rjmccall added inline comments.
================ Comment at: clang/include/clang/AST/ASTContext.h:1095 CanQualType FloatComplexTy, DoubleComplexTy, LongDoubleComplexTy; - CanQualType Float128ComplexTy; + CanQualType Float128ComplexTy, Ibm128ComplexTy; CanQualType VoidPtrTy, NullPtrTy; ---------------- Huh. I don't know why we precompute these. I mean, it's not a huge start-up overhead for the compiler, but it's not nothing, and the use in `getFloatingTypeOfSizeWithinDomain` is not at all worth optimizing. Would you mind writing up a patch that rips all of these out and makes the use-sites just build the appropriate complex type as needed? And then we can land a patch for this that just does the one-liner to allow `_Complex __ibm128`. ================ Comment at: clang/lib/Sema/Sema.cpp:1898 + (&Sem == &llvm::APFloat::PPCDoubleDouble() && + !Context.getTargetInfo().hasIbm128Type())) LongDoubleMismatched = true; ---------------- Why is this change in this patch? Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D109948/new/ https://reviews.llvm.org/D109948 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits