak suppressions based on GH output.
* Resolve memory leaks.
* Fix broken RadeTextTest.
* Additional suppressions from GH.
* Go back to processing TxNumSpeechSamples every TX cycle.
* Revert "Go back to processing TxNumSpeechSamples every TX cycle."
This reverts commit abe816d00b.
* Fix 1600 mode dropouts.
* Make TX single-threaded again.
* Revert back to last working pipeline logic (for fullduplex_* anyway).
* Revert most EOO behavior.
* Run AddressSanitizer tests separately as they interfere with real time audio.
* Need to start Xvfb etc. before asan tests.
* Only a subset of tests need to run with asan enabled.
* Fix issue causing PTT input not to work.
* Maybe we don't need to restart servers after all.
* Try using rtkit to ask for high priority instead of RT.
* Clear FIFOs a different way.
* Windows: determine minimum number of audio channels supported.
* WIP patch to wxWidgets to enable color emojis on Windows.
* WIP rewrite of FreeDV Reporter window to use wxDataViewCtrl.
* Apparently we need to associate the model before adding columns.
* Force render the Msg column using Direct2D, causing emojis to display in color.
* Move ReportMessageRenderer to a separate file.
* Only use ReportMessageRenderer on Windows. Also, make sure we don't show ellipsis unless for the Msg field.
* Bring back preservation of sort column.
* Bring back preservation of Msg column width.
* Support Request QSY functionality again.
* Fix rendering issues on macOS.
* Fix Windows compiler errors.
* Fix rendering issue on Windows from previous commits.
* Use correct selection foreground color when rendering with Direct2D.
* Set minimum column widths.
* Column headings should be centered.
* Batch up item changed notifications during refresh.
* Reenable clipboard and tooltip support.
* Ellipsize message text on Windows.
* Allow tooltips to at least show when clicking on a row.
* Have emojis display in color in tooltips as well.
* Add warning message when not building wxWidgets ourselves.
* Only notify of adding new connection if visible.
* Add additional checks around ItemAdded/ItemDeleted.
* Add/remove items individually every refresh to avoid intermittent Linux crashes.
* Defer memory deallocation until after the UI actually removes the item from view.
* Another attempt at fixing the crash.
* Need to validate pointer in GetAttr() too.
* Undo previous changes, try locking allReporterData_ instead.
* Fix Hamlib compiler failure due to changes in 4.6.
* Add some missed locks.
* Replace locks with assertions that we're on the main thread.
* Probably don't need to defer delete.
* Add additional checking around query methods.
* Warning cleanup.
* Make sure items are unselected before deleting.
* Unselect selected user once QSY request has been sent.
* Force a complete redraw when rows go away.
* Try another way of forcing a redraw.
* We only need to call Cleared() once per row.
* Improve CPU usage.
* Aggressively clear selected users when not actively interacting with them.
* Sort by connect time if no columns are selected for sorting.
* PAIn thread should be RT too (missed from previous PR).
* Force a re-sort on connect.
* Resort on every update from the server.
* Sort fix that hopefully doesn't crash.
* Some minor additional changes.
* Disable PAIn thread and just rely on the normal PulseAudio callbacks for passing audio in and out.
* Make re-sorting happen only once a second.
* Try putting mutexes around all data access.
* Increase RLIMIT_RTTIME just in case.
* Add missed locks.
* RT limit was set too high.
* Sleep for less time than latency if possible.
* Fix crash in TapStep while stopping.
* Defer FreeDV Reporter initialization on startup to make sure GUI is up beforehand.
* Try again to get RTKit to work in the GH environment.
* Try capping maxlength to avoid RX FIFO full errors.
* Increase max buffer size due to overrun errors.
* Default maxlength for output devices to prevent overruns.
* Try bringing back PAin thread. WIP
* Add missed PulseAudio locking.
* Remove more missed mallocs from RT path.
* Remove unneeded code from PulseAudio logic.
* Tab cleanup.
* Increase RX FIFO size.
* Force sleep if we take much longer than expected (avoids kill by rtkit).
* Forgot to check in file.
* Fix compiler error for real.
* Fix issue causing occasionally being unable to get the mic input on macOS.
* Add timeout for EOO in case we lose input devices.
* Add additional protection to prevent the OS from killing the app.
* macOS: get device names an alternate way to avoid empty names for certain devices.
* Make RX plots more reliable when pipewire sends multiple seconds of RX audio at once.
* Oops, used wrong comparison for EOO timeout.
* Catch SIGXCPU in case we exceed RT limits.
* Prevent RX processing from occurring if there's no space in the output FIFO.
* Apparently I was setting the soft and hard rlimits to be the same.
* Lower soft RT limit to 10ms.
* Disable RT scheduling on Linux.
* Added additional protection against server-caused race condition.
* Defer add to table until after we get connection successful message.
* Add missed lock on connect successful message.
* Add missed visibility check.
* Fix issue on Linux.
* Add additional logging to help diagnose duplicate callsign issue.
* Downgrade additional logging to debug/verbose only.
* Fix bug where incorrect SNR was being reported for RADE signals after EOO received.
* Back out previous change buffering rows until fully connected as it doesn't work properly on Windows.
* Fix right-click menu regression.
* Fix issue preventing column sorting on Windows.
* Upgrade wxWidgets to 3.2.8.
* Right-click should not select items.
* Bound requested frames between both min and max.
* Request real-time scheduling for pipeline threads (currently macOS-only).
* Ensure pipeline threads never execute at the same time as audio threads.
* macOS: ask for 5ms audio blocks.
* Temporarily re-enable stress test.
* Suppress codespell for .mm files.
* Swap times/buffer sizes.
* Fix compiler error.
* Go back to previous frame size and time.
* Re-disable stress test.
* Try adjusting time quantum.
* Move audio priority handling to audio layer.
* Set helper thread to real-time on Windows.
* Temporarily disable audio workgroup code.
* Disable periodicity in macOS RT request
* Forgot to define variable.
* Only allow a max of 50% CPU time for audio helpers.
* More tuning of the duty cycle.
* Make sure we sleep some amount of time every time through the loop.
* Go back to smaller time quantum.
* Another quantum tweak.
* Increase audio sample block size to 20ms.
* Use 2.5ms for audio block time.
* Try increasing FIFO sizes for RX.
* Oops, forgot to comment out the / 1000 too.
* Get faster turnaround for macOS GH builds.
* Revert FIFO changes.
* Add additional debugging.
* Fix compiler error.
* Fix typo.
* Use semaphores to notify TX/RX worker threads.
* Try dividing duty cycle by 2 to avoid starvation.
* Reenable audio workgroups.
* No point in having parallel threads if only one mode is active.
* Ensure that ParallelStep gets added to the audio workgroup too.
* Anything for GUI consumption shouldn't be in the RT path.
* Go back to 10ms audio blocks to see if reporting tests can more reliably pass.
* Fix issue causing audio to not work on certain Macs.
* Support real-time threading on Linux when using PulseAudio/pipewire.
* Fix misspelling.
* dbus needs to be installed in the environment.
* macOS: try a smaller block size again.
* Revert "macOS: try a smaller block size again."
This reverts commit 1d21ad6934.
* Try architecture-specific audio block times.
* rtkit itself also needs to be installed in the environment.
* Revert ordering changes to macOS CI builds.
* GH user needs to be added to rtkit group.
* Implement semaphores on Windows.
* Don't exit the render/capture threads unless requested by higher level code.
* Don't take address of a pointer.
* macOS: set thread QoS for RT threads.
* Move RADE RX/TX memory allocations out of real-time path.
* Moving additional memory allocations outside of RT context.
* Move all remaining allocations in pipeline outside of RT except for legacy FreeDV modes.
* Longer audio block times don't seem to be making a difference.
* Move legacy mode memory allocations out of RT context.
* Prevent unbounded mutex locks inside pipeline.
* Windows: fallback to a simple sleep if we can't lock the semaphore.
* Increase maximum wait time to match what Windows returns for buffer size.
* PulseAudio: fallback to simple sleep if we can't wait on semaphore.
* Prevent unbounded locking in TapStep.
* Initial implementation of macOS audio handling using native APIs.
* Fix minor calculation issue in Easy Setup when generating sine waves.
* Fix spelling error.
* Fix issue causing distorted output audio.
* Refactor audio device info retrieval to a separate method.
* Get channel information for discovered macOS audio devices.
* Add support for retrieving default audio devices.
* Try using AVAudioSinkNode for getting mic/RX audio.
* Use AVAudioSourceNode for output side too.
* Fix compiler error.
* Set minimum macOS version to 10.15 due to PyTorch and new audio code requiring it.
* Disable universal build in GH action to make it run a bit faster.
* Additionally use Homebrew packages as deps instead of building own for GH action.
* Don't build hamlib if BUILD_DEPS is 0.
* Use matrix to test macOS build on both x86_64 and ARM64.
* Experiment: always sleep for 10ms every time through the TX/RX loop.
* Temporarily enable longer tests to make sure previous pass wasn't a fluke.
* Remove logic not available in 10.15.
* Revert "Experiment: always sleep for 10ms every time through the TX/RX loop."
This reverts commit 6019c9a522.
* Need to query valid sample rates from Core Audio.
* Validate provided sample rate and channel data when creating audio object.
* Set device sample rate to user-configured option.
* Sample rates could be different between input and output.
* Revert "Temporarily enable longer tests to make sure previous pass wasn't a fluke."
This reverts commit f90c4cd081.
* Remove now unused code.
* Harmonize native vs. PortAudio config across build scripts.
* Zero out audio by default to prevent buzzing.
* Update README to use new environment variable for audio driver control.
* Add COM wrapper class around ThreadedObject. Prerequisite for WASAPI audio support.
* Disable OLE 1.0 per Windows documentation.
* WIP WASPI device enumeration.
* Forgot to check in some changes.
* Fix compiler errors.
* Avoid deadlock during sample rate list retrieval.
* s/WASPI/WASAPI/g
* WIP WASAPIAudioDevice design.
* Full WASAPIAudioDevice implementation.
* Increase priority of audio tasks to prevent dropouts.
* Add missed nullptr initialization.
* Increase buffer time to 40ms
* Fix issue where non-English device names were getting corrupted.
* Add additional logging.
* Need to create macOS audio device name just like with Windows.
* Update README to document new config options.
* Embed sioclient source code into FreeDV codebase.
* Exclude 3rdparty from spellcheck.
* Revert "Embed sioclient source code into FreeDV codebase."
This reverts commit 2021118544.
* Embed ZIP file with sioclient code instead of individual files.
* Generate tarball on Linux as CMake behaves weird with a macOS-created ZIP file.
* Add json and websocketpp source files.
* Add ThreadedTimer (prerequisite for own socket.io implementation).
* Adapt ezDV implementation of sioclient.
* Update FreeDVReporter to use our SocketIoClient.
* Fix bugs preventing connection to FreeDV Reporter server.
* Read all data available every 50ms.
* Fix bug preventing hide/show from working properly.
* Remove sioclient tarball and build logic.
* Fix Linux build issue.
* Work around exception bug on object destruction.
* Fix Windows build failure.
* Ensure that connection is actually closed if we stop being in contact with the server.
* Fix additional segfault on exit.
* Sockets on Windows aren't actually fully POSIX.
* Use ulog for websocketpp logging.
* Update comments in recently added files.
* Additional logging cleanup.
* Show IP address that failed connection.
* Reduce websocketpp buffer size to match what's used in our read logic.
* Fix spelling error.
* Fix issue with corrupted IP addresses being logged.
* Add additional logic to make absolutely sure we're done with DNS.
* Make sure whichIndex is valid once we complete DNS.
* Widen Mode field on bottom of main window.
* Prevent EOO from being repeatedly sent during TX stop.
* Make sure we don't accidentally overflow the output FIFO when transmitting EOO.
* Add additional logging to verify EOO behavior.
* Fix issue preventing EOO from being sent on voice keyer completion.
* Scale EOO the same as the primary waveform.
* Remove #ifdef code per PR comment.
* Use main branch for RADE.
* Add initial support for RADE callsign handling.
* Add additional logging and try different way to calculate amps.
* More debugging. We should also make sure all of EOO is initialized to 'zero'.
* Testing and other changes:
1. Add RadeTextTest to rule out any issues with freedv-gui.
2. Calculate RMS magnitude across all EOO bits, not just the ones used by RADE text.
* Fix test so it actually runs on Linux.
* Disable temporary fprintf of symbols.
* GH environment gave different BER, so let's try positive ROT45.
* set_eoo_bits should be receiving OFDM symbols.
* Don't fault if we don't have a RX callback.
* Fix issue preventing ctest from working in GH environment.
* Reenable interleaving.
* Initial code cleanup.
* Reformat code to make it easier to read.
* Inject noise into the TX sample in the ctest.
* Add initial test scripts for verifying RADE reporting.
* Add some sleeps since the UT framework can issue GUI events faster than the audio can escape FreeDV.
* Add ctest for test_rade_reporting.sh.
* Add RADE reporting noise test.
* Use codec2 ch tool to generate noise.
* Add raw/coded BER debugging output for tests.
* Fix formula used for BER.
* Update ctest to use mpp instead of AWGN, similar to RADE repo.
* Bring back AWGN test.
* Add Octave to build environment.
* Update cmake-macos.yml
* Fix buffer overrun from previous changes.
* Accidentally included previous change.
* Ensure we have defaults for txattempts and txtime.
* Coded BER should use output of LDPC decode.
* Install octave-signal in macOS CI environment.
* Need to pass -forge in to install signal properly.
* Need octave-control too.
* Need gcc to build control/signal.
* Some additional stuff is needed to get gfortran working.
* Need to use gcc 14 instead.
* Add debugging.
* Fix typo.
* Homebrew uses /opt/homebrew.
* Clarify logging as we also report to FreeDV Reporter.
* Forgot to update ctest acceptance criteria due to logging change.
* Add logic to output TX/RX floats to files.
* rade_text should handle stuffing of unused EOO bits, not upstream code.
* Test framework: remove hardcoded 5s delay after startup request.
* Calculate raw BER across entire EOO, not just the used portion.
* Code cleanup and EOO calculation fixes.
* FreeDV Reporter: only highlight for 2 seconds if there's no RX callsign.
* Switch RADE branch to main as EOO changes have now been merged.
* Add logic for automatically running RX/TX tests.
* Add GH action to load loopback audio driver.
* Add ctest configuration file.
* Try loading modules after installing packages.
* Add additional required packages.
* Add automated test script.
* Temporarily disable GH action changes.
* Have RX take 60 seconds as well.
* Ubuntu upgrade broke /bin/sh.
* Fix issue preventing no zero message from printing.
* Print sync status once per second and validate no resyncs.
* Fix spelling error.
* Switch to RADE main.
* Another attempt at enabling aloop in the GH action.
* Use _ instead of - for module name.
* Use pacmd to create virtual sound devices instead.
* Check sync state every 100ms instead of every 1s.
* Fix issue preventing sync changes from making it to the log.
* Add full duplex test.
* Add full duplex tests to CMakeLists.
* Generate sine wave for TX.
* Update GH action to allow new ctests to run.
* Add debugging for UT.
* pipewire-pulse is needed to actually use PipeWire.
* Install suggested packages to get additional required commands.
* Can't use --install-suggests, so explicitly mention packages.
* Enable pipewire service.
* Fix call to helper Python script.
* Show status of pipewire prior to running tests.
* Remove status display from GH action.
* Use module-loopback to simulate TX plugged into RX.
* Try explicitly starting pipewire.
* Use strict 20ms cadence for TX/RX thread instead of relying on the audio subsystem.
* Need to install wireplumber as well.
* Update regular expression in ctest as sync change message isn't printed unless there's a problem.
* Add RADE to ctests.
* RADE ctest pass wasn't actually a pass.
* Need to explicitly check for LINUX for PulseAudio based ctests.
* Reenable PortAudio build.
* Invert test criteria.
* Show pipewire status just before thye test run.
* List sink info prior to each test.
* Make sure virtual devices aren't suspended.
* Need to have the suspension disable code in another file.
* Actually, installing pulseaudio isn't a good idea.
* Need a window manager for things to refresh properly.
* Display wireplumber info after FDV start.
* Add additional sleep to guarantee that FreeDV is listening when wireplumber status is printed.
* Add pw-top for additional info.
* dbus needs to be running for pipewire to work.
* Need to wait a bit for Xvfb to become available.
* Get additional pipewire logs prior to UT run.
* Install/run rtkit too.
* Attempt to get pipewire logs after test execution.
* Looks like we still need to start pipewire manually.
* Further increase pipewire debugging.
* Too much debugging output before.
* Upgrade packages first before installing new ones.
* We weren't actually running 24.04 on GH.
* Adjust packages due to using 24.04.
* Suppress spellcheck issue.
* Disable pipewire debugging.
* Test changes to PulseAudioDevice to improve dropouts.
* Update tests to reflect 1 resync per 120s proposed threshold.
* Disable additional debugging output.
* Allow override of test audio devices from the command line.
* Enable ctests on macOS.
* getSync() isn't thread safe, causing false UT failures.
* Install virtual audio devices in CI environment for macOS.
* Use script to build drivers instead.
* Oops, forgot to remove some code first.
* Adjust permissions of script.
* Make tests work on Linux again with loopback devices.
* Attempt change to make virtual audio devices actually build.
* Try building on x86 macOS.
* Fix quote escape issues.
* More quote escape issues.
* Forgot missed folder change.
* Enable mirrored virtual audio devices.
* Fix incorrect device names in ctest execution.
* Need to append '2ch' to the other device names too.
* Clear sync when transmitting.
* Grant microphone access to FreeDV.
* Need to back out one of the changes due to voice keyer problems.
* Attempt to fix CI failures.
* Take screenshot of macOS CI environment for debugging.
* Let's try running the script separately to get a screenshot.
* Screenshot was in the wrong place.
* Need to grant access to provisioner for microphone.
* Alternate way of granting permissions.
* Oops, remove call to tccutil.
* Per PLT meeting, we shouldn't drop sync at all.
* Shorten test time to 60s.
* More tweaks to make sure that the devices are correct on Linux.
* Disable screen capture.
* Additional fixes to resolve script problems on macOS.
* I don't think previous PulseAudio changes made anything better.
* Improve dropouts when using PulseAudio/pipewire.
* Revert "Improve dropouts when using PulseAudio/pipewire."
This reverts commit f16c27f4d3.
* Have stream write callback get audio data from higher level code.
* Add PowerShell script for Windows testing.
* Move ctest files into a separate location for organization purposes.
* Add parameters for PowerShell test script.
* Add help text for the script.
* Update conf file path due to previous commits.
* Fix syntax errors in script.
* Experiment: Test Windows installer on GH action.
* Try actually running full duplex test inside Windows.
* Add 5 seconds to allow Xvfb to fully come up.
* Update other actions to use latest checkout.
* Need VS command prompt.
* Add additional debugging.
* Maybe we need to throw an exception to make the GH action fail.
* List devices another way.
* Add yet more debugging.
* Fix audio devices.
* Try with 700D to at least make sure we're on the right track.
* Fix GH action.
* Fix syntax errors.
* Try installing VAC instead as FreeDV is crashing with Scream.
* Fix syntax error.
* Try different hardware ID.
* Ignore error code that actually means success.
* Oops, fix syntax error.
* Try earlier version of VAC.
* Fix archive commands.
* Fix path.
* Probably need to start AudioEndpointBuilder too.
* Stop services first.
* Print out any errors that might be getting thrown.
* Revert "Print out any errors that might be getting thrown."
This reverts commit 15e3908c80.
* Disable test GH action for Windows and investigate failure later.
* Reenable Windows testing with new way of debugging.
* Add logging based on clues from Windows event log.
* Revert "Add logging based on clues from Windows event log."
This reverts commit de4f11362e.
* Use procdump to get a crash dump.
* Full dump seems to take a while.
* Run procdump in the background.
* Log audio errors.
* Print device name in error message.
* Make sure message box doesn't display.
* Enable additional PortAudio debugging.
* Add yet more logging.
* Add additional Registry key to ensure FreeDV has microphone permissions.
* Need to add Registry key.
* Add more keys to try to force microphone permission.
* Clean up debugging, add tests for RADE/700E/1600.
* More cleanup.
* Include very basic Python install in the FreeDV Windows installer.
* Bump version to 2.0.0 devel.
* Adjust pth file to allow pip to work.
* Create post-install script. Not tied into the installer yet.
* Install needed packages during setup.
* Fix issue preventing uninstall commands from being properly added to the installer.
* Add RADE binaries to installer for initial tests.
* Suppress RADAE/RADE in code spelling check.
* Fix additional compiler errors.
* Fix GitHub actions on Linux and macOS.
* Print API version later on startup so it'll actually show up in the Windows console.
* Enforce minimum 16 kHz sample rate for voice keyer files.
* Disallow sample rates lower than 16000 in audio config.
* Initial support for RX using librade.
* Adjust dependencies in CMake infrastructure to make sure RADE/Opus build before FreeDV code.
* Add GUI for RADE.
* Add forgotten files.
* Add initialize/finalize calls for RADE.
* Add progress text to post-install script.
* Adjust paths based on recent RADE commits.
* Add basic support for RADE TX.
* Output EOO on end of TX.
* Fix error preventing Windows installer from being created without 2020 mode.
* Change working directory to folder containing freedv.exe.
* Remove quotes from messages in post-setup script.
* Fix compiler errors.
* Fix assertion failure in PlaybackStep.
* Fix assertion failure in FreeDVInterface.
* Suppress warnings during post-install.
* Get missed bugfixes from previous PR.
* Test changes to get GH actions working again for Windows.
* Forgot to install WINE in GH action.
* Try win64 instead of win32 for WINE.
* wine32 and not wine64
* Try different way of installing wine.
* Remove packages that don't exist.
* Oops, actually installed Python 3.12.7.
* Use normal Python installer instead of wenv.
* Need to make sure wine runs in console-only.
* Use Wine from Ubuntu.
* Install wine32 too.
* Ignore return value.
* Use xvfb since trying to bypass doesn't work for winecfg.
* s/xvfb/xvfb-run/
* Separate WINE install from Python.
* More tweaks to the GH workflow.
* wine64 instead of wine
* Use WINEARCH when creating env.
* test changes
* Start xvfb separately.
* xvfb -> Xvfb
* Add debugging
* Forgot :
* Make WINE think it's Windows 10.
* Back to using wine-staging.
* Reenable xvfb.
* Fix cmake invocation.
* Integrate BuildOpus changes from RADE.
* Ensure Opus is built for minimum supported FreeDV version.
* Warning cleanup.
* Ensure RADE is actually built as a universal binary.
* Include basic Python into the .app file.
* macOS version now builds. Still need to figure out how to get modules installed.
* Resolve problems preventing proper execution of the macOS binary.
* torchvision isn't necessary.
* Fix reporting so that we report RADE instead of 'unk'.
* Expire devel versions of FreeDV after six months.
* Attenuate RADE TX by 4dB to avoid needing to adjust TX Attenuation dial just for that mode.
* Disable 2020/2020B on all platforms.
* Stop building 32 bit Windows builds.
* Move RADE mode to the top of the mode list.
* Tweak preprocess function for RADE.
* Update Windows Python to 3.12.7.
* Re-disable ARM Windows builds while we figure out how to build it.
* Build Wine in signing environment.
* Update wxWidgets to 3.2.6.
* Workaround time display issue in FreeDV Reporter on macOS.
* Limit number of numpy/OpenBLAS threads due to known issues with their interaction with Python/C++ threading.
* Add additional guarding around Codec2 FIFO calls.
* Split off DMG creation into separate target to save time during development.
* Use dr-reset for initial release.
Mooneer Salem