Filippo Cucchetto
Filippo Cucchetto
@arkanoid87 No i'm ok in a workaround. Maybe i'm wrong but i recall having some issue in removing the finalizers forward declarations. Compiler seemed buggy in there. I had some...
the feature is experimental for a reason i would say...we need a proper fix with forward declarations imho
i don't get why you want those changes and why it metters here. Again this is a compiler problem not of NimQml. If ORC/ARC is broken is not our fault
The leaks are due to the compiler generated code not again due to NimQml. Just use refc since it's the default GC even for Nim 1.6. The use of include...
I would not accept: - changing setup and delete to be methods - changing the signature of setup/delete due to destructors needing a `var` argument instead of a ref -...
Thank you I’ll take a look as soon as possible. However due to my work day job i can’t assure you to give this a look until this weekend Il...
Just store the qvariant instead of the float or store both > Il giorno 11 gen 2022, alle ore 19:52, arkanoid87 ***@***.***> ha scritto: > > > np, I...
@arkanoid87 simply don't delete nothing if don't need explicit deletion. This will be due automatically by nim. Otherwise if you need more deterministic deletion use ARC or ORC. That said:...
Regarding why most of examples have the explicit delete is just because initial release of nim use a not deterministic GC. If you latest version and use ORC or ARC...
If you test leaks with valgrind you need: 1) To understand where leaks are coming 2) Qml use a Javascript engine that is not deterministic so memory will be delete...