benedekaibas wrote: > @benedekaibas Can you post an example FP html finding here? I could maybe > reproduce and create a test case for you in the background.
[report-3efbc2.html](https://github.com/user-attachments/files/30700407/report-3efbc2.html) In the meantime I could reproduce the error. As a note for the AI policy, after my implementation of the test case that triggered the FP I asked AI to ask me questions which would help me to reduce the number of lines since I had a pretty long test case. This code snippet is the result of that refactor: ``` namespace repro { struct PunnedPointer { unsigned char Data[sizeof(void *)]; PunnedPointer(long V = 0) { *this = V; } PunnedPointer &operator=(long V) { __builtin_memcpy(Data, &V, sizeof(Data)); return *this; } }; } // namespace repro struct PointerUnion { repro::PunnedPointer Val; PointerUnion(const int *V) { Val = reinterpret_cast<long>(V); } }; void foo(PointerUnion arg); void no_dangling_by_value_argument() { int local = 0; foo(&local); // no-warning } ``` https://github.com/llvm/llvm-project/pull/213779 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
