[Core] fix frame int_cast for 24 bit types

[MFransen69]

Some remarks.

Okay, see the removed assert below, in the end the whole cast did not add too much value but okay it is there now, working correctly for re-use :)

I fixed it locally in Frame as the existing Max infrastructure was already there so that seemed the easiest solution (alternative could have been to add a std::numeric_limits<>::min and max specialization for SInt24/UInt24 but seen the already existing Max infrastructure that did not make much sense, but this as a hint might we have a future similar challenge with int_cast)

I left the int_cast in Number.h, you might never know if someone would need it in the future but I think it is unused now (considered adding a static_assert to check if NEW_TYPE is a scalar_type to prevent a similar issue from going unnoticed but as numeric_limits can be used with more types, e.g. with custom specializations, I did not do that)
One remark I see that int_cast is now in the global namespace, guess because you can then use it like the other language casts (like static_cast) but I would just place it inside Thunder::Core. The name is quite generic so possible collision can easily happen and what is the problem with typing Core::int_cast, then it is at least also clear we created it :) but okay my 2 cents.
(oh and I was a little lazy, the type for the Frame::Min must next to the Min member also have a Max member now, which I think is logical, and the check even makes it better and now, min cannot be higher than max :) , but of course you could solve this without that requirement but that was more work and not needed)

Oh I also sneaked in the Request::INCOMPLETE: , almost forgot about it. I'll create the Jira issue once this PR is merged so I can link to the code in the issue