> We have in one hand beat maps, where basically everything beat is a \"change marker\". On the other hand we have real \"bar changes\" (Taktwechsel) that would be annotated as a double bar line.
...
Ah, I just spotted it now. Well the problem is that we do call `delete` on it. So its very easy to mistake it as properly owning. I also fear that the pointer can easily outlive the RenderingEngine...
Any reason why introducing a smart pointer on a Qt-owned reference here? Just in case you've missed the comment, the pointer is managed by the QML scene, we don't have any control over it.
lets avoid introducing manual memory management. You can wrap this in a unique_ptr. If that doesn't work because QHash needs the key to be copy-able, use `std::unordered_map` instead.
yeah, but I don't think its worth the complexity. Taking a quick glance at the source, there is no heap allocation or anything else expensive that is worth reusing. We can KISS for now and optimize...
### Bug Description
I was running Mixxx, and using my Traktor X1 MK2 with a Traktor Audio 6 interface and a AKAI Midimix controller, when out of no where, my X1 MK2 freezes (lights stay on) and al...
for sure, we want to get this in and don't want to have forks that need maintenance. I do like the approach of using a controller setting to enable / disable the new API. That limits the blast radi...
> Oh no, since 6.4, the set is opened with `QStringConverter`. That's why my last commit uses `QTextCodec` prior to 6.4. Virtually any encoding supported by UCI should be available. See my last com...
Actually, there is probably a little more we can do here. Apparently there are more optional codecs available using [`QStringConverter::availableCodecs()`](https://doc.qt.io/qt-6/qstringconverter.h...