[clang] [analyzer] [MallocChecker] suspect all release functions as candite for supression (PR #104599)

[Donát Nagy via cfe-commits]

https://github.com/NagyDonat commented:

I like the overall idea of this commit, but I noticed that it may be a step backwards in some situation where there is a destructor on the stack and the old code would've suppressed the result, but the new code doesn't suppress it. (See inline comments for details.)

I also suggested some code simplifications and comment cleanup -- this logic was heavily commented, and your changes don't update all the comments that refer to 'destructors'. (Deleting some comments is also a valid choice, if you feel that that they are not useful enough and keeping them up-to-date is just a pointless chore.)

https://github.com/llvm/llvm-project/pull/104599