================ @@ -0,0 +1,7 @@ +// RUN: %clang_analyze_cc1 -std=c++20 -analyzer-checker=optin.performance -verify %s ---------------- redcode wrote:
> Have you used AI, and read the policies around that? I have not used AI. > What alternatives have you considered? To be honest, I picked the field-level check because it was the direct fix for the crash, as it happens while laying out a field whose type is dependent, and `IsTrickyField` is already the place that skips records whose fields can't be laid out. The `isDependentType()` bail-out at the top of `shouldSkipDecl` already exists but was not catching this; I hadn't chased down why at the time. I didn't really weigh the traversal-level option before opening the PR. > I presume the checker already have at least one test file. I think we should > squeeze this into one of those tests and suffix the line with `// no-crash` > to illustrate that this used to crash. I could add it to `clang/test/Analysis/padding_cpp.cpp` but I would need to bump it to C++20, since the repro needs abbreviated function templates. Is it OK? https://github.com/llvm/llvm-project/pull/232986 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
