================ ---------------- ellishg wrote:
I think this makes a code a little harder to follow. If you are concerned about performance, I am assuming `isTargetStubsAndInRange()` will be inlined, or at least partially inlined, so that if `estimatedStubsEnd` is `std::nullopt` then the rest of the check will be skipped. In general I am a fan of avoiding premature optimization. If we want to improve performance, then we should profile with a tool and see where the bottleneck is. But I'm pretty sure other parts of the linker are slower than this for large binaries. https://github.com/llvm/llvm-project/pull/197049 _______________________________________________ llvm-branch-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
