nickdesaulniers added a comment. In D76096#1922239 <https://reviews.llvm.org/D76096#1922239>, @rsmith wrote:
> it's substantially more efficient for CodeGen to walk the AST representation > (the `InitListExpr`) and directly generate an IR constant than it is to > create an `APValue` representation of the array. (`APValue` is not especially > space-efficient, and the extra copying and data shuffling can be quite slow.) Isn't that what my patch is doing? (Codegen walking the AST/`InitListExpr`, generating `Constant`s)? Uploading what I have; handling arrays is trickier than structs; the index and the base both end up having complex subtrees to fetch values from. As much fun as it is to build a compile time evaluator, it sounds like I should stop pursing `CGExprConstant` visitors in favor of `ExprConstant`? (I guess it's not clear to me whether @rsmith is in agreement with @eli.friedman on whether we want to be more aggressive in compile time evaluation or not). 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
