http://llvm.org/bugs/show_bug.cgi?id=22426
Bug ID: 22426
Summary: generic lambdas, decltype(auto), and rvalue
references, oh my!
Product: clang
Version: trunk
Hardware: PC
OS: Windows NT
Status: NEW
Severity: normal
Priority: P
Component: C++14
Assignee: [email protected]
Reporter: [email protected]
CC: [email protected]
Classification: Unclassified
Clang (trunk) rejects the following code (with -std=c++1y):
#include <utility>
int main()
{
using std::pair;
using std::declval;
using X = decltype(declval<pair<int&&,int&&>>().first);
auto f = [](auto && p) -> decltype(auto) //((decltype(p)&&)p).first)
{
return ((decltype(p)&&)p).first;
};
using Y = decltype(f(declval<pair<int&&,int&&>>()));
}
The error I get is:
main.cpp:13:16: error: rvalue reference to type 'int' cannot
bind to lvalue of type 'int'
return ((decltype(p)&&)p).first;
^~~~~~~~~~~~~~~~~~~~~~~~
There should *never* be an error when using decltype(auto). The compiler should
use the actual type of the return expression as the return type. It seems clang
is getting confused about the value category here.
I observe that X is an alias for int&& in the above code, and I feel that the
generic lambda is doing the same thing. Both the type of the expression in the
lambda and the return type of the lambda should also be int&&.
I also observe that if I replace decltype(auto) with
decltype(((decltype(p)&&)p).first), then the code compiles and Y is an alias
for int&&, which seems right to me.
For the record, gcc also gets this code wrong, but in a different way.
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
LLVMbugs mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/llvmbugs
