aaron.ballman added inline comments.
================
Comment at: clang/include/clang/Basic/AttrDocs.td:3525
+ let Content = [{
+On classes, the ``equality_operator_compares_members_lexicographically``
+attribute informs clang that a call to ``operator==(const T&, const T&)`` is
----------------
The docs don't make it clear what the ramifications are of using the attribute.
Do you get better codegen? Different diagnostics? That sort of thing.
It'd probably help for the docs to have a short code example showing what the
difference the attribute makes.
Also, should we explicitly define it as UB if the user applies the attribute to
something and the equivalence requirements aren't met?
If the attribute is written on a structure that's used as a base class, does
the derived class automatically "inherit" the behavior as well? Or do derived
classes have to be redeclared with the attribute? Is it reasonable for a base
class which does not have the attribute to be inherited by a derived class
which adds the attribute?
(I'm kind of wondering whether there are situations in which we want to
diagnose use of the attribute to help programmers avoid mistakes.)
================
Comment at: clang/include/clang/Basic/AttrDocs.td:3531-3533
+On enums, the attribute guarantees that there are either no custom comparison
+operators defined, or any that defined are quivalent to the builtin comparison
+operator.
----------------
================
Comment at:
clang/test/SemaCXX/attr-equality-operator-compares-members-lexicographically.cpp:1
+// RUN: %clang_cc1 -triple x86_64-apple-darwin10 -fsyntax-only -verify %s
-std=c++20
+
----------------
Is the triple necessary?
================
Comment at:
clang/test/SemaCXX/attr-equality-operator-compares-members-lexicographically.cpp:72
+
+// TODO: This is trivially equality comparable
+struct [[clang::equality_operator_compares_members_lexicographically]]
NotTriviallyEqualityComparableBitfieldFilled {
----------------
FIXME?
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D151625/new/
https://reviews.llvm.org/D151625
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
