[PATCH] D77150: [Analyzer] New Option for ContainerModeling: AggressiveEraseModeling

[Balogh, Ádám via Phabricator via cfe-commits]

baloghadamsoftware added a comment.

I am trying to measure the amount of extra false positive on real code (not artificial examples) of this option. The modeling of container size may reduce it, of course, but we have no other option to find this bug: even loop-unrolling or loop-widening is not be able to recognize the last before the last iteration of a loop. I cannot imagine any solution to recognize it, this really sounds like voice recognition in the 80's as @NoQ wrote. I do not expect such thing in the next 10 years. Refining and enhancing container modeling is expected to reduce false positives generally in the iterator checkers and will eventually result in moving them out of alpha state.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D77150/new/

https://reviews.llvm.org/D77150