Added dependencies needed for rtk LV2 UI building to README info
Github workflow: removed test trial
CMake: added missing install example files to /usr/share/qmidiarp
Cmake install: be consistent with .lv2 when installing in home directory
Github workflow: more trials
Github workflow: missing lupdate dependency added
Update workflow
One more try with github workflows
Remove github workflow
Version pushed to 0.7.2
qmidiarp-0.7.1 release
Prepare for 0.7.1 release
Fixed some configure/Makefile.am options
MidiSeq: try to fix github issues #16 and #4 resizing midiseq data vector
Seq LV2 RTK: Approach original display size of Qt plugins
Arp LV2 RTK: Changed default pattern mechanism again and optimized layout
Arp LV2 RTK: Read patterns from .qmidiarprc file and add selector to GUI
Arp and Seq LV2 RTK: Better default settings (Arp didn't have pattern)
Remove unused midiworker from Qt LV2 UIs and Version->0.7.1
cmake build: fixed LV2 RTK UI global symbol visibility
qmidiarp-0.7.0 release
qmidiarp-0.7.0 release
qmidiarp-0.7.0 release
qmidiarp-0.7.0 release
qmidiarp-0.7.0 release
NEWS file updated
Version pushed to 0.7.0
CMake: added robtk toolkit to dist tarball and small adjustments
Package: Add rtk ui sources as extra files to autotools tarball
LV2 RTK ARP: Fixed crash with jalv.qt5 (sending pattern after receive)
LFO LV2 RTK: implemented offset dial following drawn wave minimum
cmake: check for robtk presence during configure
Cmake: fixed translation generator checks (lrelease, lupdate)
Fixed dependency check logic
restore compatibility with Qt < 5.15
Added RTK UIs to cmake build environment as well
Arp and Seq LV2 RTK: use uint8_t where possible
LV2 Arp RTK: Improved pattern string security
Arp: Fixed influence of random that was not audible anymore (regression)
Arp and Seq LV2: do not update pattern on mouse release
Added RTK build to autoconf/make
LV2 modules: Added new UIs based on the robtk toolkit
Changed module type identification to be consistent with earlier changes
GlobalStore: Reset repetion counter when starting at zero time offset
cmake: fixed typo exposed by last commit
Hi David, I think the last git commit (33b9849 ) should force the right flags for avoiding overlinking. I'm sorry I found out only now when the hotfix release is already out. Best Frank
Package: Try to force --as-needed linker flag in libtool (sf bug #35)
cmake: fixed requirement for MIDI backend when only building plugins
Added generated automake files to cmake make dist tarball, Version now 0.6.8
qmidiarp-0.6.7 release
NEWS file updated
LFO: Fixed waveform display in application after realtime fix
cmake: give the configure file a chance to be added to make-dist
Application: Fixed a small non-rt function call in the jack thread
Package: Added CMake files to Makefile.am, too, and updated .gitignore
Cmake build: added handling of translation generation and installation
Cmake build: added a git-based package generation for dist packages
Some cmake output added to .gitignore
Package: added a cmake build environment (no make dist yet)
Fixed sed option on BSD again (still github issue #1)
LFO: Fixed malloc()s introduced during removal of Qt libs from DSPs
Hi David, I pushed some changes to the repo that fix some of the lv2lint (0.17.71) fails. You are perfectly right that dynamically linking to Qt isn't a good way for LV2 UIs. But since the whole Qmidiarp application was built with it initially (and the plugins derived later), changing this would require either a rewrite with a statically linkable (and leaner) toolkit or without toolkit, which is substantial. It now reduces to this except the LFO plugin which has a realtime problem in addition (TODO):...
Update issue templates
Fixed some (not all) lv2lint fails
Fixed (hopefully) sed warning on BSD (github issue #1)
lv2 plugins are overlinked
Hi David, I'm reopening the issue. Unfortunately this is hard to reproduce outside a packaging system's compilation. When I build on my system and check the .so binaries I don't get unused libraries reported when running ldd -u -r qmidiarp_seq_lv2.so and any of the other .so files produced. I remember from some past packaging work that this doesn't mean it will be the same on a distro build system, but I wonder (or forgot) what I can do in the corresponding Makefile.am to avoid this (i.e. avoid libtool...
Version pushed to 0.6.7
qmidiarp-0.6.6 release
NEWS file updated
Package: configure.ac updated
Translation files updated
Copyright dates updated
LV2: Fixed screen mute not updating immediately
LV2: Swapped order of ui type in desriptors: X11 now comes before Qt5
Seq Loopmarker. Fixed marker sometimes snapping back after mouse click
LV2: Added idle and show interface LV2 extensions (github #14)
Restore compatibility with Qt5 < 5.14
NEWS file updated
Logwidget: fixed arg values
Jack-Session removed due to 'official deprecation' for a long time now
Maintenance: replaced some deprecated Qt5 functions
Hi, Sorry, I'm discovering your request only now due to a failed notification. Handling such long cycles is indeed beyond the initial idea (which is around loops and patterns), and it would be unpractical to handle with the current implementation (data storage, drawing, internal memory handling, etc....). If you don't need a high time resolution, setting the tempo to something very low may go in that sense. But why don't you use a linear sequencer such as qtractor. That would do a good job there,...
German and French translations updated, all the best!
German and French translations updated
German and French translation sources updated
German and French translation sources updated
Revert "NSM: Added client_id returned to NSM in case of ALSA backend"
NSM: Added client_id returned to NSM in case of ALSA backend
Make naming consistent for InOutBoxes show/hide
LV2 includes: row back to keep compatibility with earlier LV2 versions
Fixed typo
French and German translations updated
French and German translations updated
Arp: make random mode not repeat notes and honor octave changes
Code: more cleanups
LFO: fixed old bug where recording with high resolutions stopped advancing
Arp: come up with better initial settings: 8ths random 1 octave up
Start rework module interface to Engine to gain some structure
Renamed the InOutBox class to ModuleWidget to indicate usage