hubert.reinterpretcast added a comment.
In D142867#4108079 <https://reviews.llvm.org/D142867#4108079>, @eaeltsin wrote:
> The warning now fires even if overflow is prevented with `if constexpr`:
>
> if constexpr (width <= 64) {
> if constexpr (width == 64) {
> return 1;
> }
> return -static_cast<int64_t>(uint64_t{1} << (width - 1));
> }
>
> https://godbolt.org/z/M3xdcKd3M
For reference, actually placing the second return into an `else` block (thus
making it a discarded statement) does suppress the diagnostic:
https://godbolt.org/z/Kb6Md5PrK
I also question the focus on `if constexpr`. Replacing with `if (true) { return
1; }` should be as effective in the non-`else`-block case in suppressing the
warning in terms of QoI.
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D142867/new/
https://reviews.llvm.org/D142867
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
