Add suloea-labs

[paulfd]

It's a LV2 wrapper around the Aeolus engine, trimmed down to a single division and with the builtin reverb.

[falkTX]

Hi there, thanks for the pull request! (and the nice little plugin too)

2 things to do here

1. SULOEA_LABS_POST_INSTALL_TARGET_TTLFILES define is not needed, and unused in your mk file. please remove it
2. there is a mismatch new/free usage within the plugin. seems started by suloea.cpp:204, by using C++ new operator and freeing that memory with C-style free instead of delete.

==3651137== Mismatched free() / delete / delete []
==3651137==    at 0x4E05288: free (vg_replace_malloc.c:538)
==3651137==    by 0x4AD869: CarlaBackend::CarlaPluginLV2::~CarlaPluginLV2() (CarlaPluginLV2.cpp:742)
==3651137==    by 0x4ADF43: CarlaBackend::CarlaPluginLV2::~CarlaPluginLV2() (CarlaPluginLV2.cpp:846)
==3651137==    by 0x4CBD6F: std::_Sp_counted_ptr<CarlaBackend::CarlaPluginLV2*, (__gnu_cxx::_Lock_policy)2>::_M_dispose() (shared_ptr_base.h:376)
==3651137==    by 0x41A637: std::_Sp_counted_base<(__gnu_cxx::_Lock_policy)2>::_M_release() (shared_ptr_base.h:154)
==3651137==    by 0x419DE2: std::__shared_count<(__gnu_cxx::_Lock_policy)2>::~__shared_count() (shared_ptr_base.h:684)
==3651137==    by 0x41991D: std::__shared_ptr<CarlaBackend::CarlaPlugin, (__gnu_cxx::_Lock_policy)2>::~__shared_ptr() (shared_ptr_base.h:1123)
==3651137==    by 0x419939: std::shared_ptr<CarlaBackend::CarlaPlugin>::~shared_ptr() (shared_ptr.h:93)
==3651137==    by 0x41BED3: void __gnu_cxx::new_allocator<std::shared_ptr<CarlaBackend::CarlaPlugin> >::destroy<std::shared_ptr<CarlaBackend::CarlaPlugin> >(std::shared_ptr<CarlaBackend::CarlaPlugin>*) (new_allocator.h:140)
==3651137==    by 0x41B3F8: void std::allocator_traits<std::allocator<std::shared_ptr<CarlaBackend::CarlaPlugin> > >::destroy<std::shared_ptr<CarlaBackend::CarlaPlugin> >(std::allocator<std::shared_ptr<CarlaBackend::CarlaPlugin> >&, std::shared_ptr<CarlaBackend::CarlaPlugin>*) (alloc_traits.h:487)
==3651137==    by 0x43849A: std::vector<std::shared_ptr<CarlaBackend::CarlaPlugin>, std::allocator<std::shared_ptr<CarlaBackend::CarlaPlugin> > >::_M_erase(__gnu_cxx::__normal_iterator<std::shared_ptr<CarlaBackend::CarlaPlugin>*, std::vector<std::shared_ptr<CarlaBackend::CarlaPlugin>, std::allocator<std::shared_ptr<CarlaBackend::CarlaPlugin> > > >) (vector.tcc:159)
==3651137==    by 0x438237: std::vector<std::shared_ptr<CarlaBackend::CarlaPlugin>, std::allocator<std::shared_ptr<CarlaBackend::CarlaPlugin> > >::erase(__gnu_cxx::__normal_iterator<std::shared_ptr<CarlaBackend::CarlaPlugin> const*, std::vector<std::shared_ptr<CarlaBackend::CarlaPlugin>, std::allocator<std::shared_ptr<CarlaBackend::CarlaPlugin> > > >) (stl_vector.h:1180)
==3651137==  Address 0x557f510 is 0 bytes inside a block of size 2,712 alloc'd
==3651137==    at 0x4E045E4: operator new(unsigned long) (vg_replace_malloc.c:342)
==3651137==    by 0x7002C66: SuloeaPlugin::instantiate(double, char const*, _LV2_Feature const* const*) (suloea.cpp:204)
==3651137==    by 0x7003810: instantiate(_LV2_Descriptor const*, double, char const*, _LV2_Feature const* const*) (suloea.cpp:606)
==3651137==    by 0x4C4272: CarlaBackend::CarlaPluginLV2::init(std::shared_ptr<CarlaBackend::CarlaPlugin>, char const*, char const*, unsigned int, char const*&) (CarlaPluginLV2.cpp:6764)
==3651137==    by 0x4A687D: CarlaBackend::CarlaPlugin::newLV2(CarlaBackend::CarlaPlugin::Initializer const&) (CarlaPluginLV2.cpp:8238)
==3651137==    by 0x40D193: CarlaBackend::CarlaEngine::addPlugin(CarlaBackend::BinaryType, CarlaBackend::PluginType, char const*, char const*, char const*, long, void const*, unsigned int) (CarlaEngine.cpp:693)
==3651137==    by 0x4DEDE0: carla_add_plugin (CarlaStandalone.cpp:1247)
==3651137==    by 0x408D40: main (CarlaBridgePlugin.cpp:675)
==3651137== 

[falkTX]

A few more comments:

• plugin does not respond to all-notes-off, which can lead to some stuck notes
• it should have at least 1 preset, could be "default"
• takes a while to load, because "retuning" is done on start. but any subsequent change pushes such changes to worker thread. this is a little inconsistent, and if loading a PB with some tweaked parameters makes the plugin take 2x what it should before it produces sound. a suggestion is to make plugin load right away and use worker thread on 1st run.
• retuning LED is a bit too simplistic, I suggest to make it blink/animate (can be done with pure CSS classes)
• related to previous point, would be great to have it use https://github.com/moddevices/mod-lv2-extensions/blob/main/mod-hmi.lv2/mod-hmi.h to let user know when retuning is taking place (assuming a suloea parameter is addressed to the device/HMI). see https://github.com/DISTRHO/DIE-Plugins/blob/main/plugins/a-fluidsynth.lv2/a-fluidsynth.cc for inspiration, look around MOD_EXTENDED bits

[falkTX]

That said, we can already push the plugin to beta if you wish, so it gets wider testing.
Just let us know if that is something you want.

[paulfd]

I'll look into it, thanks for the comments! No need to publish it until it's in shape, no worries.

[falkTX]

@paulfd I would like to move with this plugin, do you have time to have a look at the points I mentioned before? If not, would a patch/PR be acceptable?