Revisiting the Intent of This Project

[stachurr]

Hello all,

I see many discussions about the caveats of this project (object lifetime, ownership, parsing LVGL, etc.), but no real discussion on the purpose/intent of this project. I also see that most attempts have died or gone stale. I think there are some things we should discuss before tackling this, such as:

1. What is the use case of a C++ binding? i.e. what should the binding offer that the C code does not?
2. Should the C++ objects to take ownership of the underlying lv_*_t? Or should they simply be containers which let you interact with the underlying object (no memory management)? The latter would essentially be object-oriented syntactic sugar.
3. Should C++ memory management (new/delete) replace the existing management methods?

I think after some of these questions are answered, we can begin to look at details.
Thank you.

[maciekr1234]

Hi,
in my opinion there is no purpose for C++ bindings, especially for embedded devices is usually asking for troubles with unsafe memory etc. I think if somebody would like to use lvgl in c++ they would write wrappers themselves, and after they did they would realise like I did and probably everybody who started doing it, that there is practically no benefits, and a whole lot of problems.
Another reason would be, there is many ways to do the same thing in c++, and everybody has different opinions.
I don't see benefits of c++ over c in lvgl, it has simple API which makes sense, and with today's tools, auto completion in IDE's it's easy to use lvgl in c, with a benefit of one less abstraction layer.