Mem leaks?

[patrolez]

Hello, I can see some C++ STL misconception in usage in the code.

https://github.com/PotatoOfDoom/CyberFSR2/blob/main/CyberFSR/CyberFsr.cpp#L29

Objects created dynamically by the new keyword while being referenced (stored) as pointers in STL containers (e.g. std::map, std::unordered_map, std::vector) are not being itself a subject of clearing/erasing done by the container (their destructor would not be triggered).

In this case you need to chose among:

1. to call explicitly delete on such pointer to delete the object, then remove the pointer entry from the container,
2. to use some smart pointer from the standard library unique_ptr/shared_ptr,
3. just store whole objects in an STL container (but not just pointers) and get familiar with reference (address) invalidation cases for this specific STL container - https://stackoverflow.com/questions/6438086/iterator-invalidation-rules-for-c-containers,
   • extracted suggestion that links to some tabular summary about references invalidation: https://en.cppreference.com/w/cpp/container#Iterator_invalidation,
   • at the first glance unordered_map seems to fit for Contexts, but not for pointers, but objects directly.
4. write own RAII variation of a smart pointer which would denote that a lifetime of the object is being managed by the container intentionally.

---

An STL container of pointer will NOT clean up the data pointed at

https://stackoverflow.com/a/383084/12755962

[PotatoOfDoom]

thanks for the heads-up. I'll fix it asap.