prj- wrote: @owenca, I actually think that this introduced a regression that was previously fixed in the nightly snapshot of `clang-format-21`. With `clang-format-20`: ``` $ cat foo.cpp template <class F, class... Args> using invoke_result_t = typename std::result_of<F&& (Args &&...)>::type; $ clang-format-20 foo.cpp template <class F, class... Args> using invoke_result_t = typename std::result_of<F&& (Args &&...)>::type; ``` With `clang-format-21` prior to `++20250509052510+a6385a87a2e5-1~exp1~20250509172712.2830`, e.g., `++20250501111520%2B21aa7b8138a1-1\~exp1\~20250501111702.2812`: ``` $ clang-format-21 foo.cpp template <class F, class... Args> using invoke_result_t = typename std::result_of<F &&(Args &&...)>::type; ``` But now, we are back to the wrong formatting of `clang-format-20`.
https://github.com/llvm/llvm-project/pull/138633 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
