rsmith added a comment.
This is an interesting idea. Have you looked into whether you can defer the
check for conversion via multiple user-defined conversions until we come to
call `CompleteNonViableCandidate` after finding no viable functions? That'd
give me a lot more confidence that this won't introduce a noticeable
performance regression for pathological cases (eg, considering a quadratic --
or worse -- number of combinations of constructors and conversion functions).
================
Comment at: clang/include/clang/Basic/DiagnosticSemaKinds.td:4068
+ "%diff{from $ to $|from argument type to parameter type}3,4 "
+ "applies more than one implicit user-defined conversion"
+ >;
----------------
applies -> would apply
================
Comment at: clang/include/clang/Sema/Sema.h:2947-2956
+ enum UserDefinedConversionsKind {
+ /// Allow user-defined conversions.
+ UDC_Allow,
+
+ /// Find user-defined conversions but mark as invalid.
+ UDC_Suppress,
+
----------------
Please use an `enum class` here, so that call sites are more meaningful
(`UserDefinedConversions::Allow` vs `UDC_Allow`).
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D74009/new/
https://reviews.llvm.org/D74009
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
