AaronBallman wrote: > @AaronBallman See test results from compile-time-tracker here: > https://llvm-compile-time-tracker.com/compare.php?from=693a458287d019c5c6a66fe3019d099df2978cdb&to=dbb5e29d136a18060ba6759b328ad80fa9cea649. > > It looks like that there is a statistically meaningful difference, but it's > only about 0.05..0.25% depending on the test. Is it considered OK?
Yeah, this seems to have noticeable impact on compile times for every compilation; out of curiosity, have you tried an approach where this information is stored in `ExtQuals` instead? That's heap allocated, but would mean that the only folks paying the cost are the ones using the functionality. https://github.com/llvm/llvm-project/pull/84384 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
