mirror of
https://github.com/LMMS/lmms.git
synced 2026-01-17 10:58:09 -05:00
master
3 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
Tres Finocchiaro
|
51529cefb1 |
Add Qt6 Support (#7339)
* Rebase against master Co-authored-by: michaelgregorius <michael.gregorius.git@arcor.de> Co-authored-by: Rossmaxx <74815851+Rossmaxx@users.noreply.github.com> * Fix Qt6 DMG on Apple (#7240) - Fix linking issues with Qt Framework files - Fix qmake detection * Fixes after rebase * Fix embed.cpp compilation Fix implicit conversion from int when using QString.arg(...) * Fix Qt6 signature change for nativeEventFilter (#7254) * Adds win32EventFilter a wrapper for nativeEventFilter on Windows * win32EventFilter is currently used to intercept top-level Window events (currently, to avoid VSTs setting transparency of the parent application) * fix broken signal slot connections (#7274) QComboBox activated() replaced with textActivated() since Qt 5.14 * Enabled VSTs on Qt 6 (#7273) * enabled VST support for Qt 6 builds * Note : Embedding on QT6 will be buggy on linux as a result of using qt embedding, which unfortunately is a qt bug which hasn't been resolved. * Changed bar lines to follow snap size (#7034) * Added lines in between bars * Changed bar lines to follow snap size * Changed default zoom and quantization value * Added constants for line widths * Added QSS configuration for new grid line colors * Tied line widths to QSS properties * Changed default quantization to 1/4 * Removed clear() from destructor model * Removed destructor in ComboBoxModel.h * Changed member set/get functions to pass by value * Updated signal connection with newer syntax * Fix compilation * Fix MSVC builds * fix nullptr deref in AudioFileProcessor (qt6 branch) (#7532) * ensured mouse event != nullptr before deref * separation of concerns: AFP WaveView updateCursor extract check to pointerCloseToStartEndOrLoop() * marked some function parameters as const * Remove Core5Compat usage * Fix bad merge * Fixes after rebase * Simplify QTX_WRAP_CPP call * Remove comments that are obvious to a developer * Whitespace * Try using Qt 6 for MSVC CI I chose Qt 6.5 because it's the last Qt LTS release with declared support for Visual Studio 2019. Once we upgrade to Visual Studio 2022, we could upgrade Qt as well. * Fix MSVC build Also fixes two memory leaks in MidiWinMM * Fix GuiApplication on MSVC * Fix interpolateInRgb * Try building with patched Calf * Fix submodule * Fix OpulenZ build * Try to fix zyn * Fix comment * Ty to fix zyn (again) * Ty to fix RemotePluginBase * Revert "Ty to fix RemotePluginBase" This reverts commit 92dac44ffb11e19d1d5a21d9155369f017bd59e9. * Update plugins/ZynAddSubFx/CMakeLists.txt Co-authored-by: Dalton Messmer <messmer.dalton@gmail.com> * Fix vertical & horizontal scroll wheel in SongEditor * AppImage: Fix finding of Qt6 libs * Fix implicit QString --> QFileInfo conversion * Point submodule to lmms * Fix multiple deprecation warnings * Fix for Clang compiler * Build with latest Qt LTS version now that we use MSVC 2022 * Update jurplel/install-qt-action to v4.3.0 * Bump minimum Qt6 version for MSVC * Fix incorrect Qt version checks Some comparisons were using ">" rather than ">=" * `QSize()` != `QSize(0, 0)` * Fix more deprecation warnings * Fix style * Simplify Spectrum Analyzer mouse events The Qt bug that used to be present appears to have been fixed, so the workaround can be removed * Minor changes * Fix deprecated QCheckBox signal * Fix setContent helper functions * Remove QMultiMap usage from ControlLayout * Remove SIGNAL and SLOT macros * Revert TrackView.cpp changes * Remove Q_DISABLE_MOVE usage since it does not seem to be available in Qt6 --------- Co-authored-by: michaelgregorius <michael.gregorius.git@arcor.de> Co-authored-by: Rossmaxx <74815851+Rossmaxx@users.noreply.github.com> Co-authored-by: BoredGuy1 <66702733+BoredGuy1@users.noreply.github.com> Co-authored-by: Hyunjin Song <tteu.ingog@gmail.com> Co-authored-by: Lisa Magdalena Riedler <git@riedler.wien> Co-authored-by: Dalton Messmer <messmer.dalton@gmail.com> |
||
|
Fawn
|
e328a136fc |
Use C++20 std::numbers, std::lerp() (#7696)
* use c++20 concepts and numbers for lmms_constants.h * replace lmms::numbers::sqrt2 with std::numbers::sqrt2 * replace lmms::numbers::e with std::numbers::e Also replace the only use of lmms::numbers::inv_e with a local constexpr instead * remove lmms::numbers::pi_half and lmms::numbers::pi_sqr They were only used in one or two places each * replace lmms::numbers::pi with std::numbers::pi * add #include <numbers> to every file touched so far This is probably not needed for some of these files. I'll remove those later * Remove lmms::numbers Rest in peace lmms::numbers::tau, my beloved * Add missing #include <numbers> * replace stray use of F_EPSILON with approximatelyEqual() * make many constants inline constexpr A lot of the remaining constants in lmms_constants.h are specific to SaProcessor. If they are only used there, shouldn't they be in SaProcessor.h? * ok then, it's allowed to be signed * remove #include "lmms_constants.h" for files that don't need it - And also move F_EPSILON into lmms_math.h - And also add an overload for fast_rand() to specify a higher and lower bound - And a bunch of other nonsense * ok then, it's allowed to be inferred * ok then, it can accept an integral * fix typo * appease msvc * appease msvc again * Replace linearInterpolate with std::lerp() As well as time travel to undo several foolish decisions and squash tiny commits together * Fix msvc constexpr warnings * Fix msvc float to double truncation warning * Apply two suggestions from code review Co-authored-by: Dalton Messmer <messmer.dalton@gmail.com> * Apply suggestions from code review Co-authored-by: Dalton Messmer <messmer.dalton@gmail.com> * fix silly mistake * Remove SlicerT's dependence on lmms_math.h * Allow more type inference on fastRand() and fastPow10f() * Apply suggestions from code review Co-authored-by: Dalton Messmer <messmer.dalton@gmail.com> * Clean up fastRand() a little bit more --------- Co-authored-by: Dalton Messmer <messmer.dalton@gmail.com> |
||
|
Michael Gregorius
|
8e40038a2d |
Scalable envelope graph (#7194)
Make the graph scalable by adjusting the painting code of the envelope so that it does not assume fixed widths and heights anymore. Remove the setting of a fixed size from the envelope graph and only set a minimum size. Make three scaling modes available which can be selected via a context menu in the graph: * "Dynamic": This modes corresponds to the rendering strategy of the previous implementation. Initially 80/182 of the available width is used as the maximum width per segment. This can be interpreted like a "zoomed" version of the absolute mode. If the needed space becomes larger than the full width though then it falls back to relative rendering. * "Absolute": Each of the five segments is assigned 1/5 of the available width. The envelopes will always fit but might appear small depending of the current settings. This is a good mode to compare envelopes though. * "Relative": If there is at least one non-zero segment then the whole width is always used to present the envelope. The default scaling mode is "Dynamic". ## Technical details The new painting code is more or less divided into two parts. The first part calculates `QPointF` instances for the different points. In the second part these points are then used to draw the lines and markers. This makes the actual rendering code much more straight forward, readable and maintainable. The interpolation between the line color of an inactive and an active envelope has also been restructured so that it is much more obvious that we are doing an interpolation in the first place. The colors at both ends of the interpolation are explicit now and can therefore be adjusted much easier. The actual color interpolation is done in the helper function `interpolateInRgb` which is provided by the new class `ColorHelper`. This class will later also be needed when the LFO graph is made scalable. The line is rendered as a polyline instead of single line segments. The drawing of the markers has been abstracted into a lambda (with some outside captures though) so that it can be easily adjusted if necessary. The markers are rendered as circles instead of rectangles because that looks much nicer especially when the widget is rendered at a larger size. The width of the lines and marker outlines is determined using the size of the widget so that it scales with the size. A `lerp` function has been added to `lmms_math.h`. |