[clang] [analyzer] Fix false positive when a lifetimebound method is called during destruction (PR #210801)

[Benedek Kaibas via cfe-commits]

benedekaibas wrote:

> Oh, I think this might not be the false positive fix I anticipated.

During our meeting I talked about the false positive case for move operations, but that got actually fixed during the refactoring of the checkers. I meant that we have a current problem with destructors not being understood by the checker. I messed the two cases of false positives, my bad. Now both of them are fixed.


https://github.com/llvm/llvm-project/pull/210801