This reworks the auto-quit feature by introducing a new AudioBuffer class which keeps track of which channels are currently silent as audio flows through the effects chain.
When track channels going into an effect's input are not marked as quiet, it is assumed a signal is present and the plugin needs to wake up if it is asleep due to auto-quit. After a plugin processes a buffer, the silence status is updated.
When the auto-quit setting is disabled (that is, when effects are always kept running), effects are always assumed to have input noise (a non-quiet signal present at the plugin inputs), which should result in the same behavior as before.
Benefits:
- The auto-quit system now closely follows how it is supposed to function by only waking plugins which have non-zero input rather than waking all plugins at once whenever an instrument plays a note or a sample track plays. This granularity better fits multi-channel plugins and pin connector routing where not all plugin inputs are connected to the same track channels. This means a sleeping plugin whose inputs are connected to channels 3/4 would not need to wake up if a signal is only present on channels 1/2.
- Silencing channels that are already known to be silent is a no-op
- Calculating the absolute peak sample value for a channel already known to be silent is a no-op
- The silence flags also could be useful for other purposes, such as adding visual indicators to represent how audio signals flow in and out of each plugin
- With a little more work, auto-quit could be enabled/disabled for plugins on an individual basis
- With a little more work, auto-quit could be implemented for instrument plugins
- AudioBuffer can be used with SharedMemory
- AudioBuffer could be used in plugins for their buffers
This new system works so long as the silence flags for each channel remain valid at each point along the effect chain. Modifying the buffers without an accompanying update of the silence flags could violate assumptions. Through unit tests, the correct functioning of AudioBuffer itself can be validated, but its usage in AudioBusHandle, Mixer, and a few other places where track channels are handled will need to be done with care.
---------
Co-authored-by: Sotonye Atemie <sakertooth@gmail.com>
Closes#7869
This PR aims to fix linking bugs and it aims to make linking code faster. In the future I would like to replace controller and automation code with linked models, so it is essential for this feature to work as efficiently as possible.
Before this PR:
- AutomatableModels store a list of AutomatableModels that they are linked to.
- setValue() and other functions make recursive calls to themself resulting in the #7869 crash.
- Each AutomatableModel can unlink from other AutomatableModels, unlinking is the inverse operation to linking.
After this PR:
- AutomatableModels store a pointer to an other AutomatableModel making a linked list. The end is connected to the first element resulting in a "ring".
- setValue() and others are now recursion free, the code runs for every linked model, more efficiently than before.
- Each AutomatableModel can NOT unlink form other AutomatableModels, unlinking is NOT the inverse operation to linking. AutomatableModels can unlink themself from the linked list, they can not unlink themself from single models.
---------
Co-authored-by: allejok96 <allejok96@gmail.com>
Previously, this PR simply added a new signal to the Timeline class and had it emitted by the TimeLineWidget class in order to solve #7351.
However, as messmerd pointed out in his review comments, that was quite a hacky solution, and ideally the positionChanged signal would be solely managed by the backend Timeline class, while the frontend TimeLineWidget class would connect to those signals, instead of the other way around.
This PR is no longer a simple bugfix, but instead a refactoring of the Timeline/TimeLineWidget signal/slots.
Changes
- The positionChanged signal and updatePosition slot were removed from TimeLineWidget and moved to Timeline.
- Removed PlayPos, and instead store the timeline position in a TimePos along with a separate frame offset counter variable. The functions to set the ticks/timepos in Timeline emit the positionChanged signal. (Also, may emit positionJumped signal if the ticks were forcefully set--see below)
- The pos() method and PlayPos m_pos were removed from TimeLineWidget;
- The constructor for TimeLineWidget no longer requires a PlayPos reference.
- Since each TimeLineWidget stores a reference to its Timeline, a new method was added, timeline(), for other classes to access it.
- Removed array of PlayPos'es in Song. Now each Timeline holds their respective TimePos.
- Song's methods for getPlayPos were changed to return the TimePos of the respective Timeline. The non-const versions of the methods were removed because Timeline does not expose its TimePos to write.
- All of the places where Timelines are used were updated, along with their calls to access/modify the PlayPos. For example, occurrences of m_timeline->pos() were replaced with m_timeline->timeline()->pos(), and calls to m_timeline->pos().setTicks(ticks) were changed to m_timeline->timeline()->setTicks(ticks).
- ALSO: Removed m_elapsedMilliseconds, m_elapsedBars, and m_elapsedTicks from Song. The elapsed milliseconds is now handled individually by each Timeline.
- NEW: The m_jumped variable has been removed from Timeline. Now jumped events emit Timeline::positionJumped automatically whenever the ticks are forcefully set. This means it is no longer necessary to call timeline->setJumped(true) or timeline->setFrameOffset(0) whenever the playhead position is forcefully set. Many places in the codebase were already missing these calls, so this may fix some unknown bugs.
---------
Co-authored-by: Dalton Messmer <messmer.dalton@gmail.com>
Co-authored-by: saker <sakertooth@gmail.com>
Co-authored-by: Alex <allejok96@gmail.com>
Follow-Up of 7db3fa94a1 .
This was done by setting `CMAKE_C_INCLUDE_WHAT_YOU_USE` and
`CMAKE_CXX_INCLUDE_WHAT_YOU_USE` to (broken down into multiple lines here,
note, all below `FL/x.h` is not required for C):
```
include-what-you-use;
-Xiwyu;--mapping_file=/usr/share/include-what-you-use/qt5_11.imp;
-Xiwyu;--keep=*/xmmintrin.h;
-Xiwyu;--keep=*/lmmsconfig.h;
-Xiwyu;--keep=*/weak_libjack.h;
-Xiwyu;--keep=*/sys/*;
-Xiwyu;--keep=*/debug.h;
-Xiwyu;--keep=*/SDL/*;
-Xiwyu;--keep=*/alsa/*;
-Xiwyu;--keep=*/FL/x.h;
-Xiwyu;--keep=*/MidiApple.h;
-Xiwyu;--keep=*/MidiWinMM.h;
-Xiwyu;--keep=*/AudioSoundIo.h;
-Xiwyu;--keep=*/OpulenZ/adplug/*;
-Xiwyu;--keep=QPainterPath;
-Xiwyu;--keep=QtTest
```
FAQ:
* Q: Does this speed-up a completely fresh compile?
* A: No, I measured it.
* Q: Does it speed up anything else?
* A: Yes. If you change one header, it can reduce the number of CPP files
that your compiler needs to recompile, or your IDE has to re-scan.
* Q: What other reasons are for this PR?
* A: It's idiomatic to only include headers if you need them. Also, it will
reduce output for those who want to use IWYU for new PRs.
Background:
This is just a remainder PR of what I planned. My original idea was to setup
a CI which warns you of useless includes (but not of all issues that IWYU
complains about). However, I could not see that this was favored on Discord.
A full IWYU CI also has the problem that it (possibly??) needs to compile
with `make -j 1`, which would make CI really slow.
However, for that plan, I had to fix the whole code base to be IWYU
compliant - which it now is.
Reworks how note detuning copying works so as not to perform a clip duplication and allocation by default in the constructors
---------
Co-authored-by: Dalton Messmer <messmer.dalton@gmail.com>
* Compile in C++20 mode
* Fix implicit lambda captures of `this` by `[=]`
Those implicit captures were deprecated in C++20
* Silence MSVC atomic std::shared_ptr warning
Unfortunately std::atomic<std::shared_ptr> (P0718R2) is not supported by
GCC until GCC 12 and still is not supported by Clang or Apple Clang, so
it looks like we will continue using std::atomic_load for the time being
* Use C++20 in RemoteVstPlugin
* Simplification
* Add comment
* Fix bitwise operations between different enumeration types
* Revert "Fix bitwise operations between different enumeration types"
This reverts commit d45792cd72.
* Use a helper function to combine keys and modifiers
* Remove AnalyzeTemporaryDtors from .clang-tidy
AnalyzeTemporaryDtors was deprecated in clang-tidy 16 and fully removed
in clang-tidy 18
* Use C++20 in .clang-format
* Use bitwise OR
Prevents issues if any enum flags in `args` have bits in common
Caused by not calling `Engine::destroy`, as well as creating tracks/clips on the heap and not freeing the memory at the end of the test.
Instead of creating tracks/clips on the heap, they are now made on the stack, similar to how the surrounding tests create its objects.
This improves the way digits are calculated for display in
`Lv2ViewProc`:
- More than 2 digits are now recognized
- Minus signs are now recognized
- Tests have been added
* Add `ArrayVector` class template and tests
* Fix counting of failed test suites
* Support detuning and panning with Sf2 Player
* Restrict panning to supported FluidSynth versions
* Fix data array cast type
* Fix tests for Qt<5.10 and correct mistaken test
* DIsplay warning for FluidSynth < 2
* Remove unnecessary clamp
* Update include guard name
PR #6438 does 2 things:
1. Add check-namespace
2. Fix verify script
This PR contains only part 2 (and does some preparations for part 1). The goal of the PR is to make CI succeed on master.
This PR places all LMMS symbols into namespaces to eliminate any potential future name collisions between LMMS and third-party modules.
Also, this PR changes back `LmmsCore` to `Engine`, reverting c519921306 .
Co-authored-by: allejok96 <allejok96@gmail.com>
This adds a script `check-strings` that checks whether strings, such as file paths or class names, are valid in files outside of the source code, e.g. in translations or themes. This also adds a verify script to verify `check-strings` on a constant git commit. Both scripts are under CI.
* Update ringbuffer submodule to fix includes
* Remove cyclic includes
* Remove Qt include prefixes
* Include C++ versions of C headers
E.g.: assert.h -> cassert
* Move CLIP_BORDER_WIDTH into ClipView
This allows to remove includes to TrackView.h in ClipView cpp files.
* Elliminate useless includes
This improves the include structure by elliminating includes that are
not used. Most of this was done by using `include-what-you-use` with
`CMAKE_C_INCLUDE_WHAT_YOU_USE` and `CMAKE_CXX_INCLUDE_WHAT_YOU_USE`
set to (broken down here):
```
include-what-you-use;
-Xiwyu;--mapping_file=/usr/share/include-what-you-use/qt5_11.imp;
-Xiwyu;--keep=*/xmmintrin.h;
-Xiwyu;--keep=*/lmmsconfig.h;
-Xiwyu;--keep=*/weak_libjack.h;
-Xiwyu;--keep=*/sys/*;
-Xiwyu;--keep=*/debug.h;
-Xiwyu;--keep=*/SDL/*;
-Xiwyu;--keep=*/alsa/*;
-Xiwyu;--keep=*/FL/x.h;
-Xiwyu;--keep=*/MidiApple.h;
-Xiwyu;--keep=*/MidiWinMM.h;
-Xiwyu;--keep=*/AudioSoundIo.h
```
* Fixup: Remove empty #if-#ifdef pairs
* Remove LMMS_HAVE_STD(LIB|INT)_H
This replaces `set(CMAKE_CXX_STANDARD 14)` by `set(CMAKE_CXX_STANDARD 17)`
wherever it is required.
Additionally:
* raise `CMAKE_MINIMUM_REQUIRED(VERSION ...)` to `3.8` (the minimum
that supports C++17)
* `stdshims.h` is now unused and thus removed
* Fix ProjectVersion handling of pre-releases
* Add workaround for old, non-standard version
* Attempt to fix versioning
* More consistent comments
* Apply suggestions from code review
- Set CompareType's underlying type to int and revert change to ProjectVersion::compare's parameters
- Add "None" and "All" as names elements of CompareType enum
- Preserve hyphens in prerelease identifiers
- Pad invalid (too short) versions to prevent crashes or nasty behavior
- Compare numeric identifiers to non-numeric ones correctly
- Don't interpret identifiers of form "-#" as numeric (where '#' is any number of digits)
- Add tests to ensure fixes in this commit work and won't regress in the future
* CMAKE fixes from code review
Co-authored-by: Tres Finocchiaro <tres.finocchiaro@gmail.com>
* Remove unnecessary changes to CMake logic
* More const, more reference
* Apply suggestions from code review
Co-authored-by: Tres Finocchiaro <tres.finocchiaro@gmail.com>
* Create PathUtils
* Replace old SampleBuffer calls
* Fix automatic track names
* Fix things
* Remove accidental duplicate file
* Add includes
* More incldues
* PhysSong's code review + style
* Fix vestige loading?
Seems more reasonable to convert from relative on load and to relative on save than vice versa.
* Typo fix
* More Bases
* Enable more bases
* Add missing semicolons in prefixes
* Nicer sample track tooltip
* Use correct directories
"userXDir" gives the default dir for ladspa, sf2, and gig. "xDir" gives the user dir.
* Make relative to both default and custom locations
Part 1
* Make relative to both default and custom locations
Part 2
* Typofix
* Typofix
* Fix upgrade function after base renaming
* Fix Tests
* Update tests/src/core/RelativePathsTest.cpp
Co-Authored-By: Hyunjin Song <tteu.ingog@gmail.com>
* Choose UserXBase over DefaultXBase if identical
toShortestRelative sticks with the first base found if two bases give the same path length. By placing UserXBase Bases before DefaultXBase Bases in the relativeBases vector, toShortestRelative will prioritize them.
* Ensure baseLocation always has trailing slash
Otherwise, a user configuring a path without one will break things.
* Move loc declaration out of switch
* Semicolon
* Apply suggestions from code review...
* Include PathUtil and sort includes
* More granular includes
* Apply suggestions from code review
Co-Authored-By: Hyunjin Song <tteu.ingog@gmail.com>
* Update include/PathUtil.h
* Leave empty paths alone
* Fix stupid merge
* Really fix merge. Hopefully
* Switch Base from enum to class enum
* Don't pass Base by reference
* Use QStringLiteral for static QString allocation in basePrefix method
* Make VST loading more similar to previous implementation
* Fix tests after enum change
* Attempt to fix VST loading, nicer name for sample clips
* Fix last review comment
Don't append a "/" that will be removed by cleanPath later
* Apply suggestions from code review
Co-authored-by: Dominic Clark <mrdomclark@gmail.com>
Co-authored-by: Hyunjin Song <tteu.ingog@gmail.com>
Co-authored-by: Dominic Clark <mrdomclark@gmail.com>
Fixes regression from 75077f6200 that caused
global automation tracks to have priority in processing.
Adds a test checking for the desired behaviour.
Fixes#4268
