olestrohm added inline comments.
================
Comment at: clang/lib/Sema/SemaCast.cpp:2359
- if (SrcType == DestType) {
+ if (SrcType == Self.Context.removeAddrSpaceQualType(DestType)) {
// C++ 5.2.10p2 has a note that mentions that, subject to all other
----------------
rjmccall wrote:
> I think the upshot of the conversation Richard and I just had is that there's
> a bug here for all qualifiers; that is, `reinterpret_cast<const int>(5)`
> should be allowed. Probably the right thing to do is to strip qualifiers
> from DestType in the CastOperation constructor when the type is not a class
> or an array. Richard, do you agree? This isn't strictly implied by the C++
> wording since the semantic analysis of the cast is logically prior to the
> introduction of a qualifier pr-value expression.
Ah, I see, I did indeed misunderstand the conversation above. I see now that
with what Richard mentioned with `If a prvalue initially has the type “cv T”,
where T is a cv-unqualified non-class, non-array type, the type of the
expression is adjusted to T prior to any further analysis.` it makes sense to
strip qualifiers for those cases. I also checked out gcc to compare to another
implementation, and it does work there, so this behaviour doesn't seem
unreasonable.
I'll rework this to strip all qualifiers somewhere in the CastOperation
constructor as you suggested then.
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D102689/new/
https://reviews.llvm.org/D102689
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
