[clang] Clean up array allocation logic in Sema::BuildCXXNew (PR #186617)

[Roberto Turrado Camblor via cfe-commits]

rturrado wrote:

> I don't believe this is a good change sorry - it's a relatively large abstraction layer to "simplify" maybe two actual expressions.
> 
> If we really wanted to do something here I would probably just replace the optional<> with
> 
> ```c++
> bool IsArraySizeDependent = false;
> bool ... = false;
> if (ArraySize) {
>   set all the "there is an array size" variables as appropriate
> }
> ```
> 
> It seems the main concern is the compound conditions, which is consistent with what other code does, but I would prefer to reduce that in general.
> 
> But the issue here is not made better with an abstraction (the better option would be actual language support for optional in c++)
> 
> I've marked as "request changes" but I think this is probably the wrong tack, and I'm not sure the problem with new/delete Sema is this bit. There's so much more that is so much worse.
> 
> While not an option for clang any time soon, I have wondered if reflection could be used in a way to make this kind of thing more expensive.

Many thanks for the thorough review. And from my side, same comments I did to @tbaederr. You're very right that this current code is overkill. I will explore the variant solution.

https://github.com/llvm/llvm-project/pull/186617