http://gcc.gnu.org/bugzilla/show_bug.cgi?id=99
Manuel López-Ibáñez <manu at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |dodji at gcc dot gnu.org
--- Comment #22 from Manuel López-Ibáñez <manu at gcc dot gnu.org> ---
(In reply to Manuel López-Ibáñez from comment #21)
> (In reply to Christian Ehrhardt from comment #5)
> > From: "Christian Ehrhardt" <[email protected]>
> > To: [email protected], [email protected], [email protected],
> > [email protected], [email protected], [email protected]
> > Cc:
> > Subject: Re: c++/99: [2003-03-26] Bug in type in error message.
> > Date: Tue, 6 May 2003 15:04:23 +0200
> >
> > Here's a further reduced testcase, along with some more analysis:
> >
> > template<typename S> class X {};
> > template<typename Q> int f(X<int>, X<Q>);
> > template<typename B> int f(X<B>, X<int>);
> >
> > int main(void) {
> > return f(X<int>(), X<int>());
> > }
> >
>
> For this testcase, the problem is that:
>
> (gdb) p debug_tree(arglist)
> <tree_vec 0x7ffff7570d40
> elt 0 <template_type_parm 0x7ffff7566348 Q VOID
> align 8 symtab 0 alias set -1 canonical type 0x7ffff7566150
> index 0 level 1 orig_level 1
> chain <type_decl 0x7ffff755fc38 Q>>>
>
> and
>
> (gdb) p debug_tree(arglist)
> <tree_vec 0x7ffff7570e60
> elt 0 <template_type_parm 0x7ffff7566690 B VOID
> align 8 symtab 0 alias set -1 canonical type 0x7ffff7566150
> index 0 level 1 orig_level 1
> chain <type_decl 0x7ffff755ff18 B>>>
>
> have the same hash:
>
> (gdb) p hash
> $128 = 3284467468
>
> I am not sure whether we should change the hashing function or whether we
> should add some additional check. But what kind of check?
And this triggers also with simple typedefs:
template<typename S> class X {};
template<typename Q> int f(X<int>, X<Q>);
template<typename B> int f(X<B>, X<myint>);
int foo(void) {
return f(X<int>(), X<int>());
}
gives:
/home/manuel/test1/pr99.cc:8:30: error: call of overloaded ‘f(X<int>, X<int>)’
is ambiguous
return f(X<int>(), X<int>());
^
/home/manuel/test1/pr99.cc:8:30: note: candidates are:
/home/manuel/test1/pr99.cc:4:26: note: int f(X<int>, X<Q>) [with Q = int]
template<typename Q> int f(X<int>, X<Q>);
^
/home/manuel/test1/pr99.cc:5:26: note: int f(X<Q>, X<int>) [with B = int]
template<typename B> int f(X<B>, X<myint>);
^
Perhaps the only solution is to not pretty-print the candidates, which is
exactly what clang does:
pr99.cc:8:10: error: call to 'f' is ambiguous
return f(X<int>(), X<int>());
^
pr99.cc:4:26: note: candidate function [with Q = int]
template<typename Q> int f(X<int>, X<Q>);
^
pr99.cc:5:26: note: candidate function [with B = int]
template<typename B> int f(X<B>, X<myint>);
^
Dodji, as the diagnostics maintainer, would that be an acceptable fix?
