efriedma added a comment. Basically, I don't want order-of-magnitude compile-time regressions with large global variables. There are basically two components to that:
- That the fast path for emitting globals in C continues be fast. - That we rarely end up using evaluateValue() on large global arrays/structs in C, for reasons other than emitting them. If we end up calling evaluateValue() on most globals anyway, the codegen fast-path is a lot less useful; most of the cost of the APValue codepath is generating the APValue, not lowering it. This is not quite as concrete as I'd like, but I'm not sure how to describe it. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D76096/new/ https://reviews.llvm.org/D76096 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
