https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97922
Bug ID: 97922
Summary: inline namespace caused confusing error message
Product: gcc
Version: 10.1.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c++
Assignee: unassigned at gcc dot gnu.org
Reporter: harry.linxd at gmail dot com
Target Milestone: ---
For a type that's defined in inline namespace, forward declaration and
definition appear in the same scope cause confusing error messages, while
clang++'s diagnostic messages are more clear
```
$cat type1.h
namespace test {
inline namespace internal {
struct Type1 {
Type1() = default;
};
}
}
```
```
$cat test.cpp
#include <utility>
#include "./type1.h"
namespace test {
struct Type1; //Type1 is actually in an inline namespace
}
using Type = std::pair<test::Type1, int>;
void f(Type& t);
```
```
$ g++ -I. test.cpp
test.cpp:8:40: error: template argument 1 is invalid
8 | using Type = std::pair<test::Type1, int>;
| ^
test.cpp:10:6: error: variable or field 'f' declared void
10 | void f(Type& t);
| ^
test.cpp:10:8: error: 'Type' was not declared in this scope
10 | void f(Type& t);
| ^~~~
test.cpp:10:14: error: 't' was not declared in this scope
10 | void f(Type& t);
| ^
```
Here gcc's error message is sort of confusing, while clang's diagnostic message
is more straight:
```
$ clang++ -I. test.cpp
test.cpp:8:30: error: reference to 'Type1' is ambiguous
using Type = std::pair<test::Type1, int>;
^
././type1.h:3:8: note: candidate found by name lookup is
'test::internal::Type1'
struct Type1 {
^
test.cpp:5:8: note: candidate found by name lookup is 'test::Type1'
struct Type1;
^
1 error generated.
```
