Instead of including the files directly in plugins/MidiImport/portsmf,
that path is now a git submodule (https://github.com/portsmf/portsmf).
Several other changes have been made to clean up MidiImport as well.
* Renamed lmms_basics.h to LmmsTypes.h and scoped it down for that purpose.
* Shifted the LMMS_STRINGIFY macro to it's own header.
* Removed the debug.h header file and use cmake to handle the macro logic.
* Remove some unused headers and include directives
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>
* 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
Summary:
* `NULL` -> `nullptr`
* `gui` -> Function `getGUI()`
* `pluginFactory` -> Function `getPluginFactory()`
* `assert` (redefinition) -> using `NDEBUG` instead, which standard `assert` respects.
* `powf` (C stdlib symbol clash) -> removed and all expansions replaced with calls to `std::pow`.
* `exp10` (nonstandard function symbol clash) -> removed and all expansions replaced with calls to `std::pow`.
* `PATH_DEV_DSP` -> File-scope QString of identical name and value.
* `VST_SNC_SHM_KEY_FILE` -> constexpr char* with identical name and value.
* `MM_ALLOC` and `MM_FREE` -> Functions with identical name and implementation.
* `INVAL`, `OUTVAL`, etc. for automation nodes -> Functions with identical names and implementations.
* BandLimitedWave.h: All integer constant macros replaced with constexpr ints of same name and value.
* `FAST_RAND_MAX` -> constexpr int of same name and value.
* `QSTR_TO_STDSTR` -> Function with identical name and equivalent implementation.
* `CCONST` -> constexpr function template with identical name and implementation.
* `F_OPEN_UTF8` -> Function with identical name and equivalent implementation.
* `LADSPA_PATH_SEPARATOR` -> constexpr char with identical name and value.
* `UI_CTRL_KEY` -> constexpr char* with identical name and value.
* `ALIGN_SIZE` -> Renamed to `LMMS_ALIGN_SIZE` and converted from a macro to a constexpr size_t.
* `JACK_MIDI_BUFFER_MAX` -> constexpr size_t with identical name and value.
* versioninfo.h: `PLATFORM`, `MACHINE` and `COMPILER_VERSION` -> prefixed with `LMMS_BUILDCONF_` and converted from macros to constexpr char* literals.
* Header guard _OSCILLOSCOPE -> renamed to OSCILLOSCOPE_H
* Header guard _TIME_DISPLAY_WIDGET -> renamed to TIME_DISPLAY_WIDGET_H
* C-style typecasts in DrumSynth.cpp have been replaced with `static_cast`.
* constexpr numerical constants are initialized with assignment notation instead of curly brace intializers.
* In portsmf, `Alg_seq::operator[]` will throw an exception instead of returning null if the operator index is out of range.
Additionally, in many places, global constants that were declared as `const T foo = bar;` were changed from const to constexpr, leaving them const and making them potentially evaluable at compile time.
Some macros that only appeared in single source files and were unused in those files have been removed entirely.
* Fixes createTCO method on some classes
Classes that inherit from "Track", also inherit the createTCO method. That method takes a MidiTime position as an argument, but except on the class SampleTrack that argument was ignored. That lead to unnecessary calls to TCO->movePosition after creating a TCO in many parts of the codebase (making the argument completely redundant) and even to a bug on the BBEditor, caused by a call to createTCO that expected the position to be set on the constructor (see issue #5673).
That PR adds code to move the TCO to the appropriate position inside the constructor of the classes that didn't have it, fixes the code style on the SampleTrack createTCO method and removes the now unneeded calls to movePosition from source files on src/ and plugins/. On Track::loadSettings there was a call to saveJournallingState(false) followed immediately by restoreJournallingState() which was deleted because it's redundant (probably a left over from copying the code from pasteSelection?).
* Fix code style issues
Fixes code style issues on some files (except for ones where the current statements already had a different code style. In those the used code style was kept for consistency).
* Fixes more code style issues
* Fixes code style issues on the parameter name
Fixes code style issue on the parameter name of createTCO, where _pos was supposed to be just pos. The existing code had the old style and I ended up replicating it on the other methods.
* Code style fixes
Fixes code style in the changed lines.
* Fixes bug with dragging to negative positions
There was a bug (already present before this PR) where dragging
a selection before MidiTime 0 would result in some TCOs being placed on
negative positions. This PR fixes this bug by applying the following
changes:
1) TrackContentObject::movePosition now moves the TCO to
positions equal or above 0 only.
2) Because of the previous change, I removed the line that
calculated the max value between 0 and the new position on
TrackContentObjectView::mouseMoveEvent when dragging a single TCO (and
added a line updating the value to the real new position of the TCO so
the label displays the position correctly).
3) Unrelated to this bug, but removed an unnecessary call to
TrackContentWidget::changePosition on the left resize of sample TCOs
because it will already be called when movePosition triggers the
positionChanged signal.
4) Added some logic to the TrackContentWidget::pasteSelection
method to find the left most TCO being pasted and make sure that the
offset is corrected so it doesn't end up on a negative position (similar
to the logic for the MoveSelection action).
5) Removed another line that calculated the max between 0 and
the new position on Track::removeBar since it's now safe to call
movePosition with negative values.
* Uses std::max instead of a conditional statement
As suggested by Spekular, we use std::max instead of a
conditional statement to correct the value of offset if it positions a
TCO on a negative position.
This should keep LMMS responsive from the window managers POV, unless
the user selects a huge enough default sound font that loading it
takes too long.
Name the automation tracks like "[MIDI trackname] CC ##". And don't
call the automation track creation function for every CC event, as
it now will involve constructing a QString.
Too short notes had their duration rounded down to 0 on import,
as MIDI precision allows way shorter notes than LMMS, and
portsmf reports the duration as a double, where a beat == 1.0.
Was going to use ceil() first, but that might round some notes
up to a slightly longer duration.
Should be enough to add the object to the automation pattern only once,
as the objModel variable that is added will always be the same for the
same value of ccid, which indexes the ccs array.
This will speed up creation of automation tracks for MIDI CCs and pitch
bend.
This commit adds checks for conditions that are asserted during calls to
get_atom_value.
It might fix a crash that is described in 1981. Unfortunately no files
have been attached to that issue. However, I was able to crash LMMS
using a local file and this crash is gone with this fix. So hopefully this
change also fixes the crashes described in 1981.
