jmorse wrote: > There are a few large servers (the Borg orchestration server, some YouTube > servers) that had notable regressions in both linker memory and final binary > size with debug info. I was able to verify that with these changes to > constructor homing eligibility, linker memory + binary size returned to > normal with the experiment enabled.
Sorry to be obtuse, but could you quantify the size of the savings (percentage change in `.debug_info` perhaps)? If this turns out to cause difficulties for other developers then it's best to have good justification for the change, to reduce the chance of it being reverted in the future. https://github.com/llvm/llvm-project/pull/218165 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
