[clang] [analyzer] Revert #115918, so empty base class optimization works again (PR #157480)

[Donát Nagy via cfe-commits]

NagyDonat wrote:

Sounds reasonable.

I feel a bit guilty that I wasn't able to predict this regression when I reviewed the reverted commit as "LGTM, clean little patch", but I don't think that I could've done better without investing drastically more effort into the review.

By the way does this revert affect any of the commits that were related to the reverted one?

https://github.com/llvm/llvm-project/pull/157480