Sirraide wrote: > > Iirc when I was testing this it did 1 million expansions w/o hitting any > > existing limits; that was w/ an empty expansion statement but that already > > took a few seconds. > > That's fine - the limits are there to avoid overwhelming the compiler, not > because something might be slow
I’m more worried about it taking even longer and someone mistakenly thinking that we have a bug in Clang or that the compiler is hanging when the bug would really be in user code in this case; iirc @AaronBallman has said in the past that the `constexpr` steps limit exists for a similar reason, so I thought it’d make sense to implement something similar for this. (side note: I don’t believe 1 million expansions would ever not be a bug candidly...) https://github.com/llvm/llvm-project/pull/169689 _______________________________________________ llvm-branch-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
