http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58569
--- Comment #15 from Daniel Krügler ---
(In reply to Daniel Krügler from comment #14)
Furthermore the name hiding happening here
template>>
Function(F);
needs to be fixed (it was not in your previous example):
template>>
Function(F);
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58569
--- Comment #14 from Daniel Krügler ---
(In reply to Jonathan Wakely from comment #13)
> Argh, yes, I didn't read the diagnostic properly after that last change,
> sorry!
Thanks. I think the actual problem can be resolve by moving
template
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58569
--- Comment #13 from Jonathan Wakely ---
Argh, yes, I didn't read the diagnostic properly after that last change, sorry!
Corrected:
struct false_type { enum { value=0 }; };
struct true_type { enum { value=1 }; };
template
struct add_rvalue_re
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58569
--- Comment #12 from Daniel Krügler ---
(In reply to Jonathan Wakely from comment #11)
> Further reduced (clang gives the same error for this version):
>
> template struct enable_if { typedef void type; };
> template struct enable_if { };
Jonath
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58569
--- Comment #11 from Jonathan Wakely ---
Further reduced (clang gives the same error for this version):
struct false_type { enum { value=0 }; };
struct true_type { enum { value=1 }; };
template
struct add_rvalue_reference {
using type = T&
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58569
--- Comment #10 from Jonathan Wakely ---
(In reply to Jonathan Wakely from comment #9)
>
> // template struct CheckResult :
> std::true_type { };
Sorry, bugzilla wrapped that line, you'll need to join those lines or comment
out the std::true_t
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58569
--- Comment #9 from Jonathan Wakely ---
Somewhat reduced:
#include
template
inline _Functor&
__callable_functor(_Functor& __f)
{ return __f; }
template struct Function;
template
struct Function
{
template
using Invoke = decltype(_
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58569
--- Comment #8 from Jonathan Wakely ---
This might be a front-end bug, I'm struggling to reduce it though.
If I produce prepreocessed code then clang++ compiles it fine.
Adding this to the global namespace (so that version gets found by unqualif
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58569
--- Comment #7 from Daniel Krügler ---
(In reply to Jonathan Wakely from comment #5)
> And it should only happen when instantiating the converting constructor
> template or converting assignment operator function, not when instantiating
> the clas
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58569
--- Comment #6 from Daniel Krügler ---
(In reply to Jonathan Wakely from comment #4)
Why would LWG 2132 require any such check during the instantiation of the class
template or during the instantiation of the default constructor? The
constraints
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58569
--- Comment #5 from Jonathan Wakely ---
And it should only happen when instantiating the converting constructor
template or converting assignment operator function, not when instantiating the
class template.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58569
--- Comment #4 from Jonathan Wakely ---
(In reply to Daniel Krügler from comment #3)
> I don't see a good
> reasons why std::function should check std::is_convertible foo> when instantiating the class template.
It's done to check the constraint i
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58569
--- Comment #3 from Daniel Krügler ---
(In reply to Paolo Carlini from comment #1)
[Somehow bugzilla lost my response]
Yes, foo is incomplete within a data member declaration of type foo and this
breaks std::is_convertible, because the arguments
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58569
--- Comment #2 from Jonathan Wakely ---
We can fix it in the library by adding this to std::function
template
struct _CheckResult<_CallRes, _CallRes>
: true_type { };
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58569
--- Comment #1 from Paolo Carlini ---
We badly need a reduced testcase for this. The issue doesn't seem to involve
std::function in a fundamental way, seems more about using
std::is_convertible inside foo.
Adding Daniel in CC, in case he wants to
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58569
Jonathan Wakely changed:
What|Removed |Added
Status|UNCONFIRMED |NEW
Last reconfirmed|
16 matches
Mail list logo
