[clang] [analyzer] Don't fold wide string literals in CStringChecker (PR #213281)

[Balázs Benics via cfe-commits]

steakhal wrote:

> Would it be more reasonable to fix endiannes of the buffer returned by `getBytes` with respect of `Lit->getCharByteWidth()`?

I've considered. But it's not worth it IMO. If I'm wrong, we will see bug reports, but I doubt that.
basically, `u`, `u8`, and some other prefixes would work, it's just a handful that would be not supported.
But I don't particularly think it's important. If you feel like it, you could implement the proper endianness handling.

https://github.com/llvm/llvm-project/pull/213281