whisperity added a comment. In D81272#2218175 <https://reviews.llvm.org/D81272#2218175>, @aaron.ballman wrote:
> While I agree with your observation about data and flow sensitivity, I > approved on the belief that the check as-is provides enough utility to > warrant adding it as-is. If someone wants to improve the check into being a > CSA check, we can always deprecate this one at that point. However, if there > are strong opinions that the check should start out as a CSA check because it > requires that sensitivity for your needs, now's a good time to bring up those > concerns. It's generally harder to create big logic mistakes when it comes to more complex expressions, assuming the user does't copy-paste (which I might have done, in the above example). We do not need to solve //every// potentially equivalent conditional (it is unsolvable in the generic case anyways). I'm sure this check can be improved later with handling trivial comparisons (such as standard library `int` results not being 0, -1, etc.). CHANGES SINCE LAST ACTION https://reviews.llvm.org/D81272/new/ https://reviews.llvm.org/D81272 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
