https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100393
--- Comment #1 from Dannii Willis <curiousdannii at gmail dot com> --- Okay, I've confirmed the regression myself, using functions_unsafe.i: gcc-8: real 0m11.450s gcc-10: real 4m46.472s And for comparison clang: real 0m0.711s