[Bug c++/5660] dependant names involving templates & inheritance

[Bug c++/5660] dependant names involving templates & inheritance

[pinskia at physics dot uc dot edu]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=5660


pinskia at physics dot uc dot edu changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|2003-05-06 00:00:00         |2003-08-01 20:14:38
               date|                            |


------- Additional Comments From pinskia at physics dot uc dot edu  2003-08-01 20:14 -------
On the mainline (20030801):
It still prints `In base::f()'.

[Bug c++/5660] dependant names involving templates & inheritance

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-04-28 01:40 -------

template<typename T> struct W : public T
{
    void g( W& w )
    {
        f(w);  // <-- this is dependent should it not wait until instantiation to figure out what 
function it should call?
    }
}; 


Maybe I am just misreading the standard but I think GCC is right here.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|2004-02-23 13:33:47         |2004-04-28 01:40:28
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=5660

[Bug c++/5660] dependant names involving templates & inheritance

[bangerth at dealii dot org]

------- Additional Comments From bangerth at dealii dot org  2004-04-28 06:51 -------
Interestingly enough, icc prints 
  g/x> icc x.cc 
  g/x> ./a.out  
  In base::f() 
just like gcc, but _if I switch on the strict mode_, we get 
  g/x> icc -Xc -ansi x.cc 
  g/x> ./a.out  
  In ::f() 
 
That confuses even me... 
 
W. 

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=5660

[Bug c++/5660] dependant names involving templates & inheritance

[giovannibajo at libero dot it]

------- Additional Comments From giovannibajo at libero dot it  2004-04-28 11:13 -------
(In reply to comment #4)

>         f(w);  // <-- this is dependent should it not 
> wait until instantiation to figure out what 
> function it should call?

No, the function name "f" is non-dependent, and unqualified. w is dependent 
(its type is W<T>), but this only means that an additional koenig lookup is 
performed at instantiation time in the associated namespaces. Since the only 
namespace associated is the global one, nothing new is found.

Net result is that ::f() is called. 

This is really no different from the example shown in changes.html for GCC 
3.4.0 ("non dependent names are not looked up in dependent bases"). It would be 
very surprising and wrong if changing a parameter type from (say) "int" 
to "W<T>" made another different function to be called.


> just like gcc, but _if I switch on the strict mode_, we get 
>   g/x> icc -Xc -ansi x.cc 
>   In ::f() 

It's possible that icc non-strict is using MS- or GCC- compatibility mode, 
which disables two-stage name lookup.


::f() is the function that should be called. BTW, you can turn this into a link-
time test by removing the body of base::f().


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=5660

[Bug c++/5660] dependant names involving templates & inheritance

[bangerth at dealii dot org]

------- Additional Comments From bangerth at dealii dot org  2004-04-28 16:33 -------
Giovanni, you're right of course. So this is one of the few 
places where we don't get name lookup right. 
 
Here's the link test: 
------------ 
template<typename T> void f( T& aT ) {} 
 
struct base {  
    void f( base& ); 
}; 
 
template <typename T> struct W : public T { 
    void g( W& w ) { f(w); } 
}; 
 
int main() {  
    W<base> wb;  
    wb.g(wb);  
    return 0; 
} 
---------------- 
 
icc -Xc -ansi links this fine, gcc's call to ld is missing the function that 
the compiler erroneously calls. 
 
W. 

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=5660

[Bug c++/5660] dependant names involving templates & inheritance

[giovannibajo at libero dot it]

------- Additional Comments From giovannibajo at libero dot it  2004-04-29 00:34 -------
I forgot to say, this is also DR 213, which has TC1 status, so it's approved as 
an official defect report of C++98. The resolution adds this text to the 
standard to clarify the meaning:

"In the definition of a class template or a member of a class template, if a 
base class of the class template depends on a template-parameter, the base 
class scope is not examined during unqualified name lookup either at the point 
of definition of the class template or member or during an instantiation of the 
class template or member."


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=5660

[Bug c++/5660] [DR213] dependant names involving templates & inheritance

[giovannibajo at libero dot it]

------- Additional Comments From giovannibajo at libero dot it  2004-04-29 00:38 -------
In fact, a test for this is already present in g++.dg/tc1/dr213.C in mainline, 
and it is xfailed at the moment.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|dependant names involving   |[DR213] dependant names
                   |templates & inheritance     |involving templates &
                   |                            |inheritance


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=5660

[Bug c++/5660] [DR213] dependant names involving templates & inheritance

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-08-03 01:45 -------
To me, PR 2922 looks very much related.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
  BugsThisDependsOn|                            |2922


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=5660

[Bug c++/5660] [DR213] dependent names involving templates & inheritance

[redi at gcc dot gnu dot org]

------- Additional Comments From redi at gcc dot gnu dot org  2004-08-11 08:40 -------
I believe this should also have keyword accepts-invalid, as shown by the
following testcase, which should fail to compile but is accepted by current GCC

   template <typename T>
     struct A
     {
         typedef T Type;
         void f(Type) {}
     };

   template <typename T>
     struct B : A<T>
     {
         typedef typename A<T>::Type Type;
         void ff(Type t) { f(t); }         // XXX
     };

   template struct B<int>;

The call to f(t) shouldn't find a match, but finds A<T>::f

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |redi at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=5660

[Bug c++/5660] [DR213] dependent names involving templates & inheritance

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-06-12 19:06 -------
*** Bug 22040 has been marked as a duplicate of this bug. ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sebor at roguewave dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=5660

[Bug c++/5660] [DR213] dependent names involving templates & inheritance

[pinskia at gcc dot gnu dot org]

-- 
Bug 5660 depends on bug 2922, which changed state.

Bug 2922 Summary: [DR 197] two-stage lookup for unqualified function calls with type-dependent arguments
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=2922

           What    |Old Value                   |New Value
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=5660

[Bug c++/5660] [DR213] dependent names involving templates & inheritance

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-09-04 18:28 -------
(In reply to comment #11)
> I believe this should also have keyword accepts-invalid, as shown by the
> following testcase, which should fail to compile but is accepted by current GCC
That looks like a different bug as all other problems here are fixed.  Since the orginal example in 
comment #0 is fixed, I am closing as fixed for 4.1.0.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|---                         |4.1.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=5660

[Bug c++/5660] [DR213] dependent names involving templates & inheritance

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-09-20 16:43 -------
*** Bug 23982 has been marked as a duplicate of this bug. ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mmarek at suse dot cz


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=5660